# FROM CONSUMER EXPERIENCE TO AFFECTIVE LOYALTY: CHALLENGES AND PROSPECTS IN THE PSYCHOLOGY OF CONSUMER BEHAVIOR 3.0

EDITED BY: María Pilar Martínez-Ruiz, Mónica Gómez-Suárez, Ana Isabel Jiménez-Zarco and Alicia Izquierdo-Yusta PUBLISHED IN: Frontiers in Psychology

#### *Frontiers Copyright Statement*

*© Copyright 2007-2018 Frontiers Media SA. All rights reserved. All content included on this site, such as text, graphics, logos, button icons, images, video/audio clips, downloads, data compilations and software, is the property of or is licensed to Frontiers Media SA ("Frontiers") or its licensees and/or subcontractors. The copyright in the text of individual articles is the property of their respective authors, subject to a license granted to Frontiers.*

*The compilation of articles constituting this e-book, wherever published, as well as the compilation of all other content on this site, is the exclusive property of Frontiers. For the conditions for downloading and copying of e-books from Frontiers' website, please see the Terms for Website Use. If purchasing Frontiers e-books from other websites or sources, the conditions of the website concerned apply.*

*Images and graphics not forming part of user-contributed materials may not be downloaded or copied without permission.*

*Individual articles may be downloaded and reproduced in accordance with the principles of the CC-BY licence subject to any copyright or other notices. They may not be re-sold as an e-book.*

*As author or other contributor you grant a CC-BY licence to others to reproduce your articles, including any graphics and third-party materials supplied by you, in accordance with the Conditions for Website Use and subject to any copyright notices which you include in connection with your articles and materials.*

> *All copyright, and all rights therein, are protected by national and international copyright laws.*

*The above represents a summary only. For the full conditions see the Conditions for Authors and the Conditions for Website Use.*

ISSN 1664-8714 ISBN 978-2-88945-412-9 DOI 10.3389/978-2-88945-412-9

### About Frontiers

Frontiers is more than just an open-access publisher of scholarly articles: it is a pioneering approach to the world of academia, radically improving the way scholarly research is managed. The grand vision of Frontiers is a world where all people have an equal opportunity to seek, share and generate knowledge. Frontiers provides immediate and permanent online open access to all its publications, but this alone is not enough to realize our grand goals.

### Frontiers Journal Series

The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access, online journals, promising a paradigm shift from the current review, selection and dissemination processes in academic publishing. All Frontiers journals are driven by researchers for researchers; therefore, they constitute a service to the scholarly community. At the same time, the Frontiers Journal Series operates on a revolutionary invention, the tiered publishing system, initially addressing specific communities of scholars, and gradually climbing up to broader public understanding, thus serving the interests of the lay society, too.

### Dedication to Quality

Each Frontiers article is a landmark of the highest quality, thanks to genuinely collaborative interactions between authors and review editors, who include some of the world's best academicians. Research must be certified by peers before entering a stream of knowledge that may eventually reach the public - and shape society; therefore, Frontiers only applies the most rigorous and unbiased reviews.

Frontiers revolutionizes research publishing by freely delivering the most outstanding research, evaluated with no bias from both the academic and social point of view. By applying the most advanced information technologies, Frontiers is catapulting scholarly publishing into a new generation.

### What are Frontiers Research Topics?

Frontiers Research Topics are very popular trademarks of the Frontiers Journals Series: they are collections of at least ten articles, all centered on a particular subject. With their unique mix of varied contributions from Original Research to Review Articles, Frontiers Research Topics unify the most influential researchers, the latest key findings and historical advances in a hot research area! Find out more on how to host your own Frontiers Research Topic or contribute to one as an author by contacting the Frontiers Editorial Office: researchtopics@frontiersin.org

## **FROM CONSUMER EXPERIENCE TO AFFECTIVE LOYALTY: CHALLENGES AND PROSPECTS IN THE PSYCHOLOGY OF CONSUMER BEHAVIOR 3.0**

Topic Editors:

**María Pilar Martínez-Ruiz,** University of Castilla-La Mancha, Spain **Mónica Gómez-Suárez,** Autonoma University of Madrid, Spain **Ana Isabel Jiménez-Zarco,** Open University of Catalunya, Spain **Alicia Izquierdo-Yusta,** University of Burgos, Spain

Image: Jacek Fulawka/Shutterstock.com

This research topic for Frontiers in Psychology highlights some of the more relevant changes that have conditioned consumer behavior in recent years—among these, the paradigm shift in marketing is worth emphasizing.

Today, the market and the companies are implementing Marketing 4.0; This new marketing approach modifies both the business rules and the channels by changing the way to dialogue, interact and relation with consumers.

The present Research Topic brings together 30 studies by 76 authors who analyzed the relevance of consumer behavior changes under this new paradigm, using different theoretical and methodological frameworks. These different papers, mainly constituting original research, examine a variety of sub-topics, including online and mobile environments, value co-creation, internal marketing strategies, and diverse industries and product markets.

Given this broad selection of papers, we encourage readers to draw their own conclusions about the complex phenomena of consumer behavior. Our hope is that these different perspectives will cover various gaps in the field and prompt discussion among the audience of Frontiers in Psychology.

**Citation:** Martínez-Ruiz, M. P., Gómez-Suárez, M., Jiménez-Zarco, A. I., Izquierdo-Yusta, A., eds. (2018). From Consumer Experience to Affective Loyalty: Challenges and Prospects in the Psychology of Consumer Behavior 3.0. Lausanne: Frontiers Media. doi: 10.3389/978-2-88945-412-9

# Table of Contents

*07 Editorial: From Consumer Experience to Affective Loyalty: Challenges and Prospects in the Psychology of Consumer Behavior 3.0* María P. Martínez-Ruiz, Mónica Gómez-Suárez, Ana I. Jiménez-Zarco and Alicia Izquierdo-Yusta

#### **SECTION 1. Online and Mobile Environments**


María Pilar Martínez-Ruiz, Alicia Izquierdo-Yusta, Cristina Olarte-Pascual and Eva Reinares-Lara

*46 Music Audiences 3.0: Concert-Goers' Psychological Motivations at the Dawn of Virtual Reality*

Jean-Philippe Charron

*50 Differences of Perceived Image Generated through the Web Site: Empirical Evidence Obtained in Spanish Destinations*

Juan J. Blazquez-Resino, Ana I. Muro-Rodriguez and Israel R. Perez-Jimenez

*64 Consumer Expectations of Online Services in the Insurance Industry: An Exploratory Study of Drivers and Outcomes*

M. Dolores Méndez-Aparicio, Alicia Izquierdo-Yusta and Ana I. Jiménez-Zarco

*78 A Literature Review of Word of Mouth and Electronic Word of Mouth: Implications for Consumer Behavior*

Nuria Huete-Alcocer

#### **SECTION 2. Value Co-creation**

*82 Consumer-Brand Relationships under the Marketing 3.0 Paradigm: A Literature Review*

Mónica Gómez-Suárez, María Pilar Martínez-Ruiz and Noemí Martínez-Caraballo

*86 Consumer Participation in Co-creation: An Enlightening Model of Causes and Effects Based on Ethical Values and Transcendent Motives*

Ricardo Martínez-Cañas, Pablo Ruiz-Palomino, Jorge Linuesa-Langreo and Juan J. Blázquez-Resino

*103 A Perspective on Consumers 3.0: They Are Not Better Decision-Makers than Previous Generations*

Petr Houdek

*109 I Will Do It If I Enjoy It! The Moderating Effect of Seeking Sensory Pleasure When Exposed to Participatory CSR Campaigns*

Salvador Ruiz de Maya, Rafaela Lardín-Zambudio and Inés López-López

### **SECTION 3. Health**


Carmen Selva-Sevilla, Maria Luisa Gonzalez-Moral and Maria Teresa Tolosa-Perez

### **SECTION 4. Retailing**

*156 Influence of Customer Quality Perception on the Effectiveness of Commercial Stimuli for Electronic Products*

Álvaro Garrido-Morgado, Óscar González-Benito and Mercedes Martos-Parta

*171 Consumer Behavior in Shopping Streets: The Importance of the Salesperson's Professional Personal Attention*

Natalia Medrano, Cristina Olarte-Pascual, Jorge Pelegrín-Borondo and Yolanda Sierra-Murillo

*185 Multi-Vendor Loyalty Programs: Influencing Customer Behavioral Loyalty?* Teresa Villacé-Molinero, Pedro Reinares-Lara and Eva Reinares-Lara

### **SECTION 5. Banking and Financial Markets**

*196 Assessing the Growth of Ethical Banking: Some Evidence from Spanish Customers*

Fernando E. Callejas-Albiñana, Isabel Martínez-Rodríguez, Ana I. Callejas-Albiñana and Irene M. de Vidales-Carrasco


### **SECTION 6. Food**

*239 Attitudes vs. Purchase Behaviors as Experienced Dissonance: The Roles of Knowledge and Consumer Orientations in Organic Market*

María Hidalgo-Baz, Mercedes Martos-Partal and Óscar González-Benito

*247 Key External Influences Affecting Consumers' Decisions Regarding Food* María Pilar Martínez-Ruiz and Carmen María Gómez-Cantó

### **SECTION 7. Non-profit Organizations**

*251 Emotional Effects on University Choice Behavior: The Influence of Experienced Narrators and Their Characteristics*

Ana I. Callejas-Albiñana, Fernando E. Callejas-Albiñana and Isabel Martínez-Rodríguez

*264 Person-Organization Commitment: Bonds of Internal Consumer in the Context of Non-profit Organizations*

Emma Juaneda-Ayensa, Mónica Clavel San Emeterio and Carlos González-Menorca

### **SECTION 8. Transport**

*279 Consumer Behavior in the Choice of Mode of Transport: A Case Study in the Toledo-Madrid Corridor*

Ana I. Muro-Rodríguez, Israel R. Perez-Jiménez and Santiago Gutiérrez-Broncano

### **SECTION 9. Internal Marketing Strategies**

*294 New Strategies in the New Millennium: Servant Leadership As Enhancer of Service Climate and Customer Service Performance*

Jorge Linuesa-Langreo, Pablo Ruiz-Palomino and Dioni Elche-Hortelano

*308 Behavior of Internal Customer in Family Business: Strategies and Actions for Improving Their Satisfaction* Santiago Gutiérrez-Broncano, Pedro Jiménez-Estévez

and María del Carmen Zabala-Baños


Joan Torrent-Sellens, Pilar Ficapal-Cusí and Joan Boada-Grau

# Editorial: From Consumer Experience to Affective Loyalty: Challenges and Prospects in the Psychology of Consumer Behavior 3.0

María P. Martínez-Ruiz <sup>1</sup> , Mónica Gómez-Suárez <sup>2</sup> , Ana I. Jiménez-Zarco<sup>3</sup> \* and Alicia Izquierdo-Yusta<sup>4</sup>

<sup>1</sup> Comercializacion e Investigación de Mercados, University of Castilla-La Mancha, Albacete, Spain, <sup>2</sup> Comercializacion e Investigación de Mercados, Autonoma University of Madrid, Madrid, Spain, <sup>3</sup> Economic and Business Studies, Open University of Catalunya, Barcelona, Spain, <sup>4</sup> Comercializcion e Invetigacion de Mercados, University of Burgos, Burgos, Spain

Keywords: consumer behavior, consumer emotional journey, market, marketing 3.0, marketing 4.0

#### **Editorial on the Research Topic**

#### **From Consumer Experience to Affective Loyalty: Challenges and Prospects in the Psychology of Consumer Behavior 3.0**

#### Edited and reviewed by:

Pavlos A. Vlachos, ALBA Graduate Business School, Greece

> \*Correspondence: Ana I. Jiménez-Zarco ajimenezz@uoc.edu

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 30 October 2017 Accepted: 07 December 2017 Published: 19 December 2017

#### Citation:

Martínez-Ruiz MP, Gómez-Suárez M, Jiménez-Zarco AI and Izquierdo-Yusta A (2017) Editorial: From Consumer Experience to Affective Loyalty: Challenges and Prospects in the Psychology of Consumer Behavior 3.0. Front. Psychol. 8:2224. doi: 10.3389/fpsyg.2017.02224 Around the world, consumer markets have been fundamentally changed by various forces, such as globalization and social media. As a result, organizational managers are increasingly challenged to analyze and understand consumer behavior. Indeed, companies' long-term competitiveness may depend on how well they understand diverse product markets and can adapt to consumers' needs and demands.

This research topic for Frontiers in Psychology highlights some of the more relevant changes that have conditioned consumer behavior in recent years—among these, the paradigm shift in marketing is worth emphasizing.

Indeed, when the call for papers began in September 2015, marketing studies have already recognized the transition from marketing 1.0 to marketing 3.0. For instance, Kotler et al. (2010) suggest how along time marketing orientation within organizations passed through three different stages: marketing 1.0 oriented to the product, marketing 2.0 focused on the client (2.0) and from there to marketing 3.0, "the Values Driven Era," whereby consumers' personal values (ethical, social, spiritual, etc.) were incorporated into marketing decisions. Whilst under traditional marketing unidirectional communication was the key, currently connectivity and technology have altered the way marketing approach to consumers.

By the end of 2016, the Marketing 4.0 perspective has emerged. Its goal is to help organizations reach and engage consumers more fully than in previous years by analyzing shifts in consumers' behaviors (Kotler et al., 2017). Thus, Marketing 4.0. represents the natural evolution of Marketing 3.0., based on the use of technologies as a way to know, dialogue, interact, and establish a relationship with the consumers (Jiménez-Zarco et al., 2017). Moreover, Marketing 4.0 emphasizes the need to consider simultaneously the "new" and the "old" marketing to get consumers recommend the brand. Especially through the numerous possibilities social media and digital marketing offer in this respect, which are revolutionizing the marketing environment and consequently, the way business are being made today.

Today, the market—considering in a wide sense all stakeholders and not only end consumers—is forcing companies to implement Marketing 4.0; but also, companies push the market towards Marketing 4.0, so it is possible to affirm this phenomenon is taking place in a double direction. However, not all agents are adapting at the same time to the use of new technologies.

This new marketing approach modifies both the business rules and the channels by changing the relationship with consumers. Nowadays, clients appreciate constant communication with the organization and expect more than a relationship based on reciprocity and interactivity; they expect the company to "contribute with something else" (Gómez-Suárez et al., 2016), especially in the realm of emotional value (Kotler et al., 2017).

In particular, there is a transformation of the relation between the company and the client, as well as a modification of the playing rules and the relationship channels between both agents (Melero et al., 2016). For this reason, organizations must continue to offer value to the client, according to the digital environment (Vernuccio et al., 2015) and taking into account an inclusive marketing of attraction, oriented to create valuable contents, and offer satisfactory experiences, with the aim at increasing affective links with brands (Kaufmann et al., 2016).

While those facets of the new consumer paradigm are approached, there remain pivotal questions about consumer behavior. To address these gaps, the present Research Topic brings together 30 studies by 76 authors who analyzed the relevance of consumer behavior changes using different theoretical and methodological frameworks. These different papers, mainly constituting original research, examine a variety of sub-topics, including online and mobile environments, value co-creation, internal marketing strategies, and diverse industries and product markets. In particular, these research gaps can be described as follows:


In the last decade, the advance of the Internet and other new technologies has transformed both markets and consumers behaviors. Seven articles focusing on online and mobile environments highlight instances of companies adapting the following new strategies and tools. Omni channel (Juaneda-Ayensa et al.), mobile advertising (Hongyang and Zankui; Martínez-Ruiz et al.), virtual reality (Charron), promotional and online services through web pages (Bláquez-Resino et al.; Méndez-Aparicio et al.), and electronic word of mouth (Huete-Álcocer).

Another four articles discuss the value co-creation—whereby many companies and organizations are giving consumers a more active role in the value-creation process. Through their literature review, Gómez-Suárez et al. examine how consumerbrand relationships have received increasing attention in recent years. Martínez-Cañas et al. highlight how those brands acknowledged as ethical elicit positive emotional responses among their consumers. Houdek shows that even socially and environmentally responsible consumers (the so-called Consumer 3.0) exhibit selective recall, limited attention, and bounded search in their perceptions about the price-quality of purchases. Ruiz de Maya et al. analyze how companies' allocation of resources to corporate social responsibility (CSR) initiatives serve as a form of value co-creation with consumers.

Several articles address issues related to diverse industries and sectors. Four papers emphasize health (Múñoz et al.; Pelegrín-Borondo et al.; Selva-Sevilla et al.; Selva-Selvilla et al). Three studies focus on retailing (Garrido-Morgado et al.; Medrano et al.; Villacé-Molinero et al.). Three concentrate on banking and financial markets (Callejas-Albiñana et al.; Cano et al.; González et al.). There are another two studies about food (Hidalgo-Baz et al.; Martínez-Ruiz and Gómez-Cantó), two about non-profit organizations (Callejas-Albiñana et al.; Juaneda-Ayensa et al.), and one about transport (Muro-Rodríguez et al.).

Finally, internal marketing strategies are another important part of the new paradigm in consumer research. On this topic, three papers explain how companies are adopting new organizational strategies in order to appeal to customers in the current millennium. Linuesa-Langreo et al. explain the impact of servant leadership, which involves putting employees' needs first and serving the broader society. Gutiérrez-Boncano et al. determine the relevant aspects of family businesses that make them increasingly competitive. González and Fernández focus on the inclusion of people with disabilities, while Torrent-Sellens et al. focus on dispositional employability.

Given this broad selection of papers, we encourage readers to draw their own conclusions about the complex phenomena of consumer behavior. Our hope is that these different perspectives will cover various gaps in the field and prompt discussion among the audience of Frontiers in Psychology.

## AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

## FUNDING

This research was conducted under the framework of different research projects and groups: the Spanish Ministry of Economy and Competitiveness, Research Project reference MINECO Ref.: ECO2014-59688-R (Spain); TECHNOCOM UAM research group (E-38) (Universidad Autónoma de Madrid); and the I2TIC Research Group (Open University of Catalonia).

### REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Martínez-Ruiz, Gómez-Suárez, Jiménez-Zarco and Izquierdo-Yusta. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Omnichannel Customer Behavior: Key Drivers of Technology Acceptance and Use and Their Effects on Purchase Intention

#### Emma Juaneda-Ayensa, Ana Mosquera\* and Yolanda Sierra Murillo

Departamento de Economía y Empresa, Universidad de La Rioja, Logroño, Spain

The advance of the Internet and new technologies over the last decade has transformed the retailing panorama. More and more channels are emerging, causing consumers to change their habits and shopping behavior. An omnichannel strategy is a form of retailing that, by enabling real interaction, allows customers to shop across channels anywhere and at any time, thereby providing them with a unique, complete, and seamless shopping experience that breaks down the barriers between channels. This paper aims to identify the factors that influence omnichannel consumers' behavior through their acceptance of and intention to use new technologies during the shopping process. To this end, an original model was developed to explain omnichannel shopping behavior based on the variables used in the UTAUT2 model and two additional factors: personal innovativeness and perceived security. The model was tested with a sample of 628 Spanish customers of the store Zara who had used at least two channels during their most recent shopping journey. The results indicate that the key determinants of purchase intention in an omnichannel context are, in order of importance: personal innovativeness, effort expectancy, and performance expectancy. The theoretical and managerial implications are discussed.

Keywords: omnichannel experience, shopping motives, consumer behavior, omnishopper, fashion retail, technology acceptance model, UTAUT2

## INTRODUCTION

In recent years, advances in technology have enabled further digitalization in retailing, while also posing certain challenges. More specifically, the evolution of interactive media has made selling to consumers truly complex (Crittenden et al., 2010; Medrano et al., 2016). With the advent of the mobile channel, tablets, social media, and the integration of these new channels and devices in online and offline retailing, the landscape has continued to evolve, leading to profound changes in customer behavior (Verhoef et al., 2015).

A growing number of customers use multiple channels during their shopping journey. These kinds of shoppers are known as omnishoppers, and they expect a seamless experience across channels (Yurova et al., in press). For example, an omnishopper might research the characteristics of a product using a mobile app, compare prices on several websites from their laptop, and, finally, buy the product at a physical store. This consumer 3.0 uses new technology to search for information, offer opinions, explain experiences, make purchases, and talk to the brand.

#### Edited by:

Maria Pilar Martinez-Ruiz, University of Castilla–La Mancha, Spain

#### Reviewed by:

Maria Avello, Complutense University of Madrid, Spain Maria Dolores Reina, National University of Distance Education, Spain

#### \*Correspondence:

Ana Mosquera ana-maria.mosquera@unirioja.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

> Received: 28 April 2016 Accepted: 12 July 2016 Published: 28 July 2016

#### Citation:

Juaneda-Ayensa E, Mosquera A and Sierra Murillo Y (2016) Omnichannel Customer Behavior: Key Drivers of Technology Acceptance and Use and Their Effects on Purchase Intention. Front. Psychol. 7:1117. doi: 10.3389/fpsyg.2016.01117 In an omnichannel environment, channels are used seamlessly and interchangeably during the search and purchase process, and it is difficult if not virtually impossible for retailers to control this use (Neslin et al., 2014; Verhoef et al., 2015).

Lu et al. (2005) consider mobile commerce to be the second wave of e-commerce. We believe that omnichannel commerce could be the third wave. Most studies on end-user beliefs and attitudes are conducted long after the systems have been adopted; while initial adoption is the first step in long-term usage, the factors affecting usage may not be the same as those influencing the initial adoption, or the degree of their effect may vary (Lu et al., 2005). Few papers have addressed the issue of preadoption criteria for omnishoppers, and explanations of why users behave in a particular way toward information technologies have predominantly focused on instrumental beliefs, such as perceived usefulness and perceived ease of use, as the drivers of usage intention. Previous papers in behavioral science and psychology suggest that holistic experiences (Schmitt, 1999) with technology, as captured in constructs such as enjoyment, flow, and social image, are potentially important explanatory variables in technology acceptance.

This paper aims to advance the theoretical understanding of the antecedents of omnishoppers' technology acceptance and use in relation to early adoption of omnichannel stores. To this end, it focuses on the acceptance and use of the technology that customers use in the "information prior to purchase" and "purchase" stages. We carried out this research in the fashion word, because it is one of the earliest industries to adopt this new strategy (PwC et al., 2016). This paper presents a new model of technology acceptance and use based on UTAUT2 (Venkatesh et al., 2012), extended to include two new dimensions—personal innovativeness and perceived security—and adapted to a specific context, i.e., the omnichannel environment.

Our research has important theoretical and managerial implications since studying the drivers of omnishoppers' shopping behavior would allow firms to follow different strategies in omnichannel customer management aimed at increasing customer satisfaction by offering an integrated shopping experience (Lazaris and Vrechopoulos, 2014; Neslin et al., 2014; Lazaris et al., 2015; Verhoef et al., 2015).

To achieve this goal, this paper proceeds as follows: first, we review the literature on the topic of omnichannel consumer behavior and the drivers of omnichannel shopping. Second, we develop a new theoretical model. Third, we describe and explain the empirical study. Fourth, we examine the results and implications of the findings and derive our conclusions. Fifth and finally, we address the limitations of the research and offer further research proposals.

### LITERATURE REVIEW

## Omnichannel Retailing Context

Recent years have witnessed the emergence of new retailing channels. Thanks to new technologies, retailers can integrate all the information these channels provide, a phenomenon known as omnichannel retailing (Brynjolfsson et al., 2013).

The omnichannel concept is perceived as an evolution of multichannel retailing (**Table 1**). While multichannel retailing implies a division between the physical and online store, in the omnichannel environment, customers move freely among channels (online, mobile devices, and physical store), all within a single transaction process (Melero et al., 2016). Omnis is a Latin word meaning "all" or "universal," so omnichannel means "all channels together" (Lazaris and Vrechopoulos, 2014). Because the channels are managed together, the perceived interaction is not with the channel, but rather the brand (Piotrowicz and Cuthbertson, 2014).


Source: Based on Rigby (2011), Piotrowicz and Cuthbertson (2014), Beck and Rygl (2015), and Verhoef et al. (2015).

The dominant characteristic of the omnichannel retailing phenomenon is that the strategy is centered on the customer and the customer's shopping experience, with a view to offering the shopper a holistic experience (Gupta et al., 2004; Shah et al., 2006).

In addition, the omnichannel environment places increasing emphasis on the interplay between channels and brands (Verhoef et al., 2015). Neslin et al. (2014) describe multiple purchase routes to show how this interplay works. Thus, not only is the omnichannel world broadening the scope of channels, it also integrates consideration of customer-brand-retail channel interactions.

Another important change is that the different channels are blurring together as the natural boundaries that once separated them begin to disappear. They are thus used seamlessly and interchangeably during the search, purchase, and post-purchase process, and it is difficult or virtually impossible for firms to control this usage (Verhoef et al., 2015).

### Consumer Attitudes toward Technology in an Omnichannel Context

Due to the increasing use of new technologies in retailing, consumer shopping habits and expectations are also changing. A new multi-device, multiscreen consumer has emerged who is better informed and demands omnichannel brands. Research has shown that omnichannel consumers are a growing global phenomenon (Schlager and Maas, 2013).

Customers expect a consistent, uniform, and integrated service or experience, regardless of the channel they use; they are willing to move seamlessly between channels—traditional store, online, and mobile—depending on their preferences, their current situation, the time of day, or the product category (Cook, 2014; Piotrowicz and Cuthbertson, 2014). The omnishopper no longer accesses the channel, but rather is always in it or in several at once, thanks to the possibilities offered by technology and mobility. These new shoppers want to use their own device to perform searches, compare products, ask for advice, or look for cheaper alternatives during their shopping journey in order to take advantage of the benefits offered by each channel (Yurova et al., in press). In addition, omnichannel consumers usually believe that they know more about a purchase than the salespeople and perceive themselves as having more control over the sales encounter (Rippé et al., 2015).

Despite the increase recorded in research on information and communication technology (ICT) and multichannel, it is important to continue investigating in the field of omnichannel consumer behavior (Neslin et al., 2014; Verhoef et al., 2015) and, especially, to determine how consumers' attitudes toward technology influence the purchasing decision process in the new context (Escobar-Rodríguez and Carvajal-Trujillo, 2014).

### Theory of Acceptance and Use of Technology in an Omnichannel Context: Model and Hypothesis

Our research framework is based on an additional extension of the extended Unified Theory of Acceptance and Use of Technology (UTAUT2) model (Venkatesh et al., 2012) that seeks to identify the drivers of technology acceptance and use during the shopping journey to purchase in an omnichannel environment. Following the literature review, we chose the UTAUT2 model because it provides an explanation for ICT acceptance and use by consumers (Venkatesh et al., 2012). UTAUT2 is an extension of the original UTAUT model that synthesizes eight distinct theoretical models taken from sociological and psychological theories used in the literature on behavior (**Table 2**; Venkatesh et al., 2003). This theory contributes to the understanding of important phenomena such as, in this case, omnichannel consumers' attitudes toward technology and how they influence purchase intention in the shopping-process context. Under UTAUT2, a consumer's intention to accept and use ICT is affected by seven factors: performance expectancy, effort expectancy, social influence, facilitating conditions, hedonic motivations, price value, and habit.

As proposed by Venkatesh et al. (2012), UTAUT2 needs to be applied to different technologies and contexts, and other factors need be included, to verify its applicability, especially in the context of consumer behavior. To this end, building on previous work, in this study, we included personal innovativeness (San Martín and Herrero, 2012; Escobar-Rodríguez and Carvajal-Trujillo, 2014) and perceived security (Kim et al., 2008; Escobar-Rodríguez and Carvajal-Trujillo, 2014) to shed light on the degree to which the different factors included in the model influence consumers' purchase intentions.

#### The UTAUT2 Model Adapted to an Omnichannel Environment

As noted, our model was based on the UTAUT2 model.


Source: Based on Escobar-Rodríguez and Carvajal-Trujillo, (2014). SI, Social Influence; PE, Performance Expectancy; EE, Effort Expectancy; FC, Facilitating conditions.

Performance expectancy is defined as the degree to which using different channels and/or technologies during the shopping journey will provide consumers with benefits when they are buying fashion (Venkatesh et al., 2003, 2012). Performance expectancy has consistently been shown to be the strongest predictor of behavioral intention (e.g., Venkatesh et al., 2003, 2012; Escobar-Rodríguez and Carvajal-Trujillo, 2014) and purchase intention (Pascual-Miguel et al., 2015). In keeping with the literature, we proposed the following hypothesis:

H1. Performance expectancy positively affects omnichannel purchase intention.

Effort expectancy is the degree of ease associated with consumers' use of different touchpoints during the shopping process. Existing technology acceptance models include the concept of effort expectancy as perceived ease of use (TAM/TAM2) or ease of use (Innovation Diffusion Theory). According to previous studies (Karahanna and Straub, 1999), the effort expectancy construct is significant in both voluntary and mandatory usage contexts (Venkatesh et al., 2003) and positively affects purchase intention (Venkatesh et al., 2012). The following hypothesis was thus proposed for this construct:

H2. Effort expectancy positively affects omnichannel purchase intention.

Social influence is the extent to which consumers perceive that people who are important to them (family, friends, role models, etc.) believe they should use different channels depending on their needs. Social influence, understood as a direct determinant of behavioral intentions, is included as subjective norm in TRA, TAM2, and TPB, and as image in IDT (Fishbein and Ajzen, 1975; Schifter and Ajzen, 1985; Davis, 1989; Davis et al., 1989; Moore and Benbasat, 1991). The social influence, subjective norm, and social norm constructs all contain the explicit or implicit notion that individual behavior is influenced by how people believe others will view them as a result of having used the technology (Venkatesh et al., 2003) and positively affect purchase intention (Venkatesh et al., 2012). Therefore, the following hypothesis was proposed:

H3. Social influence positively affects omnichannel purchase intention.

Habit is defined as the extent to which people tend to perform behaviors automatically because of learning (Venkatesh et al., 2012). This concept, which was included as a new construct in the UTAUT2 model, has been considered a predictor of technology use in many studies (e.g., Kim et al., 2005; Kim and Malhotra, 2005; Limayem et al., 2007) and directly influences purchase intention (Venkatesh et al., 2012; Escobar-Rodríguez and Carvajal-Trujillo, 2014). Based on the literature, the following hypothesis was thus proposed:

H4. Habit positively affects omnichannel purchase intention.

In order to analyze consumers' motivations for adopting omnichannel behavior, we based our framework on the extended literature used in retailing. Previous research on shopping behavior suggests that customers use different channels at each stage of the shopping process to meet utilitarian and hedonic needs at the lowest cost relative to benefits, in other words, to maximize value (e.g., Balasubramanian et al., 2005; Noble et al., 2005; Konu¸s et al., 2008).

Shopping value can be both hedonic and utilitarian (Babin et al., 1994). Hedonic motivations are associated with adjectives such as fun, pleasurable, and enjoyable (e.g., Holbrook and Hirschman, 1982; Kim and Forsythe, 2007; To et al., 2007; Venkatesh et al., 2012). In contrast, utilitarian motivations are rational and task-oriented (Batra and Ahtola, 1991). Both dimensions are important because they are present in all shopping experiences and consumer behavior (Jones et al., 2006). Items such as clothing are classified in the highly hedonic product category due to their symbolic, experimental, and pleasing properties (Crowley et al., 1992). Consumers are more likely to select a physical store when they shop for hedonic fashion goods because strong physical environments elevate mood by providing opportunities for social interaction, product evaluation, and sensory stimulation (Nicholson et al., 2002). However, recent data show that consumers consider online fashion shopping to be a pleasurable activity and spend their leisure time searching for clothes using this medium (Blázquez, 2014).

In relation to technology acceptance and use, while utilitarian motivation was included as part of the performance expectancy construct in keeping with Venkatesh et al. (2003), hedonic motivation was included as a separate construct in UTAUT2 (Venkatesh et al., 2012). Hedonic motivation is defined as the fun or pleasure derived from using a technology, and it has been shown to play an important role in determining technology acceptance and use (Brown and Venkatesh, 2005). Numerous papers on ICT have demonstrated the influence of hedonic motivation on the intention both to use a technology and to purchase it (Van Der Heijden, 2004; Thong et al., 2006). Therefore, the following hypothesis was proposed:

H5. Hedonic motivations positively affect omnichannel purchase intention.

#### External Variables Applied in the Extension of UTAUT2

When shoppers come into contact with a new technology or innovation, they have the opportunity to adopt or refuse it. Prior research has shown that innovative multichannel customers prefer to explore and use new alternatives (e.g., Steenkamp and Baumgartner, 1992; Rogers, 1995; Konu¸s et al., 2008). In addition, several studies in the e-commerce literature have demonstrated the important role that innovativeness plays in purchase intention in different contexts (e.g., Herrero and Rodriguez del Bosque, 2008; Lu et al., 2011; San Martín and Herrero, 2012; Escobar-Rodríguez and Carvajal-Trujillo, 2014).

Personal innovativeness is defined as the degree to which a person prefers to try new and different products or channels and to seek out new experiences requiring a more extensive search (Midgley and Dowling, 1978). Many papers have highlighted that consumer innovativeness is a highly influential factor in ICT adoption and on purchase intention (e.g., Agarwal and Prasad, 1998; Citrin et al., 2000; Herrero and Rodriguez del Bosque,

2008; San Martín and Herrero, 2012; Escobar-Rodríguez and Carvajal-Trujillo, 2014). The following research hypothesis was thus formulated:

H6. Personal innovativeness positively affects omnichannel purchase intention.

Additionally, we included the perceived security of the online channels, referring to the belief that the Internet is a secure option for sending personal data (Escobar-Rodríguez and Carvajal-Trujillo, 2014; Bonsón Ponte et al., 2015). Perceived security can be defined as the perception by consumers that the omnichannel companies' technology strategies include the antecedents of information security, such as authentication, protection, verification, or encryption (Kim et al., 2008). If consumers perceive that the online channels have security attributes, they will deduce that the retailer's intention is to guarantee security during the purchasing process (Chellappa and Pavlou, 2002). There is some evidence that the perceived security of online channels positively affects the intent to purchase using these kind of channels (e.g., Salisbury et al., 2001; Frasquet et al., 2015). In light of these findings, it was hypothesized that perceived security is related to purchase intention as follows:

H7. Perceived security positively affects the omnichannel purchase intention.

**Figure 1** shows the theoretical model based on the seven hypotheses, reflecting how the antecedents of technology acceptance and use affect purchase intention in an omnichannel environment.

### METHODOLOGY

We designed an online survey focused on omnichannel fashion retail customers. The questionnaire was administered to a Spanish Internet panel. For the purposes of our study, we defined omnichannel shoppers as those shoppers who use at least two channels of the same retailer during their shopping journey. The panelists were screened to select those members that fit our definition of omnichannel shoppers. In all, 628 respondents indicated their behavior with regard to their most recent purchase in the 12 months prior to the collection of the data (January 2016).

To carry out the study we selected the company Zara for several reasons. First, Zara is one of the most well-known and important fashion retailers. Additionally, the brand follows an omnichannel strategy, allowing its customers to combine different online channels (the company website, social media, and the mobile app) with the offline channels throughout their customer journey. In other words, shoppers can search for information on a product using the Zara mobile app, buy the product on the Zara website (www.zara.com), and then pick up or return the product at the physical store. However, the most important reason for choosing a single company to study the factors influencing omnichannel customers' behavior was to isolate the omnichannel factor, that is, we wanted to determine the drivers for using different channels and/or technologies of a single company during a single shopping process.

The questionnaire consisted of two parts. The first part contained statements about shopping motives. Based on their

#### TABLE 3 | Theory of use and acceptance of technology in an omnichannel context.


most recent shopping process (**Table 3**), respondents were instructed to rate their agreement with each item on a sevenpoint Likert scale ranging from 1 (strongly agree) to 7 (strongly disagree).

The second part of the questionnaire was used to gather sociodemographic information, such as gender, age, employment status, and education (**Table 4**). The sample was highly representative of the distribution of online shoppers according to recent surveys (Corpora 360 and iab Spain, 2015).

Because of the novelty of the field of application, the measurement scales were then translated into Spanish using a back-translation method, whereby one person translated the items into Spanish and two others translated them back into English, making it possible to check for any misunderstandings or misspellings resulting from the translation (Brislin, 1970). In addition, we conducted a pretest with 25 participants to ensure the comprehensibility of the questions.

We used IBM SPSS Statistics 19 to perform the exploratory factor analysis. Subsequently we undertook a regression analysis of latent variables based on the partial least squares (PLS) technique.

The aim of this research was to explore technology acceptance and use in an omnichannel context. To achieve this aim, fundamentally, theory development, we chose to use the PLS technique to evaluate the structural model before testing the causal model. Next, we estimated a confirmatory factor model to study the validity of the scale and examined the underlying structure. To this end, we created a causal model and used




structural equations to evaluate the scale and the effect of technology acceptance and use on omnichannel shoppers' purchase intentions.

The study was approved by the Head of Ethics at the Faculty of Business Administration of La Rioja University. All participants provided informed consent.

#### DATA ANALYSIS AND RESULTS

#### Measurement Model

We performed a confirmatory factor analysis to which we made a few amendments. It was likewise verified that the loadings of all the standardized parameters were greater that 0.7 (Hair et al., 2013). The item innovativeness3 had a value lower that 0.7 and a t-value lower that 1.96. We thus decided to exclude it to improve the model's convergence, as recommended by Anderson and Gerbing (1988). The model confirms that the indicators converge with the assigned factors.

The model was verified in terms of construct reliability (i.e., composite reliability and Cronbach's alpha), convergent validity, and discriminant validity. The composite reliability and Cronbach's alpha values were >0.70, and the constructs' convergent validity was also confirmed, with an average variance explained (AVE) >0.50 in all cases. The discriminant validity of the constructs was measured by comparing the square root of the AVE of each construct with the correlations between constructs (Roldán and Sánchez-Franco, 2012). The square root of the AVE (diagonal elements in italics in **Table 5**) had to be larger than the corresponding inter-construct correlation (off-diagonal elements in **Table 5**). This criterion was also met in all cases. Furthermore, each item's loading on its corresponding factor was greater than the cross-loadings on the other factors.

#### Structural Model

Bootstrapping with 5000 resamples was used to assess the significance of the path coefficients obtained by PLS-SEM (Hair et al., 2011). The model explains the intention to purchase in the omnichannel context well, with an R <sup>2</sup> of 47.9% (**Table 6**). Stone-Geisser's cross-validated redundancy Q <sup>2</sup> was >0, specifically, 0.406. This result confirmed the predictive power of the proposed model (see Hair et al., 2011).

The sign, magnitude, and significance of the path coefficients are shown in **Table 6**. Three hypotheses were supported by the results: H1 (regarding the influence of performance expectancy), H2 (regarding the influence of effort expectancy), and H6 (regarding the influence of personal innovativeness). In contrast, H3 (regarding social influence), H4 (regarding the influence of habit), H5 (regarding the influence of hedonic motivation), and H7 (regarding the influence of perceived security) were rejected, as the relationships were not significant.

### CONCLUSION AND MANAGERIAL IMPLICATIONS

Today's increasingly competitive retail world has given rise to a new phenomenon known as omnichannel retailing (e.g., Rigby et al., 2012; Neslin et al., 2014; Beck and Rygl, 2015; Verhoef et al., 2015). This phenomenon can be defined as the customer management strategy throughout the life cycle of the customer relationship whereby the shopper interacts with the brand through different devices and channels (mainly the physical store, the online channel, the mobile channel, and social media), and, thus, all touchpoints must be integrated to provide a seamless and complete shopping experience, regardless of the channel used. Omnichannel retailing stands to become the third wave of e-commerce.

Few studies have analyzed the antecedents of omnishopper behavior (e.g., Lazaris et al., 2014; Neslin et al., 2014; Verhoef et al., 2015). The main goal of the present research was to identify the drivers of technology acceptance and use among omnichannel consumers and to analyze how they affect purchase intention in an omnichannel context. To this end, we proposed a new model based on the extended Unified Theory of Acceptance and Use of Technology (UTAUT2) model (Venkatesh et al., 2012), which we further extended to include two new factors: personal innovativeness and perceived security. Both personal



EE, Effort Expectancy; H, Habit; HM, Hedonic Motivation; PI, Personal Innovativeness; PE, Performance Expectancy; PS, Perceived Security; SI, Social Influence; PUR\_IN, Purchase Intention.

#### TABLE 6 | Results of the structural model.


PE, Performance Expectancy; EE, Effort Expectancy; SI, Social Influence; H, Habit; HM, Hedonic Motivation; PI, Personal Innovativeness; PS, Perceived Security; PUR\_IN, Purchase Intention.

innovativeness and perceived security have been found to be important for the adoption of new technologies in the literature on consumer behavior (e.g., Salisbury et al., 2001; Herrero and Rodriguez del Bosque, 2008; Escobar-Rodríguez and Carvajal-Trujillo, 2014; Frasquet et al., 2015). The present paper helps to advance the theoretical understanding of the antecedents of consumer 3.0 technology acceptance and use in the early adoption of omnichannel stores.

The model was found to predict omnichannel purchase intention (R <sup>2</sup> = 47.9%). Our findings show that a consumer's intention to purchase in an omnichannel store is influenced by personal innovativeness, effort expectancy, and performance expectancy. In contrast, contrary to our hypotheses based on the broader previous literature, habit, hedonic motivation, social influence, and perceived security do not affect omnichannel purchase intention.

Personal innovativeness is the strongest predictor of purchase intention in the omnichannel context. This factor plays an important role as a direct driver of omnichannel purchase intention. This finding is consistent with those of previous papers (e.g., Herrero and Rodriguez del Bosque, 2008; Lu et al., 2011; San Martín and Herrero, 2012; Escobar-Rodríguez and Carvajal-Trujillo, 2014). Thus, individuals who are more innovative with regard to ICT will have a stronger intention to purchase using different channels and devices in an omnichannel environment. Our findings show that omnishoppers seek out new technology in order to experiment with it and be the first to try it among their family and friends. Managers should thus take this technological profile into account and constantly roll out new technologies in different ways in order to attract and surprise these kinds of shoppers.

Our findings also show that effort expectancy and performance expectancy are significant factors in explaining attitude and purchase intention, with a positive effect on behavioral intention, as has been widely reported in the literature (e.g., Childers et al., 2001; Verhoef et al., 2007; Rose et al., 2012). Effort expectancy is the second strongest predictor and has a direct positive influence on purchase intention (e.g., Karahanna and Straub, 1999; Venkatesh et al., 2003, 2012). This could be because omnishoppers are more used to using multiple channels and are more task-oriented, using different channels or technologies to look for better prices or maximize convenience at any given time. In keeping with previous research (e.g., Venkatesh et al., 2003, 2012; Escobar-Rodríguez and Carvajal-Trujillo, 2014), performance expectancy was found to be the third strongest predictor of behavioral intention in an omnichannel environment.

Although the literature has recognized the influence of normative factors such as social influence on people's attitude, intentions, and behavior (e.g., Fishbein and Ajzen, 1975; Bagozzi, 2000; Venkatesh et al., 2012), our results show that this factor does not influence the intention to purchase in an omnichannel environment. On the contrary, in line with previous work (e.g., Casaló et al., 2010; San Martín and Herrero, 2012), social influence was found not to affect purchase intention. This finding contrasts with those reported elsewhere (Kim et al., 2009; Venkatesh et al., 2012; Escobar-Rodríguez and Carvajal-Trujillo, 2014; Pelegrín-Borondo et al., 2016). This may be because technology use is not conditioned by other people's opinions; it could also be due to the specific sector under study. In either case, it is a topic that should be studied further.

Contrary to previous studies (e.g., Venkatesh et al., 2012; Escobar-Rodríguez and Carvajal-Trujillo, 2014), our results indicate that habit does not influence omnichannel purchase intention. This could be because customers are not used to using different channels due to the relatively low number of companies that allow customers to use multiple channels simultaneously. In keeping with authors such as Valentini et al. (2011) and Melero et al. (2016), we believe this variable will increase in importance in the coming years, as more and more retailers implement true omnichannel strategies.

In our research, the hypothesized influence of hedonic motivation on purchase intention was found to be low. Previous work in other contexts has found a positive relationship between these variables (e.g.,Van Der Heijden, 2004; Thong et al., 2006; Venkatesh et al., 2012; Escobar-Rodríguez and Carvajal-Trujillo, 2014). These findings are probably because, when omnishoppers use different channels and touchpoints, they expect a seamless, holistic experience throughout their shopping journey. In other words, hedonic and utilitarian motivation are part of the same construct (Melero et al., 2016). In addition, technology acceptance and use is more of a new experience related to the innovativeness profile than a hedonic one, i.e., excitement over discovering how something will work rather than expected enjoyment based on prior experience.

Finally, contrary to previous findings (e.g., Salisbury et al., 2001; Frasquet et al., 2015), perceived security did not influence omnichannel purchase intention. We interpreted these results to mean that the possibility of buying in an omnichannel context offsets the influence of the need for security, an important factor in e-commerce, by offering the option of traditional instore payment, which nullifies the effect of perceived risk in ecommerce. In this sense, omnichannel stores offer an opportunity to attract more conservative consumers who perceive an increased risk in e-commerce to a more interactive scenario in which retailers can use new technologies to manage customer relationships based on direct contact in the physical store.

Our study contributes to the current literature on omnichannel consumer behavior by adapting the previous UTAUT models to include two new factors in order to determine how the technologies used during the shopping process affect

#### REFERENCES


the intention to purchase in an omnichannel context. The results have practical implications for omnichannel retailer managers regarding the best management and marketing strategies for improving a key part of their business, namely, the creation of a holistic shopping experience for their customers (Lemon and Verhoef, in press). Specifically, retailers need to properly define not only which technologies they will invest in, but also how they will encourage the acceptance thereof, as this acceptance is an important predictor of purchase intention. In particular, in-store technology has to be focused on creating a new integrated customer experience, using technology that is practical, enjoyable, and interesting in order to ensure that innovative customers perceive that the new omnichannel stores facilitate and expedite their shopping journey.

Our paper has some limitations. Our data are related to consumer behavior in a particular case: the buying process in the fashion retailer Zara. It would be interesting to replicate this study in another product category or country to compare the results.

Our research also suggests interesting lines of future research, such as identifying omnichannel consumer profiles in order to personalize the customer shopping experience. Likewise, future studies could investigate the new role of technology in the physical store in an omnichannel environment. In addition, the influence of sociodemographic variables, such as age or gender, as moderator variables to complement the current model should be explored. In keeping with Chiu et al. (2012), we think it would also be interesting to examine habit as a moderator variable in purchase intention.

Finally, fashion companies need to determine which factors matter most to consumers 3.0 when they set out on their shopping journey in order to adapt their strategies to shoppers' motivations. This study has sought to shed light on the new omnichannel phenomenon. Technology is changing the future of retailing. The key will lie in successfully integrating all channels in order to think about them as consumers do and try to offer shoppers an integrated and comprehensive shopping experience.

#### AUTHOR CONTRIBUTIONS

The three authors have equally participated in literature review, data analysis and writing of the paper.

#### ACKNOWLEDGMENTS

The authors would like to thank the referees for their constructive comments on previous versions of this article. This work was funded by the Chair in Commerce at the University of La Rioja (Spain).


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Juaneda-Ayensa, Mosquera and Sierra Murillo. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Effects of Mobile Text Advertising on Consumer Purchase Intention: A Moderated Mediation Analysis

Lin Hongyan\* and Chen Zhankui

School of Economics and Management, Xiamen University of Technology, Xiamen, China

Mobile shopping is increasing in prevalence and has become a necessary part of many people's daily lives. However, one main channel for mobile shopping, mobile shopping applications (apps), has not been thoroughly investigated. This study focused on mobile text advertising delivered from mobile shopping apps using the intention to purchase as the dependent variable for testing its marketing effect. In the context of a promotion focus vs. a prevention focus, we used Higgins' regulatory focus theory combined with Ajzen's TPB and Herzog's U&G to analyze the mechanism by which consumers formulate an intention to purchase in a mobile advertising context. This empirical study surveyed 320 consumers who had made a purchase using a mobile shopping app in the previous month. The results showed that infotainment, irritation, and subjective norms were significantly associated with attitudes; in turn, attitudes mediated the impact of these three factors on the intention to purchase. Moreover, a high promotion focus not only strengthened the positive effect of infotainment on attitudes but also intensified the mediation effect of attitudes between infotainment and the intention to purchase. A high prevention focus also consolidated the negative effect of irritation on attitudes as well as reinforced the mediation effect of attitudes between irritation and the intention to purchase. Furthermore, attitudes, subjective norms, and perceived behavioral control collectively impacted the intention to purchase. These findings shed light on ways to customize goods information in mobile advertising and have strong theoretical and practical implications.

#### Edited by:

Monica Gomez-Suárez, Universidad Autonoma de Madrid, Spain

#### Reviewed by:

Carmen Pacheco-Bernal, Open University of Catalonia, Spain Gonzalo Moreno, Saint Louis University Madrid Campus, Spain

> \*Correspondence: Lin Hongyan linhongyan@xmut.edu.cn

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 23 February 2017 Accepted: 02 June 2017 Published: 23 June 2017

#### Citation:

Hongyan L and Zhankui C (2017) Effects of Mobile Text Advertising on Consumer Purchase Intention: A Moderated Mediation Analysis. Front. Psychol. 8:1022. doi: 10.3389/fpsyg.2017.01022 Keywords: mobile text advertising, infotainment, irritation, regulatory focus, attitudes, intention to purchase

## INTRODUCTION

With the rapid development of the mobile internet, consumers use smartphones not only for communication but also for other activities, such as surfing the web and participating in mobile business. Mobile shopping has become a natural extension of traditional online shopping, thereby achieving multi-scene synchronous shopping (Ngai and Gunasekaran, 2007). According to the latest statistics from the China Internet Network Information Center(CNNIC, 2017), Chinese cell phone cybercitizens now number 695 million, which is 95.1 percent of the total number of people in China who use the internet. Moreover, 441 million users are mobile shoppers. Online shopping platforms, such as Amazon, Taobao, and JD, have exerted considerable effort toward developing mobile business, including developing mobile shopping applications (apps) (Yang et al., 2015). In general, downloading these apps can only be done with the permission of users. Thus, the mobile apps in a user's smartphone have generally been downloaded voluntarily. Such apps are widely used (Lee and Santanam, 2014). For example, Taobao is an open market similar to Amazon. It was developed by the largest Chinese retail trading platform, the Alibaba Group. The Taobao app was launched in 2011. Chinese mobile apps have gradually become commonplace in the past few years. Furthermore, a number of apps facilitate the delivery of advertising content (Grewal et al., 2016). A statistical analysis showed that the United States had more than \$19 billion in investments in the mobile advertising market in 2015, a figure expected to exceed \$65 billion by 2019 (eMarketer, 2015). The growth of the world market is similar to that of the United States. eMarketer (2015) predicted that \$100 billion will be invested in mobile advertising in the world market by 2016. Yet, to our knowledge, there are no previous studies on the effectiveness of mobile text advertising from apps.

Smartphones and other mobile devices have achieved a high degree of personalization and have become vital communication tools; most users constantly keep them nearby, even when they sleep (Bacile et al., 2014). Such devices grant consumers ubiquitous access to digital information at any time and any place; thus, these devices also allow marketers a great opportunity to reach consumers directly and constantly and to analyze their needs (Grewal et al., 2016). Hence, these devices make personalized advertising services possible. But, the consumer's attention is limited. Companies that sent a large number of advertisements to users found that this was not a good marketing method (Stewart and Paul, 2002). The issue of how to optimize the sending of suitable mobile advertising to targeted users is thought-provoking.

Most research on mobile advertising used either the theory of reasoned action or the theory of acceptance model to examine the influence of a technique's characteristics, (e.g., its perceived usefulness and perceived ease of use), information characteristics, and perceived value on consumers' attitudes, adoption, or intention to use (Tsang et al., 2004; Kim et al., 2007; Liu et al., 2012; Muk and Chung, 2015). However, a few empirical studies have developed their mobile advertising research from the perspective of other individual characteristics, such as regulatory focus. Recent sociological studies indicate that the regulatory focus level of individuals considerably affects their behavior, a finding which has been partially validated for environmental advertising (Bhatnagar and McKay-Nesbitt, 2016). But, studies about the impact of commercial mobile advertising on the intention to purchase have seldom been conducted (Bart et al., 2014).

This research attempted to tackle this deficiency by applying regulatory focus theory (RFT), theory of planned behavior (TPB), and uses and gratifications theory (U&G) to investigate the purchase intention of consumers after they had received commercial mobile advertising. The subjects were people who had recently purchased products using a mobile shopping app. The research model is shown in **Figure 1**. Specifically, (1) in the antecedents of attitudes section, as illustrated on the left side of the diagram, this study investigated whether infotainment and irritation would be the key factors affecting consumer attitudes toward mobile text advertising. (2) The outcomes section portrayed on the right side investigated whether the TPB applies to a mobile advertising environment, that is to say, whether consumer attitudes toward mobile text advertising, subjective norms, and perceived behavioral control influence a consumer's intention to purchase mobile advertised goods. (3) At the same time, based on the RFT, as also indicated on the left side of the model diagram in **Figure 1**, this study investigated whether a high promotion focus would not only strengthen the positive effect of infotainment on attitudes toward mobile text advertising but would also intensify the mediation effect of attitudes between infotainment and the intention to purchase. Would a high prevention focus also consolidate the negative effect of irritation on attitudes toward mobile text advertising as well as reinforcing the mediation effect of attitudes between irritation and intention to purchase? In addition, in the empirical section, the research model was examined by structural equation modeling (Anderson and Gerbing, 1988) and a bootstrapping method (Edwards and Lambert, 2007).

Different types of mobile advertising, such as video, multimedia, text advertisements, mobile display advertising, and short message service (SMS), are available. In general, the communication methods used in mobile advertising can be categorized into push and pull types (Park et al., 2008). (1) Pushbased advertising is frequently delivered as an SMS. If a user is interested in the advertisement, s/he can click a link in the SMS to browse for additional information. (2) Pull-based advertising, which can be browser-based or within-app advertising, guides users to a mobile website and then provides interactive choices to understand their habits and preferences (Molitor et al., 2012; Zubcsek et al., 2016). The present study focused on push text advertising from a mobile shopping app for two reasons. First, such advertising is convenient for consumers to read and save, but, browsing a website and selecting goods generally take considerable time. However, push advertising is sent by companies and can be saved. Users can read these advertisements at any time. Second, permission-based advertising information can reduce consumer irritation (McCoy et al., 2008). Mobile push advertising is generally sent with the users' permission. Thus, push advertising is beneficial for forming a positive attitude and intention to purchase. Therefore, this study is consistent with actual marketing situations and can satisfy the practical needs of companies.

The contributions of this research are primarily in two areas. First, this study employed RFT, TPB, and U&G to explore the formation of consumer purchase intention in mobile shopping apps. Secondly, it tested for moderated effects and moderated mediation effects in the mobile text advertising context.

## LITERATURE REVIEW AND HYPOTHESES Uses and Gratifications Theory (U&G)

U&G theory examines the psychological and behavioral utility produced by using mass communication and analyzes the use

**Abbreviations:** Apps, applications; SMS, short message service; U&G, uses and gratifications; TPB, theory of planned behavior; RFT, regulatory focus theory.

of motivating and need-satisfying information by audiences. Previous studies concluded that entertainment, informativeness, and irritation are vital dimensions of U&G (Ha et al., 2014). This theory has been widely used in consumer motivation research on traditional media and has made a substantial contribution to the field (Eighmey and McCord, 1998). Moreover, Ducoffe (1996) and Tsang et al. (2004) assumed that entertainment, informativeness, and irritation were related to consumer attitudes when browsing website and mobile advertisements. In this study, we investigated the impacts of informativeness, entertainment, and irritation on consumer attitudes after receiving mobile advertising from an app. This application to mobile advertising is a reasonable extension of U&G theory.

Entertainment and informativeness are interpenetrations in the new technologies (Wang and Sun, 2010). Hence, they can be integrated into one concept: infotainment (Okazaki, 2004). Studies show that whether the internet or mobile networks are involved, infotainment has a positive impact on attitudes, but irritation has a negative one (Okazaki, 2004; Shankar and Balasubramanian, 2009). Recent research also suggested that, as the intelligence of mobile functions increases, users will be able to increase their control of advertising information; consequently, the negative influence of irritation on consumer attitudes will also decrease (Watson et al., 2013). However, the way that these effects change has seldom been discussed. In this study, a promotion focus was investigated as a possible moderator of the positive effect of infotainment on attitudes toward mobile text advertising, whereas a prevention focus was investigated as possibly moderating the negative effect of irritation on attitudes toward mobile text advertising.

#### Infotainment

Information is a valuable incentive in mobile marketing because recipients of information tend to respond positively to advertising (Aitken et al., 2008). Therefore, informativeness, which primarily refers to the degree of information richness and the usefulness of a provided advertising media, is part of the infotainment construct (Ducoffe, 1996). Consumers expect information content that fits their interests (Robins, 2003) as well as information that is related to their own preferences and interests (Milne and Gordon, 1993) when they connect with mobile services. Some researchers have maintained that receiving suitable information from advertisers does not anger consumers (Varshney, 2003). At the same time, other researchers have shown that if advertising providers provide pleasant, relevant, and informative advertisements, the consumers' intention to purchase may increase (Scharl et al., 2005). Entertainment can deeply move customers and further familiarize them with the available services or goods (Liu et al., 2012). Thus, entertainment constitutes the other aspect of the infotainment concept.

Early research revealed that interesting and pleasurable advertising can actively influence attitude (Shimp, 1981). Moreover, a feeling of enjoyment was found to be central to a consumer's attitudes (Shavitt et al., 1998). Zhang and Mao (2008) held that perceived informativeness and perceived entertainment were components of perceived utility, and Rohm et al. (2012) explored the issue of whether perceived utility was a determinant of attitude toward mobile marketing. Furthermore, recent research showed that advertisements with more information than usual and a high level of entertainment can be readily accepted (Lin et al., 2014, 2016).

In this study, we investigated the attitudes toward mobile text advertising (branded communication). This topic is worth exploring because mobile advertising is a new and effective form of media advertisement. According to the latest forecast by eMarketer (2017), most of the Chinese advertising expenditures in the future will be invested in mobile advertising. Because a mobile device is a private good, such advertising can easily be delivered to targeted customers. At the same time, because it can conveniently be accessed and saved, mobile text advertising via apps has become a common marketing tool for a number of

companies. Thus, discussing customers' attitudes toward mobile text advertising is valuable.

Furthermore, users seem to feel that voluntarily downloaded apps are useful and valuable (Kang, 2014), and their attitudes toward them may be positive, even without any mobile text advertising. In addition, users' attitudes toward branded products are heavily influenced by their attitudes toward mobile advertising (Drossos et al., 2013). According to U&G theory, if users are satisfied with one type of mobile advertising, they will be likely to accept the included products brand information and buy the product. Therefore, attitudes toward mobile text advertising are the determinant of the intention to purchase. Focusing on the attitudes toward mobile text advertising is, therefore, worthwhile and suitable.

Based on the above, Hypothesis 1 was:

H1: Upon receiving a text advertisement from a previously voluntarily downloaded shopping app, consumers' attitudes toward mobile text advertising will be positively related to their perception of the infotainment level of this text advertisement.

#### Irritation

Irritation in the context of advertising can be described as the way that such advertising annoys consumers (Liu et al., 2012). The primary cause of criticisms of advertising is that it is an irritation (Bauer and Greyser, 1968).

When mobile advertising first began, the negative impact of irritation on consumer attitudes was considered to be great. "When advertising employs techniques that annoy, offend, insult, or are overly manipulative, consumers are likely to perceive it as an unwanted and irritating influence" (Ducoffe, 1995). Subsequently, Siau and Shen (2003) claimed that common people were unfamiliar with the concept of mobile commerce and were skeptical about the feasibility and security of mobile advertising. Moreover, when consumers received an excessive number of advertisements from a provider, it disturbed them (Bruner and Kumar, 2007). More recently, with the improvement of the smartphone, others argued that, although the irritation of SMS negatively affects attitudes in a mobile context (Liu et al., 2012), the extent is less than before (McCoy et al., 2008). Thus, the following hypothesis was formulated:

H2: Upon receiving a text advertisement from a previously voluntarily downloaded shopping app, consumers' attitudes toward mobile text advertising will be negatively related to their perception of the irritation level of this text message.

### Theory of Planned Behavior (TPB)

Ajzen (1985) proposed a general model, the TPB, to explain the unique behavior of individuals. It reflects an individual's beliefs, attitudes, and intentions. The TPB is an extension of the theory of reasoned action (Ajzen and Fishbein, 1975, 1980). In the expectations–value model (Fishbein and Ajzen, 1974), attitudes are considered to be the consequence of individual beliefs. Subjective norms refer to the extent that the perception of opinions from his/her reference groups, such as friends and colleagues, impact an individual's behavior (Schofield, 1974). Perceived behavioral control refers to an individual's subjective appraisal of their own ability and their opportunities to carry out a specific behavior (Ajzen and Madden, 1986). Intention is another vital indicator of an individual's behavior (Ajzen, 1991). The TPB holds that an individual's behavior is impacted by her/his beliefs, attitudes, subjective norms, and other uncontrollable factors (Crespo and Bosque, 2008). In this study, these three variables were used to investigate the consumer's intention to purchase the advertised product. These variables have been used in several previous studies to explore online consumer behavior intention (Luarn and Lin, 2005; Kim et al., 2011; Lin and Chen, 2015).

An individual's attitude tends to be a determinant of his/her behavioral intention (Ajzen and Fishbein, 1975). Many studies have showed a positive relationship between the attitude toward an advertisement and an intention to purchase (Drossos et al., 2013). Subjective norms were found to play an important role in predicting people's intention to consume fast food (Bagozzi et al., 2000) and have been perceived as positively shaping young Chinese people's intention to use mobile services (Zhang and Mao, 2008). Following the TPB, this study assumed that attitudes, subjective norms, and perceived behavioral control affect a consumer's purchase intention. Moreover, attitude is typically influenced by the opinions of others (Kim et al., 2013). Thus, the following hypotheses were investigated:


### Regulatory Focus Theory (RFT)

RFT is a goal pursuit theory. It was articulated by Higgins et al. (1997) to address people's perceptions during their decisionmaking process. Higgins (2000) believed that the consumer decision process stems from a variety of motives. Motivation can be divided into two types: one is to achieve a desired goal, and the other is to avoid an appearance of negative results (Lockwood et al., 2002; Das, 2016). RFT proposes that regulatory focus is a special propensity of individuals when they self-regulate to achieve their goals (Higgins et al., 1997; Brockner et al., 2004). RFT is a chronic manifestation and includes promotion focus and prevention focus, two independent orientations. Promotion focus refers to people's motivation to seek pleasure and is related to an ideal target, that is, that they can obtain a desired outcome. During self-regulation, a radical strategy is required to maximize the probability of obtaining a positive outcome, such as improving the benefit (Crowe and Higgins, 1997; Freitas, 2002). Meanwhile, prevention focus refers to people's motivation to avoid pain and is related to a "should be" target, that is, that they should experience no loss or punishment. A prevention focus usually generates a cautious strategy to minimize the occurrence of a negative outcome, such as reducing a loss (Crowe and Higgins, 1997; Kirmani and Zhu, 2007; Lee et al., 2014). Although a person's regulatory focus is a long-term personality trait, it can be triggered to impact a temporary motivation in a specific situation (Wilson and Ross, 2000; Lockwood et al., 2002).

RFT has been widely used in the field of consumer behavior (Avnet and Higgins, 2006; Wang and Sun, 2010; Ouyang et al., 2015). Scholars have found that promotion and prevention focuses significantly influence consumer decision making (Lee and Santanam, 2014; Das, 2016). According to RFT, a prevention focus causes consumers to perceive a loss from utilizing bundled sales to be higher than it really is (Lin and Huang, 2011), whereas a promotion focus encourages consumers to be more mindful of future results (Joireman et al., 2012). Furthermore, individuals' promotion and prevention focuses have different effects on their intention to purchase, on their participation in online product reviews, and on their willingness to spread a positive word of mouth (Das, 2016). Most of these studies primarily involved Western consumers and are more directed toward traditional and internet marketing environments. Relatively little research exists on Asian consumers in the mobile marketing environment (Zhang and Mao, 2008; Lin and Chen, 2015). This study focused on the different motivations of Chinese consumers in response to mobile text advertising and explored the manner in which a promotion focus or a prevention focus can affect purchasing decisions.

#### Moderation Effect

According to RFT, when the type of regulatory focus matches the implemented strategy, individuals evaluate that strategy highly and have positive attitudes (Wang and Lee, 2006). Therefore, some researchers proposed that companies should emphasize different aspects when facing different consumers. Specifically, acquiring greater potential revenue will attract promotion focus consumers, whereas avoiding potential losses will draw prevention focus consumers (Higgins, 2000; Wang and Lee, 2006; Park and Morton, 2015). Moreover, other scholars argued that a high level of promotion focus will result in the consumers' high evaluation of hedonic goods (Hassenzahl et al., 2008). Thus, we supposed that high promotion focus individuals will be inclined to take positive measures to add to their enjoyment; thus, the influence of infotainment on attitudes will be intensified. However, people with a high prevention focus will tend to take protective actions to avoid possible loss, thus the negative impact of irritation on their attitudes will be strengthened. The hypotheses are as follows:

H7a: Upon receiving a text advertisement from a previously voluntarily downloaded shopping app, a high promotion focus will strengthen the positive relationship between infotainment and attitudes toward mobile text advertising. H7b: Upon receiving a text advertisement from a previously voluntarily downloaded shopping app, a high prevention focus will strengthen the negative relationship between irritation and attitudes toward mobile text advertising.

#### Moderated Mediation Effect

Consequently, promotion focus consumers will be inclined toward using promotion strategies to increase pleasure and attain their goals, whereas prevention focus consumers will tend to use protection strategies to avoid losses. In particular, the infotainment of mobile text advertising can be expected to allow the promotion focus consumers to form more positive attitudes and a greater intention to purchase, whereas irritation stemming from mobile text advertising will generate more negative attitudes and reduce the intention to purchase in prevention focused individuals. Thus, we assumed:


## RESEARCH DESIGN

One common form of current commercial activity—mobile shopping apps—was selected as the source of mobile advertising information. A survey that was aimed to measure their response was given to a group of consumers after they received a text advertisement. An electronic questionnaire was used for data collection. Each participant was chosen randomly if he or she had previously had a shopping experience with a famous Chinese mobile shopping app, Taobao, and had purchased products from the same app at least once in the previous month.

### Data Collection

The investigated app was the Taobao app. Its usage percentage in China was 24.1% by the end of 2016 (CNNIC, 2017). By the latest statistics from China Internet Network Information Center (CNNIC, 2017), the structure of total Chinese cybercitizens is as follows: the male to female ratio is 52:48; 30.3% are aged 20–29, 23.2% are aged 30–39; 11.5% hold a bachelor's degree; 25% are students, and 64.3% are currently employed.

According to investigational results from one Chinese mobile data business service company, Questmobile, the Taobao app had up to 406 million subscribers by the first quarter of 2017 (QuestMobile, 2017). The demographic profile of its users in China, as of May 2016 was as follows: 49% male, 51% female and 72% of the users were aged 18–34. We have no reason to doubt that this is representative of Chinese shopping apps.

The number of people who responded to the questionnaire was 431, but 81 did not meet the two requirements after answering the first two questions, so they were not permitted to continue with the survey. Thus, a total of 350 respondents finished our survey with 320 valid questionnaires (a response rate of 74%). 171 (53%) were female, and 149 (47%) were male. Most of the participants were young; 240 (75%) respondents were aged 22 to 35 years. In terms of educational attainment, 234 (73%) of the respondents had a bachelor's degree. Moreover, 243 (76%) respondents were employed. The characteristics of these investigated data are broadly consistent with the total population of Taobao app users, so we believe that our data is a representative sample.

#### Measurement

All of the measurement items in this study were adopted from previous research. The promotion and prevention focus scales are from Higgins et al. (2001) as translated by Ouyang et al. (2015) into a Chinese context to give them better credibility and reliability. This study directly used their scales to appropriately assess Chinese consumers. We translated the remaining scales from English to Chinese and implemented a back-translation to ensure the consistency and clarity of the questions' expression (Brislin, 1980). To examine the content validities, the translated questionnaire was assessed by two bilingual faculty members. A pretest was conducted to simplify the questions and to ensure that they precisely measured the concepts. Based on the feedback, confusing items were revised. The Cronbach's alpha for all the items was 0.89, which is considerably higher than 0.70, which was the level suggested by Hollis et al. (1978) as acceptable. The main variable measurement items are listed in **Table 1**.

Related research previously showed that demographic variables (gender, age, education level, and job) are among the important factors influencing attitudes (Wolin and Korgaonkar, 2003; Habuchi et al., 2005). Thus, these factors were used as control variables.

All the items were measured using a five-point Likert scale. The measurement for regulatory focus was from never (1) to always (5), whereas the others were from strongly disagree (1) to strongly agree (5).

#### DATA ANALYSIS AND RESULTS

#### Data Analysis

#### Reliability and Validity

All of the factor loadings were highly significant (p < 0.01) to their respective latent constructs; the composite reliabilities of all of the constructs were > 0.80; and all of the average variance extracted estimates were > 0.60 (**Table 1**; Fornell and Larcker, 1981). Therefore, the measurements had adequate convergent validity and composite reliability (Bagozzi and Yi, 1988).

#### Confirmatory Factor Analysis (CFA)

CFAs were implemented to examine the discriminant validity of the key variables. First, an eight-factor CFA model (see **Table 2**) was tested. This model included infotainment, irritation, promotion focus, prevention focus, attitudes, subjective norms, perceived behavioral control, and intention to purchase. The measurement model fitted the data (χ <sup>2</sup> = 382.33, df = 247, TLI = 0.97, CFI = 0.94, RMSEA = 0.04), confirming the unidimensionality of the measures (Anderson and Gerbing, 1988). Subsequently, the eight-factor model was compared with alternative models that sequentially combined infotainment with each of the 7 other factors (**Figure 2**). A model comparison revealed that the eight-factor model fit the data considerably better than the alternative model. The means, standard deviations (SDs), and correlations are shown in **Table 3**.

#### Common Method Issues

Given that infotainment, irritation, promotion focus, prevention focus, subjective norms, perceived behavioral control, attitudes, and intention to purchase were obtained from the same sources, a potential common method variance was a concern. Several techniques were used to minimize any common source bias. First, the confidentiality of the responses was guaranteed to the participants to limit any concerns, such as the participants' evaluation apprehension or social desirability. Second, a psychological separation was constructed in the survey by using different instructions and positioning the variables in various parts of the survey with a number of filler items between them to try to minimize the respondents' perception of any direct connection between the variables (Podsakoff et al., 2003). Finally, the variance explained by the Harmon one-factor test was 19.63%, which was below the cut-off value of 25%. Hence, common method variance was an unlikely constraint of the results in this study (Peng et al., 2016).

## Test of Hypotheses

#### Hypothesized Model

We used the structural equation modeling technique (Anderson and Gerbing, 1988) to examine the hypotheses of this study. This technique allows a simultaneous examination of an entire system of variables in a hypothesized model (Byrne, 2009). As presented in **Table 4**, χ 2 (356) <sup>=</sup> 512.33, <sup>p</sup> <sup>&</sup>lt; 0.001, CFI <sup>=</sup> 0.97, TLI = 0.96, and RMSEA = 0.04, the empirical results indicated that the hypothesized model fit the data well.

#### Model Comparisons

The change in chi-square test (Bentler and Bonett, 1980) was utilized to compare the hypothesized model with nested models, which were less likely but nevertheless plausible based on theoretical arguments (Wang and Chen, 2005). Nested Models 1–6 were included.

Nested Model 1 assessed the same path as in the hypothesized model, along with the direct path from infotainment to intention to purchase. Nested Model 2 also tested the direct path from infotainment∗promotion focus to intention to purchase in addition to the hypothesized paths. Nested Model 3 tested the direct path from promotion focus to intention to purchase in addition to the hypothesized paths. Nested Model 4 examined the direct path from irritation to intention to purchase based on the hypothesized model. Nested Model 5 examined the direct path from irritation∗prevention focus to intention to purchase in addition to the paths of the hypothesized model. Nested Model 6

#### TABLE 1 | Main variable measurement items, reliability and factor loadings.


N = 320; C.R. = composite reliability; AVE = average variance extracted; α = Cranach's alpha.

also checked the direct path from prevention focus to intention to purchase along with the paths of the hypothesized model.

The change in chi-square tests indicated that Nested Model 1 (1χ<sup>2</sup> = 0.01; 1df = 1; p > 0.05), Nested Model 2 (1χ<sup>2</sup> = 1.49; 1df = 1; p > 0.05), Nested Model 3 (1χ<sup>2</sup> = 1.15; 1df = 1; p > 0.05), Nested Model 4 (1χ<sup>2</sup> = 2.28; 1df = 1; p > 0.05), Nested Model 5 (1χ<sup>2</sup> = 1.33; 1df = 1; p > 0.05), and Nested Model 6 (1χ<sup>2</sup> = 0.07; 1df = 1; p > 0.05) were not significantly better than the hypothesized model and were less parsimonious (see **Table 4**). Hence, the hypothesized model was the most parsimonious.

**Table 5** shows the results of the hypothesized model. The path coefficients were all significant. Infotainment (β = 0.21, p < 0.01) was positively related to attitudes, whereas irritation (β = −0.18, p < 0.01) was negatively related to attitudes. Thus, Hypotheses 1 and 2 were both supported. In addition, intention to purchase was related to attitudes (β = 0.27, p < 0.01), perceived behavioral control (β = 0.24, p < 0.01), and subjective norms (β = 0.27, p < 0.01). Hence, Hypotheses 3, 4, and 5 were supported. In addition, subjective norms were positively connected with attitudes (β = 0.19, p < 0.01). Thus, Hypothesis 6 was supported.

TABLE 2 | Results of the CFA for the measures of the studied variables.


N = 320; \*\*p < 0.01 (2-tailed).

#### Moderation Effect

As shown in **Table 5**, promotion focus (β = 0.26, p < 0.01) was positively related to attitudes, and the interaction between infotainment and promotion focus (β = 0.23, p < 0.01) was also positively related to attitudes. Thus, H7a was supported. Prevention focus (β = −0.20, p < 0.01) was negatively related to attitudes, as was the interaction between irritation and prevention focus (β = −0.23, p < 0.01). Hence, H7b was supported.

Furthermore, the interaction effects were plotted using procedures from Aiken et al. (1994) and Dawson (2014). As shown in **Figure 2**, infotainment was positively related to attitude when promotion focus was high (β = 0.49, p < 0.01), and this positive influence was lessened when promotion focus was low (β = 0.14, p < 0.05), further supporting Hypothesis 7a. Moreover, **Figure 3** indicates that irritation was negatively related to attitude (β = −0.44, p < 0.01) when prevention focus was high but was unrelated to attitude (β = −0.05, n.s.) when prevention focus was low (n.s. = non-significant). Hence, Hypothesis 7b was further supported.

#### Moderated Mediation Effects

A moderated path approach was utilized to test the moderated mediation hypotheses (Edwards and Lambert, 2007), and 1,000 samples were bootstrapped to compute the bias-corrected confidence intervals. An indirect effect can be viewed as statistically significant if the 95% confidence interval excludes zero (Edwards and Lambert, 2007). As shown in **Table 6**, an indirect effect of infotainment on intention to purchase via attitude was positive and significant when promotion focus was high (path coefficient (P) = 0.20, 99.5% CI [0.07, 0.39]) and was also significant when promotion focus was low (P = 0.05, 95% CI [0.01, 0.12]; difference = 0.15, 99.5% CI [0.02, 0.33]). Similarly, **Table 7** shows that the indirect effect of irritation on intention to purchase via attitude was negative and significant when prevention focus was high (P = −0.11, 99.5% CI [−0.22, −0.01]) but was non-significant when prevention focus was low (P = −0.01, 95% CI [−0.06,0.02]; difference = −0.10, 95% CI [−0.17, −0.03]). These results provide support for Hypotheses 8a and 8b.

### GENERAL DISCUSSION AND IMPLICATIONS

The goal of this study was to analyze empirically the influence of mobile text advertising on consumers' purchase intention. With


N = 320; \*\*p < 0.01; \*p < 0.05 (2-tailed), SD, Standard Deviation.

TABLE 4 | Comparisons of the structural equation models.


N = 320; \*\*p < 0.01; \*p < 0.05 (2-tailed); <sup>a</sup>Compared to the hypothesized model.


N = 320, \*\*p < 0.01 (2-tailed), AT = attitudes toward mobile text advertising, INTER1 = infotainment \* promotion focus, INTER2 = irritation \* prevention focus, SN = subjective norms, PBC = perceived behavioral control, ITP = intention to purchase.

the rapid development of smart mobile and internet technologies, mobile shopping apps have become convenient and efficient. Mobile marketing is now essential to many companies' total marketing strategies. The question arises whether mobile text advertising, sent by mobile shopping apps, facilitates recipients'

purchase intention and how this process works. The findings of this study address these questions and provide insights from both theoretical and practical perspectives.

#### TABLE 6 | The results of moderated path analysis for Infotainment.


N = 320; \*\*p < 0.01; \*p < 0.05 (2-tailed).

PMX: path from infotainment to attitudes; PYM: path from attitudes to intention to purchase; PYX : path from infotainment to intention to purchase. Low promotion focus refers to one standard deviation below the mean of promotion focus; high promotion focus refers to one standard deviation above the mean of promotion focus. Tests of differences between the indirect and total effects were based on bias-corrected confidence intervals derived from bootstrap estimates.

#### TABLE 7 | The results of moderated path analysis for Irritation.


N = 320; \*\*p < 0.01; \*p < 0.05 (2-tailed).

PMX:path from irritation to attitudes; PYM: path from attitudes to intention to purchase; PYX : Path from irritation to intention to purchase. Low prevention focus refers to one standard deviation below the mean of prevention focus; high prevention focus refers to one standard deviation above the mean of prevention focus. Tests of differences for the indirect and total effect were based on bias-corrected confidence intervals derived from bootstrap estimates.

### Conclusions

First, infotainment and irritation influenced attitudes toward mobile text advertising in opposing directions in accordance with prior research (H1, H2). The present study shows that two main dimensions of U&G, infotainment and irritation, impacted attitudes in opposing directions. The impact of infotainment was greater than that of irritation. This finding confirms previous views about mobile advertising (Tsang et al., 2004). Mobile advertisements tend to have more infotainment and less irritation to increase consumer contentment (McCoy et al., 2008).

Second, the study reveals the moderation effects of regulatory focus between information content characteristics and attitudes toward mobile text advertising (H7a, H7b). In accordance with RFT, this study showed that at a high level of promotion focus, the positive relationship between infotainment and attitudes was strengthened; conversely, at a high level of prevention focus, the negative relationship between irritation and attitudes was strengthened. These results can deepen the understanding of attitudes and benefit subsequent research.

Finally, the mechanism behind the intention to purchase mobile-advertised products was explored. After receiving mobile text advertisements from mobile shopping apps, the consumers' attitudes toward mobile text advertising, together with their subjective norms and perceived behavioral controls, influenced the intention to purchase (H3, H4, and H5). Moreover, this study suggested that subjective norms indirectly impacted the intention to purchase through their attitudes toward mobile text advertising (H6). According to reference group theory, the views of others considerably affect an individual's beliefs and their purchase decisions. Our study indicated that this was true in a mobile advertising context. Additionally, two moderated mediation effects were identified (H8a, H8b). One mediation effect of attitudes toward mobile text advertising between infotainment and intention to purchase was moderated by the promotion focus; whereas, the other mediation effect of attitudes toward mobile text advertising between irritation and intention to purchase was moderated by the prevention focus.

In addition, due to the similarities between the Taobao app and other mobile shopping apps, the above conclusions are also applicable to similar apps. The Taobao app has the largest number of Chinese shoppers (QuestMobile, 2017) and also has a range of products and services which covers most online items. Estimates are that 82% of Chinese internet advertising will be invested in mobile trading platforms by the year 2021 (eMarketer, 2017). Therefore, delivering text advertising via apps can be expected to become a prosperous method. Our findings about mobile text advertising via the Taobao app can be expected to be basically stable.

### Theoretical Implications

First, our findings shed light on the use of (promotion focus and prevention focus) regulatory focus in mobile advertising activities. Previous studies on mobile advertising were mostly based on either the single theory of reasoned action (Tsang et al., 2004) or the theory of acceptance model (Muk and Chung, 2015). In the present study, the regulatory focus was found to not only significantly influence the relationship between information characteristics and attitudes but also to moderate the mediation effects of attitudes on information characteristics and the intention to purchase. The empirical test showed that a high promotion focus strengthened the impact of infotainment on attitudes as well as that attitudes mediated the relationship between infotainment and the intention to purchase. Moreover, a high prevention focus increased the impact of irritation on attitudes in addition to reinforcing the mediated effect of attitudes between irritation and the intention to purchase. The results can not only contribute to the further understanding of consumer behavior characteristics with regard to the mobile field but can also shed light on related research based on the RFT in the context of mobile commerce. Furthermore, this study extends the application of RFT to the mobile context and other mobile service activities.

Second, this study helps to elucidate the mechanism of the purchase intention in the mobile context, indicating that it is an extension of the TPB. Most extant studies have discussed the intention to purchase only from the perspective of the TPB. This study, which investigated the intention to purchase from the perspective of regulatory focus, identified a moderating role for regulatory focus, mediating effects of attitudes, and moderated mediation effects. The results of this study improve the understanding of the mechanism underlying the intention to purchase in a mobile context. The results also highlighted the importance of consumer motivation in the mobile field. Moreover, this study laid a foundation for conducting related behavioral intention research in different contexts.

Third, the results of our research appear to be representative and universal. Extant studies on mobile advertising have primarily focused on no-income groups, such as university students. In this study, 95% of the respondents had had mobile shopping experience for at least 1 year and 76% were employees from various companies and thus had a certain economic foundation. Hence, we believe that our data is representative and our findings will be valuable. Furthermore, the empirical results showed that the participants had either a positive attitude (**Table 4**, the attitudes mean = 3.32) or a near neutral irritation (**Table 4**, irritation mean = 2.94), a finding which is consistent with the notion that the negative impact of irritation has become weaker than in the past (Xu et al., 2009).

### Practical Implications

In addition to the above theoretical implications, this study has the following practical implications.

First, the findings highlighted the importance of subjective norms. Although attitudes are usually emphasized in consumer behavior intention research, managers must carefully attempt to

Second, the findings can enlighten designers about the design of mobile advertising. Managers often try to avoid irritating consumers. Our results suggest that they should also concentrate on increasing the infotainment to please consumers. Avoidance alone is not enough. A high level of infotainment is beneficial to the formation of a positive attitude and to the purchase intention. Advertising designers may need to increase the informativeness and entertainment of mobile advertising content.

Third, the findings can contribute to perfecting the marketing strategy of various companies. In general, marketing managers employ strategies aimed at managing the information content characteristics of mobile advertising (Tsang et al., 2004; Xu et al., 2009; Liu et al., 2012). This study suggested that they should pay attention not only to the information content of mobile advertising but also to the regulatory focus (promotion focus and prevention focus). For example, delivering enjoyable mobile text advertising to promotion focus customers may contribute to creating better attitudes and a stronger intention to purchase; but how to identify promotion focus customers and prevention focus ones is the question. The literature indicates that a promotion focus can be triggered temporarily in some situations (Lockwood et al., 2002). Managers could motivate consumers to purchase by activating each individual's promotion focus by some mobile promotion activities. Furthermore, the method we used in this article can serve as a reference for managers. By sending the online testing items of about a regulation focus to their customers, they may be able to distinguish between them. Analyzing the existing customer database is another possible method (Das, 2016). In addition, promotion focus individuals are inclined to change the current situation, whereas prevention focus ones are inclined to maintain the status quo (Chernev, 2006; Herzenstein et al., 2007). Hence, managers could deliver the customized information about their goods in accordance with the recipient's regulatory focus. For instance, for the promotion focus customers, receiving new or high technology products text advertising from shopping apps would help to draw attention and improve the products' brand awareness, whereas sending information about how the product will help the recipients maintain their status quo will tend to attract prevention focus individuals.

### LIMITATIONS AND FUTURE RESEARCH

Although, all of the hypotheses in this study were supported, the study still presents limitations that can provide opportunities for future research. First, only push mobile advertising was examined. Future research should explore the effectiveness of pull mobile advertising and compare the two forms. Second, the study only concentrated on the text form of mobile advertising. In the future, other types of mobile advertising, such as multimedia mobile advertising, should be investigated as well. Third, this study explored the moderation effect only in a mobile environment. The RFT's influence in the contentattitudes relationship, as well as its moderation effect in the attitudes-intentions relationship, is worth studying as a general TPB modifier as well as in a mobile advertising context in further scenarios. In addition, our analysis may lag behind the rapid development of mobile advertising and communication technology to some extent. Because of these rapid advances, the formation mechanism of the intention to purchase will need to be continually updated. Fourth, we primarily included subjects who were fairly well-educated and investigated their responses to a single shopping app. According to statistics from the China Internet Network Information Center (CNNIC, 2017), college students are the major users of Chinese cybercitizens. However, a further study utilizing a different range of the population and more shopping apps should also be performed to strengthen its robustness. Finally, the study discussed chronic promotion and prevention focuses, but these may have been triggered in some situations. Future research should examine the manner in which they can be triggered and the things that influence them.

### ETHICS STATEMENT

This research was approved by Economics and Management Department Research Ethic Committee. All participants gave

### REFERENCES


informed consent to participate. All participants were fully insured for the duration of the questionnaire.

### AUTHOR CONTRIBUTIONS

LH contributes the writing of the whole article. CZ contributes the data collection.

#### ACKNOWLEDGMENTS

The authors are grateful to Rhoda E. and Edmund F. Perozzi, Ph.D.s, for extensive editing and discussions. This work was funded by the Society and Science Union Project of Fujian Province, P.R. China, FJ2016B082 "Influence of mobile advertising on the users' behavior intention," FJ2015B231 "Research on the financial support for the development of the private economy in Fujian Province under the new normal economy"; and the Project of Education Department of Fujian Province, P.R. China, JAS150418 "Users adoption of Short Message Service advertising in the mobile commercial context."


Eighmey, J., and McCord, L. (1998). Adding value in the information age: uses and gratifications of sites on the World Wide Web. J. Bus. Res. 41, 187–194. doi: 10.1016/S0148-2963(97)00061-1

eMarketer (2015). Mobile Ad Spend to Top \$100 Billion Worldwide in 2016, 51% of Digital Market. Availble online at: http://www.emarketer.com/Article/Mobile-Ad-Spend-Top-100-Billion-Worldwide-2016-51-of-Digital-Market/1012299

eMarketer (2017). Mobile Ad Spending Share of Digital Ad Spending Worldwide 2016–2021 of Total Digital Ad Spending. Available online at: http://www.emarketer.com/Chart/Mobile-Ad-Spending-Share-of-Digital-Ad-Spending-Worldwide-2016-2021-of-total-digital-ad-spending/205576, http://www.199it.com/archives/572399.html

Fishbein, M., and Ajzen, I. (1974). Attitudes towards objects as predictors of single and multiple behavioral criteria. Psychol. Rev. 81, 59–74. doi: 10.1037/h0035872

Fornell, C., and Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. J. Mark. Res. 18, 39–50. doi: 10.2307/3151312

Freitas, A. A. (2002). Introduction Data Mining and Knowledge Discovery with Evolutionary Algorithms. Heidelberg: Springer.

Grewal, D., Bart, Y., Spann, M., and Zubcsek, P. P. (2016). Mobile advertising: a framework and research agenda. J. Interact. Mark. 34, 3–14. doi: 10.1016/j.intmar.2016.03.003

Ha, Y. W., Kim, J., Libaque-Saenz, C. F., Chang, Y., and Park, M. C. (2014). Use and gratifications of mobile SNSs: facebook and kakaotalk in Korea. Telemat. Inform. 32, 425–438. doi: 10.1016/j.tele.2014.10.006

Habuchi, I., Dobashi, S., Tsuji, I., and Iwata, K. (2005). Ordinary usage of new media: internet usage via mobile phone in Japan. Int. J. Jap. Sociol. 14, 94–108. doi: 10.1111/j.1475-6781.2005.00071.x

Hassenzahl, M., Schöbel, M., and Trautmann, T. (2008). How motivational orientation influences the evaluation and choice of hedonic and pragmatic interactive products: the role of regulatory focus. Interact. Comput. 20, 473–479. doi: 10.1016/j.intcom.2008.05.001

Herzenstein, M., Posavac, S. S., and Brakus, J. J. (2007). Adoption of new and really new products: the effects of self-regualtion systems and risk salience. J. Mark. Res. 44, 251–260. doi: 10.1509/jmkr.44.2.251

Higgins, E. T. (2000). Making a good decision: value from fit. Am. psychol. 55, 1217–1230. doi: 10.1037/0003-066X.55.11.1217

Higgins, E. T., Friedman, R. S., Harlow, R. E., Idson, L. C., Ayduk, O. N., and Taylor, A. (2001). Achievement orientations from subjective histories of success: promotion pride versus prevention pride. Eur. J. Soc. Psychol. 31, 3–23. doi: 10.1002/ejsp.27

Higgins, E. T., Shah, J., and Friedman, R. (1997). Emotional responses to goal attainment: strength of regulatory focus as moderator. J. Pers. Soc. Psychol. 72, 515–525. doi: 10.1037/0022-3514.72.3.515

Hollis, D. P., Nunnally, R. L., Taylor, G. J. IV., Weisfeldt, M. L., and Jacobus, W. E. (1978). Phosphorus nuclear magnetic resonance studies of heart physiology. J. Magn. Reson. 29, 319–330. doi: 10.1016/0022-2364(78)90156-7

Hwang, J., Yoon, Y.-S., and Park, N.-H. (2011). Structural effects of cognitive and affective reponses to web advertisements, website and brand attitudes, and purchase intentions: the case of casual-dining restaurants. Int. J. Hospitality Manage. 30, 897–907. doi: 10.1016/j.ijhm.2011.01.011

Joireman, J., Shaffer, M. J., Balliet, D., and Strathman, A. (2012). Promotion orientation explains why future-oriented people exercise and eat healthy: evidence from the two-factor consideration of future consequences-14 scale. Personal. Soc. Psychol. Bull. 38, 1272–1287. doi: 10.1177/0146167212449362

Kang, S. (2014). Factors influencing intention of mobile application use. Int. J. Mobile Commun. 12, 360–379. doi: 10.1504/IJMC.2014.063653

Kim, E., Ham, S., Yang, I. S., and Choi, J. G. (2013). The roles of attitude, subjective norm, and perceived behavioral control in the formation of consumers' behavioral intentions to read menu labels in the restaurant industry. Int. J. Hospital. Manage. 35, 203–213. doi: 10.1016/j.ijhm.2013.06.008

Kim, H., Chan, H. C., and Gupta, S. (2007). Value-based Adoption of Mobile internet: an empirical investigation. Decision Support Syst. 43, 111–126. doi: 10.1016/j.dss.2005.05.009

Kim, K. K., Shin, H. K., and Kim, B. (2011). The role of psychological traits and social factors in using new mobile communication services. Electron. Commer. Res. Appl. 10, 408–417. doi: 10.1016/j.elerap.2010.11.004

Kirmani, A., and Zhu, R. (2007). Vigilant against manipulation: the effect of regulatory focus on the use of persuasion knowledge. J. Market. Res. 44, 659–670. doi: 10.1509/jmkr.44.4.688

Lee, G., and Santanam, R. (2014). Determinants of mobile apps' success: evidence from the app store market. J. Manage. Inform. Syst. 31, 133–169. doi: 10.2753/MIS0742-1222310206

Lee, K., Choi, J., and Li, Y. J. (2014). Regulatory focus as a predictor of attitudes toward partitioned and combined pricing. J. Consum. Psychol.24, 355–362. doi: 10.1016/j.jcps.2014.01.001

Lin, C. H., and Huang, C. W. (2011). The effects of regulatory focus and bundlepricing framing on consumers' perceptions of loss. Soc. Behav. Personal. Int. J. 39, 113–117. doi: 10.2224/sbp.2011.39.1.113

Lin, H., and Chen, Z. (2015). Influence of SMS Advertising on Consumer Behavioral Intention. J. Organ. End User Comput. 27, 25–42. doi: 10.4018/JOEUC.2015100102

Lin, H., Zhou, X., and Chen, Z. (2014). Impact of the content characteristic of short message service advertising on consumer attitudes. Soc. Behav. Personal. Int. J. 42, 1409–1419(1411). doi: 10.2224/sbp.2014.42.9.1409

Lin, T. T. C., Paragas, F., Goh, D., and Bautista, J. R. (2016). Developing locationbased mobile advertising in Singapore: a socio-technical perspective. Technol. Forecast. Soc. Change 103, 334–349. doi: 10.1016/j.techfore.2015.06.002

Liu, C. L. E., Sinkovics, R. R., Pezderka, N., and Haghirian, P. (2012). Determinants of consumer perceptions toward mobile advertising—a comparison between Japan and Austria. J. Interact. Market. 26, 21–32. doi: 10.1016/j.intmar.2011.07.002

Lockwood, P., Jordan, C. H., and Kunda, Z. (2002). Motivation by positive or negative role models: regulatory focus determines who will best inspire us. J. Pers. Soc. Psychol. 83, 854–864. doi: 10.1037/0022-3514.83.4.854

Luarn, P., and Lin, H. H. (2005). Towards an understanding of the behavioral intention to use mobile banking. Comput. Hum. Behav. 21, 873–891. doi: 10.1016/j.chb.2004.03.003

McCoy, S., Everard, A., Polak, P., and Galletta, D. F. (2008). An Experimental study of antecedents and consequences of online Ad intrusiveness. Int. J. Hum. Comput. Interact. 24, 672–699. doi: 10.1080/10447310802335664

Milne, G. R., and Gordon, M. E. (1993). Direct mail privacy-efficiency tradeoffs within an implied social contract framework. J. Public Policy Market. 12, 206–215.

Molitor, D., Reichhart, P., and Spann, M. (2012). Location-based advertising: measuring the impact of context-specific factors on consumers' choice behavior. Soc. Sci. Electron. Publ. 29, 252–253. doi: 10.2139/ssrn.2116359

Muk, A., and Chung, C. (2015). Applying the technology acceptance model in a two-country study of SMS advertising. J. Bus. Res. 68, 1–6. doi: 10.1016/j.jbusres.2014.06.001

Ngai, E. W. T., and Gunasekaran, A. (2007). A review for mobile commerce research and applications. Decis. Support Syst. 43, 3–15. doi: 10.1016/j.dss.2005.05.003

Okazaki, S. (2004). How do Japanese consumers perceive wireless ads? A multivariate analysis. Int. J. Advert. 23, 429–454. doi: 10.1080/02650487.2004. 11072894

Ouyang, Y., Zhu, Y., Fan, W., Tan, Q., and Zhong, Y. (2015). People higher in self-control do not necessarily experience more happiness: regulatory focus also affects subjective well-being. Personal. Indiv. Diff. 86, 406–411. doi: 10.1016/j.paid.2015.06.044

Park, S. Y., and Morton, C. R. (2015). The role of regulatory focus, social distance, and involvement in anti-high-risk drinking advertising: a construallevel theory perspective. J. Advert. 44, 338–348. doi: 10.1080/00913367.2014. 1001503

Park, T., Shenoy, R., and Salvendy, G. (2008). Effective advertising on mobile phones: a literature review and presentation of results from 53 case studies. Behav. Inf. Technol. 27, 355–373. doi: 10.1080/01449290600958882

Peng, L., Liao, Q., Wang, X., and He, X. (2016). Factors affecting female user information adoption: an empirical investigation on fashion shopping guide websites. Electron. Comm. Res. 16, 145–169. doi: 10.1007/s10660-016-9213-z

Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., and Podsakoff, N. P. (2003). Common method biases in behavioral research: a critical review of the literature and recommended remedies. J. Appl. Psychol. 88, 879–903. doi: 10.1037/0021-9010.88.5.879


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Hongyan and Zhankui. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Do Affective Variables Make a Difference in Consumers Behavior Toward Mobile Advertising?

María Pilar Martínez-Ruiz <sup>1</sup> \*, Alicia Izquierdo-Yusta<sup>2</sup> , Cristina Olarte-Pascual <sup>3</sup> and Eva Reinares-Lara<sup>4</sup>

<sup>1</sup> Department of Business Administration, University of Castilla-La Mancha, Albacete, Spain, <sup>2</sup> Department of Business Administration, University of Burgos, Burgos, Spain, <sup>3</sup> Department of Economics and Business Studies, Universidad de La Rioja, Logroño, Spain, <sup>4</sup> Department of Business Administration, Universidad Rey Juan Carlos, Madrid, Spain

Research into permission-based mobile marketing is increasingly common due to the widespread adoption of mobile technology and its use as a communication channel. Yet few studies have attempted to analyze the factors that determine attitudes toward mobile advertising while simultaneously considering: the links among them and consumers' intentions, behavior, and/or cognitive and affective variables simultaneously. The present research therefore sought to deepen understanding of the antecedents and consequences of attitudes toward permission-based mobile advertising. More specifically, it sought to identify the antecedents of attitudes toward mobile advertising and the bridges between these attitudes and consumers' intentions upon receiving advertising on their mobile devices. To this end, a causal model was proposed and tested with a sample of 612 mobile phone users that was collected from a panel of Spanish adults who receive advertising on their mobile phones in the form of SMS text messages. The structural model used was validated using the partial least squares (PLS) regression technique. The results show that the greatest influence was that exerted by positive emotions on feelings, suggesting that positive emotions have an indirect effect on attitude toward mobile advertising. This influence was even greater than their direct effect. Another important, though less powerful, effect was the influence of attitude on behavioral intentions to receive mobile advertising. In contrast, the influence of cognitive variables on attitude was less relevant.

#### Edited by:

Guendalina Graffigna, Università Cattolica del Sacro Cuore, Italy

#### Reviewed by:

Marco Giovanni Mariani, University of Bologna, Italy CPH Myburgh, University of Johannesburg, South Africa

\*Correspondence: María Pilar Martínez-Ruiz mariapilar.martinez@uclm.es

Specialty section: This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 27 June 2016 Accepted: 12 December 2016 Published: 03 January 2017

#### Citation:

Martínez-Ruiz MP, Izquierdo-Yusta A, Olarte-Pascual C and Reinares-Lara E (2017) Do Affective Variables Make a Difference in Consumers Behavior Toward Mobile Advertising? Front. Psychol. 7:2018. doi: 10.3389/fpsyg.2016.02018 Keywords: antecedents, consequences, attitude, mobile advertising, permission marketing, emotions, feelings

### INTRODUCTION

The high potential that mobile devices offer as a medium for delivering advertising to consumers is based on several key factors. First, companies tend to believe that mobile advertising has a high capacity to reach almost anyone anywhere at any time (Haghirian and Madlberger, 2005; Richard and Meuli, 2013). This belief is supported, among other things, by the high penetration rate of mobile phones among end users, which became quite significant in the 1990s (e.g., Khalifa and Shen, 2008; Zhang and Mao, 2008). Today, it continues to increase worldwide, making it possible to understand the large amounts of time consumers spend on their devices (e.g., Bart et al., 2014).

The use of this type of mobile technology enables relatively more personal and interactive communication with consumers (e.g., Bauer et al., 2005; Sanz-Blas et al., 2015), as well as the development of specific differentiated strategies to target them (Drossos et al., 2007; Sultan et al., 2009; Olarte-Pascual et al., 2014). Moreover, due to the proliferation and growth of localization technologies, mobile advertising makes it possible to send messages to subscribers based on their geographic location. Thus, mobile advertising can range from completely undifferentiated (mass) messages to, where desired, messages tailored to each individual (Richard and Meuli, 2013). This technological resource opens the door to new business opportunities in the field of mobile advertising, offering considerable competitive advantages in terms of customization and of presenting the most relevant information to each consumer (Bauer et al., 2005; Kim and Han, 2014). However, it has also contributed to growing consumer concerns about issues related to the protection of privacy and personal data (Haghirian and Madlberger, 2005). These concerns illustrate the need for permission-based mobile advertising (PBMA), that is, advertising requiring individuals to give their permission before they can receive any type of advertising message (Godin, 1999; Varnali and Toker, 2010).

Indeed, numerous studies have confirmed this fact. For instance, Barwise and Strong (2002) observed that requesting permission in advance influenced the effectiveness of SMS text messaging as an advertising medium for reaching young adults. Likewise, following a comprehensive review of the relevant literature on mobile marketing, Varnali and Toker (2010) found that, among the top six best practices in mobile marketing, requesting permission, and addressing consumers' security and privacy concerns are particularly important. Baek and Morimoto (2012) corroborated this importance, noting that consumers' perception of the behavioral intentions behind text message advertising was influenced by the concern shown for whether their privacy was preserved. In light of all these findings, it is unsurprising that there is some consensus among mobile advertising industry operators that one of the keys to the success of a mobile marketing campaign is that it not be intrusive (Soroa-Koury and Yang, 2010).

In order to provide a solid theoretical basis for examining the adoption of mobile advertising, this paper draws on two schools of thought regarding the nomological structure (Lee, 2009) of the Theory of Reasoned Action (TRA) (Fishbein and Ajzen, 1975): (i) the Technology Acceptance Model (TAM) (Davis, 1989) and (ii) the Theory of Planned Behavior (TPB) (Ajzen, 1991). Since TAM and TPB have been used in many studies to predict and understand user perceptions of systems use and the probability of adopting an online system (Gefen et al., 2003; Wu and Chen, 2005; Hsu et al., 2006), they are the most appropriate tools for understanding mobile advertising adoption. This investigation, similar to others, (Mathieson, 1991; Igbaria et al., 1995; Taylor and Todd, 1995; Lee, 2009), proposes to integrate both models, TAM, and TPB, in order to provide a more comprehensive model of mobile advertising.

The TAM is an adaptation of the TRA by Fishbein and Ajzen (1975) and was developed by Davis (1989) to explain acceptance of information technology for different tasks. This model hypothesizes that system use is directly determined by behavioral intention of use, which is in turn influenced by users' attitudes toward using the system and the perceived usefulness of the system. Attitudes and perceived usefulness are also affected by perceived ease of use. A critical review of TAM has revealed that there is a need to include other components in order to provide a broader view and a better explanation of Information Technology (IT) adoption. As a matter of this fact, since TAM originated in work contexts where emphasis was mainly placed on variables related to job performance, it seems reasonable to give consideration to affective variables that might contribute to acceptance of technologies in more hedonic scenarios, such as the context in which consumers use their mobile devices (Van Der Heijden, 2004; Abad et al., 2010). Moreover, factors related to human and social change processes should be also incorporated. For example, in Information System (IS) literature, the TAM (Davis, 1989), the extended Technology Acceptance Model (TAM2) (Venkatesh and Davis, 2000) and the Unified Theory of Acceptance and Use of Technology (UTAUT) (Venkatesh et al., 2003) are used to explain possible adoption and acceptance patterns of new technologies among consumers. In all these models, concepts like relative advantage, compatibility, complexity, and observability, as well as perceived risk, perceived usefulness, subjective norm, and perceived ease of use play a key role in these approaches.

In the same way, in the context of mobile communications, few studies have taken a holistic approach of the interrelated antecedent factors. Instead, most research has sought to assess only one cognitive dimension. From a joint cognitiveaffective perspective, the theoretical background underlying the assumptions of the model for attitudes toward mobile advertising includes eminently cognitive variables (Karjaluoto et al., 2008; Soroa-Koury and Yang, 2010).

In light of these limitations identified in the literature, the present research aimed to fill this gap by delving deeper into the study of the advisability of treating attitudes toward mobile advertising as a two-dimensional variable including both cognitive and affective factors, and the antecedents and consequences of attitude toward PBMA. More specifically, we have proposed a model to meet the stated aims and to analyze the antecedents of attitudes toward mobile advertising and the relationship between attitude and intention in consumers who receive advertising via their mobile devices.

With these ideas in mind, the remainder of the paper is organized as follows. First, it reviews the relevant literature underlying each of the proposed research hypotheses. Next, it describes the proposed research methodology based on the study of a representative sample in Spain. Finally, it reports the main results and conclusions of the research, placing special emphasis on the implications for businesses and taking into account certain limitations.

### CONCEPTUAL FRAMEWORK

This section will review the relevant literature with a view to selecting the most important antecedents and consequences of attitude toward PBMA. To this end, amongst the antecedents of attitude toward mobile advertising, it will distinguish between cognitive and affective variables and between utilitarian and hedonic ones. This approach is in keeping with Yang et al. (2013), who proposed that taking these classifications into account facilitates a deeper understanding of the possible effects of mobile advertising. It moreover addresses the criticism TAM has received for its excessive dependence on external factors (e.g., perceived usefulness) and exclusion of internal ones (e.g., emotions). Such criticism would seem to indicate that TAM alone is not enough to explain consumers' responses to mobile advertising, but rather affective variables likely to influence attitude formation must be considered too. This is especially true given that TAM originated in work contexts, in which emphasis is placed on variables related to job performance. In contrast, in more hedonic scenarios, such as the context in which consumers use their mobile devices, consideration must also be given to the affective variables that might contribute to acceptance of the technology in question (Van Der Heijden, 2004; Abad et al., 2010). Moreover, while the benefits of considering both cognitive and affective factors in order to better understand peoples' appraisals have been widely recognized in the literature (Van Waterschoot et al., 2008; Levav and McGraw, 2009; Zielke, 2011), it is not yet known how these factors combine to influence attitudes toward mobile advertising and, thus, the intention to receive it. In this context, the first aim of this paper is to fill these gaps. As noted, both cognitive and affective factors influence subjects' appraisals (Dean et al., 2008). Affect and cognition take place through an interlocked dual system that comes together in natural human behavior (Boehner et al., 2007). Moreover, some authors (e.g., Vincent and Harper, 2003; Vincent and Haddon, 2004) have found that, even in some work settings, employees use their mobile phones for their social relationships with partners, family and friends more than with clients.

Given these ideas, it should be noted that the assumptions on which this paper is based are also consistent with the relevant literature in the field of social psychology, which establishes that attitude is influenced by both cognitive and affective variables (e.g., Bagozzi and Burnkant, 1985; Chaiken and Strangor, 1987; Weiss and Cropanzano, 1996).

### Antecedents of Attitude Toward Mobile Advertising

Perceived usefulness is generally understood to refer to the judgment customers make regarding a product's utility based on their perceptions of what they give and what they receive (Zeithaml, 1998). It thus consists of a perceived preference for and evaluation of the product's attributes, attribute performances, and the consequences arising from its use that allow the consumer to achieve his or her goals in different use situations (Woodruff, 1997).

The importance of this concept is evident in the numerous studies conducted in recent years (e.g., Wu and Wang, 2005; Kim et al., 2008; Flavián et al., 2009; Hess et al., 2014; Izquierdo-Yusta et al., 2015; Olarte-Pascual et al., 2016). By way of example, attention should be drawn to the perceived value model proposed by Zeithaml (1998), which served as the inspiration for several subsequent studies highlighting the influence of perceived value on consumers' behavioral intentions (Dodds et al., 1991; Grewal et al., 1998; Sweeney and Soutar, 2001). Among these studies, attention should be called to those carried out based on TAM. TAM clearly assigns fundamental importance to perceived usefulness, proposing it as a key antecedent of attitude toward the use of a given technology.<sup>1</sup> From this perspective, perceived usefulness becomes a very important variable for understanding user behavior in relation to mobile advertising. For instance, in their study using the extended TAM model (TAM2),<sup>2</sup> Venkatesh and Davis (2000) found that perceived usefulness predicted attitude toward mobile advertising, compared to other variables that barely influenced it at all. Kavassalis et al. (2003) observed that when consumers perceived a benefit in receiving advertising messages on their mobile phones, they were more willing to accept such advertising.

In a study analyzing the acceptance of SMS advertising among young people between the ages of 21 and 35, Zhang and Mao (2008) found that perceived usefulness was one of the most important variables for predicting the intention to use that advertising. In the context of permission-based mobile marketing, Karjaluoto et al. (2008) found that the perceived usefulness of mobile communications explained attitude toward mobile advertising. Likewise, in their analysis of a sample of 343 university students, Soroa-Koury and Yang (2010) found that perceived usefulness predicted attitude toward mobile advertising.

In light of the influence that these studies have shown perceived usefulness to have on attitude, the following research hypothesis was proposed:

H1. Perceived usefulness positively and significantly influences attitude toward mobile advertising.

The subjective norm, or influence of reference groups on an individual, is often used as a variable to address the importance of social context with regard to behavior. Its influence on attitudes and behavioral intentions has been widely demonstrated in the academic literature (c.f., Fishbein and Ajzen, 1975; Bearden and Etzel, 1982; Bagozzi, 2000; Pelegrín-Borondo et al., 2016; Versluis and Papies, 2016).

In order to measure the influence of this variable, Fishbein and Ajzen (1975) identified the potential use of two variables: normative belief, referring to what other people want the individual to do, on the one hand, and the motivation to comply with different reference groups on the other. To this end, Peter and Olson (2005) highlighted the influence of two main reference groups: normative ones, such as parents and peers (Fitzgerald and Arndt, 2002), and comparative ones, such as idols (Childers and Rao, 1992).

TAM2 highlights the influence of the subjective norm on the intention to adopt a new technology. From this perspective, Venkatesh and Davis (2000) observed that the subjective norm

<sup>1</sup> In addition to perceived usefulness, TAM tends to include the variable perceived ease of use. However, as the relevant literature has shown that it has a much smaller impact on technology acceptance than perceived usefulness (e.g., Van Der Heijden, 2004), it was not included in this study.

<sup>2</sup>TAM2 differs from the original TAM in that it includes three additional variables: the subjective norm, voluntariness, and image (c.f., Venkatesh and Davis, 2000).

has a stronger direct impact on usage intentions than other variables, such as perceived usefulness. However, drawing on previous studies conducted using TAM2 (e.g., Venkatesh and Davis, 2000), Soroa-Koury and Yang (2010) suggested that social norms are likely to influence the perceived usefulness of mobile advertising. Subsequent studies (e.g., Bauer et al., 2005; Muk and Babin, 2006; Rohm and Sultan, 2006; Lee et al., 2013) have confirmed the positive influence of reference groups on the intention to participate in mobile marketing.

However, few papers have focused on the influence of the subjective norm on attitude toward mobile advertising, considering this attitude to be an antecedent of intention and behavior. In light of the ideas discussed above, the following research hypothesis was proposed, referring to the positive and significant influence of social norms on attitude toward mobile advertising:

H2. The subjective norm positively and significantly influences attitude toward mobile advertising.

In the 1980s, individuals' emotions began to be associated with their decision-making processes and ceased to be considered external elements likely to hinder the optimal functioning of the process (Zajonc, 1980). Moreover, the literature has demonstrated the ability of emotions to stimulate individuals' behavioral intentions (Bagozzi et al., 1999; Zajonc, 2000) and the existence of groups of people who tend to react similarly to certain emotions (Mano, 2004; Pelegrín-Borondo et al., 2016). Thus, in general, objects that cause pleasant sensations are evaluated positively, while those that cause unpleasant sensations are evaluated negatively (Pham, 2007; Bargh, 2013). Likewise, individuals tend to avoid unpleasant situations, engage in activities they find pleasant (Bagozzi et al., 1999; Mano, 2004), and select alternatives that make it easier for them to experience positive, rather than negative, emotions (Bower and Cohen, 2014).

With regard to purchase decision-making processes, Oliver (2014) has shown that when consumers assess the available alternatives and find that the service quality exceeds their expectations, it influences their emotions and pleasure, generating a feeling of delight with the service and positively impacting their intention to repeat their choice. White and Yu (2005) found that positive emotions were positively correlated with the tendency to speak favorably about products, while negative ones encouraged unfavorable communication and the search for available alternatives. Likewise, Forgas and Ciarrochi (2001) found that positive moods enhanced peoples' assessments of products, while negative ones lowered them. Similarly, O'Neill and Lambert (2001) showed that surprise and enjoyment positively affect product evaluations and the act of choosing. Therefore, it can generally be stated that consumers have a natural tendency to make choices that minimize the likelihood of experiencing negative emotions (e.g., Elliott, 1998; Schwarz, 2000; Carstensen and Mikels, 2005).

In the field of online advertising, WAM assumes the existence of three antecedents—entertainment, irritation, and informativeness—as the main determinants of attitudes toward online advertising (Ducoffe, 1996). In this line of research, Tsang et al. (2004) observed that consumers prefer entertaining content over other types of content (such as informative content) when it comes to accepting mobile services. In general, the content and form of advertisements are important predictors of their value, and they are critical to the effectiveness of online advertising (e.g., Ducoffe, 1996; Berger and Milkman, 2012; Teixeira et al., 2012). These findings are consistent with those of other studies carried out decades ago in the sphere of conventional advertising, such as Mitchell and Olson (1981) and Shimp (1981), who observed that interesting and enjoyable advertisements had a positive impact on the brand, or Schlosser et al. (1999), who reported that attitudes toward online advertising were influenced by enjoyment, informativeness, and utility.

However, some studies have shown that certain emotions, such as irritation, negatively impact advertising avoidance by consumers (Ducoffe, 1996; Martí Parreño et al., 2013; Yang et al., 2013). Elliott and Speck (1998) observed that individuals reported negative responses (whether ad avoidance or the development of negative feelings toward the ad) when they were shown cluttered advertisements or when the ads hindered their search for information. More recently, Ünal et al. (2011) corroborated the fact that, if a mobile advertisement is sent with permission, and it is entertaining, informative, reliable, and personalized, it positively affects the creation of positive attitudes toward advertising, although they found some differences in the relationships between attitude, intention, and behavior depending on whether the recipients of the advertising were youths or adults.

Specifically in the field of mobile communication, rather than highlighting the functionality of the communication that mobile devices enable, many studies have stressed the emotional attachment that can potentially be established between the user and his or her mobile device (e.g., Vincent, 2005). From this perspective, Kolsaker and Drakatos (2009) examined the relationship between the strength of the emotional attachment to the mobile device, the perceived benefits of mobile advertisements, and receptiveness toward them, confirming that users with a strong emotional attachment to their mobile devices are more receptive and perceive greater potential benefits in mobile advertising than the rest.

Likewise, emotions are closely related to other affective states, such as feelings. Feelings are the conscious assessment of the perceived body state during an emotional response. In other words, feelings are the results of emotions. Therefore, feelings occur when the brain is aware of the bodily change occurring as a result of a given emotion and are thus subsequent to emotions. Moreover, although emotions are more intense, they are also more short-lived; consequently, the optimal strategy should focus not so much on generating an emotional attachment as on achieving a sentimental one, which is more enduring (Bechara et al., 2000).

Based on the above, the following working hypotheses were proposed:

H3. Positive emotions positively and significantly influence attitude toward mobile advertising.

#### TABLE 1 | Technical details of the research.


H4. Positive emotions positively and significantly influence feelings toward mobile advertising.

H5. Negative emotions negatively and significantly influence attitude toward mobile advertising.

H6. Negative emotions negatively and significantly influence feelings toward mobile advertising.

H7. Feelings positively and significantly influence attitude toward mobile advertising.

### Consequences of Attitude Toward Mobile Advertising

Attitude toward advertising has been extensively studied in recent decades in academia (e.g., Shavitt et al., 1998; Dutta-Bergman, 2006), where it is usually considered to be an antecedent of individuals' behavior and final decisions from the perspective of the aforementioned theories. The empirical evidence obtained has underscored the suitability of using the variable attitude as a determinant of the intentions and behavior of mobile phone users when it comes to accepting advertising. An initial study by Kim and Hunter (1993) on the links between attitude and behavioral intention showed that attitude is positively related to intention.

Subsequently, Lee et al. (2006) observed that favorable attitudes toward mobile advertising, correlated with strong motives, significantly influence intentions and positive actions. Barutçu (2007) found that users have positive attitudes toward certain mobile marketing tools, including advertising. In the context of TAM, Karjaluoto et al. (2008) found that attitude explained a considerable amount of the intention to receive advertising messages from a company, and that this relationship was stronger in women. Based on TAM2, Soroa-Koury and Yang (2010) and Xu (2006) found that attitude toward mobile advertising significantly predicted the intention to adopt mobile advertising. Although most of the studies conducted in this line of research have observed this positive relationship, in some cases, it could not be verified.

Based on these prior studies, in which attitude was generally found to positively influence intention, the following research hypothesis was proposed:

H8. The more positive the attitude toward mobile advertising is, the greater the intention to receive mobile advertising.

Based on the proposed hypotheses, a theoretical model was defined that integrates the various variables influencing attitudes and intentions with regard to mobile advertising in text format (**Figure 1**).

### RESEARCH METHODOLOGY

To carry out the proposed empirical research, a sample was collected from a panel of Spanish adults who receive advertising on their mobile phones in the form of SMS text messages. This was achieved with the technical support of Cint Panel Exchange, as it both enables access to a broad sample of consumers representative of the Spanish market and ensures a high level of quality. The data was obtained in two steps: (1) randomized 9,000 sending invitations with a response rate of 26% and (2) select only individuals receiving mobile advertising by taking into account the structure of the Spanish population by gender and age. The sample was of 612 individuals who were representative of the Spanish adult population and received advertising on their mobile phones. 92.3% of the sample had been using a mobile phone for more than 5 years, and 35.9% had more than one mobile. The most common mobile phone brand was Nokia (40.5%), followed by Samsung (18.8%). The predominant sociodemographic characteristics were: 51.8% were women; 33.8% were between the ages of 35 and 44; 36.6% had a college education; 60.3% were married; and maximum income level was 1201–1800 euros per month. **Table 1** shows the technical details of the research.

Two consecutive rounds of pre-testing were conducted to verify proper comprehension of the questionnaire and the adaptation thereof to the aims of the research. **Table 2** provides a description of the variables included in each construct. All

#### variables were measured by means of a 5-point Likert scale, where 1 was strongly disagree and 5 was strongly agree.

With regard to ethics approval: (1) all participants were given detailed written information about the study and procedure; (2) no data directly or indirectly related to the subjects' health were collected and, thus, the Declaration of Helsinki was not generally mentioned when the subjects were informed; (3) the anonymity of the collected data was ensured at all times; and (4) no permission was obtained from a board or committee voluntary completion of the questionnaire was taken as consent for the data to be used in research.

### RESULTS

The structural model used (**Figure 1**) was validated using the partial least squares (PLS) regression technique. PLS path modeling is frequently used in researches referred to TRA, TAM, TPB (Henseler et al., 2016). Although there is a debate about which technique is more appropriate, if PLS (e.g., Wold, 1982; Lohmöller, 1989), or covariance based structural equation modeling (SEM) (e.g., Jöreskog and Wold, 1982), we decided to use PLS due to the higher benefits it provides in this kind


of researches (e.g., possibility to use with both reflective and formative constructs; small samples).

The model was estimated using the statistical software SmartPLS 2.0, and the significance of the parameters was established by means of bootstrap resampling. To ensure convergent validity, all indicators whose factor loading was not significant or was <7 were eliminated. Thus, the resulting model presented no reliability issues with regard to any of the established criteria (Cronbach's alpha, composite reliability, and average variance extracted) (see **Table 3**).

Discriminant validity was assessed using the average variance extracted for each factor, taking into account that it should be greater than the squared correlation between each factor pair (Fornell and Larcker, 1981), as shown in **Table 4**.

Once the measuring tool's psychometric properties had been evaluated, PLS was used to estimate the structural model


#### TABLE 4 | Discriminant validity.


The off-diagonal numbers are the estimated correlations between the factors. The on-diagonal numbers in bold are the square roots of the average variances extracted.

synthesizing the proposed hypotheses shown in **Figure 1**. The same criterion used to determine the significance of the parameters (612 bootstrap subsamples the same size as the original) was used.

To assess the structural model's predictive ability, the criterion proposed by Falk and Miller (1992) was used, whereby the R <sup>2</sup> of each dependent construct must be >1. Lower values, even if significant, should not be accepted. It is thus possible to determine whether or not the proposed hypotheses are supported, considering the significance of the estimated standardized regression coefficients (see **Table 5**).

The results show that the most important effects were those generated by positive emotions on feelings (β = 0.924; p < 0.01; H4) and by attitudes toward mobile advertising on the intention to receive mobile advertising (β = 0.737; p < 0.01; H8). The positive and significant influences of feelings on attitude (β = 0.353; p < 0.01; H7), the subjective norm on attitude (β = 0.251; p < 0.01; H2), positive emotions on attitude (β = 0.195; p < 0.01; H3) and perceived usefulness on attitude (β = 0.182; p < 0.01; H1) were less important. Finally, negative emotions were found to exert a minor (negative) influence on attitude (β = −0.096; p < 0.01; H5). Therefore, all of the research hypotheses were supported, except H6, referring to the negative influence of negative emotions on feelings.

These results underscore the strong influence positive emotions can have on attitude toward mobile advertising, especially indirectly through feelings. This is a very positive finding for companies that engage in mobile advertising: since feelings last longer than emotions, it is very favorable that positive emotions help to strengthen feelings. Companies should thus try to foster sentimental ties by conveying positive emotions in their mobile advertising messages, as that was found to be the strongest relationship established among the variables. The influence of attitude on the intention to receive mobile advertising was likewise considerable. This finding sheds light on how attitude largely translates to the intention to accept mobile advertising.

The subjective norm and perceived usefulness had less of an influence on attitude. Therefore, the most important variables were those related to affective aspects of the consumers. This finding helps to confirm the aforementioned criticism of TAM, which emphasizes the need to take more affective aspects of consumers into account. Not in vain, to understand these results, it is important to consider the hedonic context in which the present research was carried out.

### CONCLUSIONS

This paper has offered a joint analysis of the antecedents and consequences of attitude toward mobile advertising, based on a context of permission-based mobile marketing, and taking into account the precepts of models such as TAM or WAM. In so doing, it sought to fill an identified gap in the literature with regard to the joint study of these factors.

To this end, the relevant literature was reviewed, in order to identify the most important variables from both


R <sup>2</sup> Feelings = 0.831; R<sup>2</sup> Attitude = 0.810; R<sup>2</sup> Intention = 0.543; \*\*\*p < 0.01; NS = not significant.

a cognitive and affective perspective. Among the cognitive variables conventionally considered to be antecedents of attitude under TAM (or, where applicable, TAM2), perceived usefulness and the subjective norm stood out; among the affective variables, special attention should be called to emotions—both positive and negative—and feelings. Finally, the most important consequence of attitude toward mobile advertising was the intention to receive it. A conceptual model was proposed and tested with a sample of 612 mobile phone users and recipients of text-message advertising.

The findings made it possible to measure attitudes toward mobile advertising, as well as their antecedents and consequences. Specifically, the greatest influence in the model was found to be that exerted by positive emotions on feelings, which refers to the indirect influence that positive emotions can have on attitude toward mobile advertising. This influence was found to be much greater than that exerted by positive emotions directly. Somewhat smaller, but nevertheless very important, was the influence of attitude on the behavioral intention to receive mobile advertising.

Additionally, the influence of cognitive variables on attitude was found to be far less important. In order to understand these results, several factors must be taken into account. First and foremost, as several authors have shown, TAM was originally applied in work contexts; however, the contexts in which end users most often use their mobile devices tend to be hedonic.

These findings clearly point to interesting opportunities for companies that engage in mobile advertising, especially since they underscore the importance of the role of affective variables, such as emotions and feelings, in consumer behavior. They are moreover perfectly consistent with the precepts of the latest approaches and theories in marketing, such as marketing 3.0, which suggests that the discipline of marketing should focus on meeting the full range of individual needs (cognitive, affective, spiritual, etc.). Furthermore, this paper has shown that positive emotions have a much stronger effect than negative ones. Although the latter should be taken into account insofar as they might indirectly influence the intention to receive mobile advertising, companies should focus their resources on emphasizing positive emotions and feelings toward mobile advertising rather than on mitigating the consequences of negative ones. They should moreover consider the strong influence of attitude on the intention to receive mobile advertising.

Therefore, since positive emotions were found to have a greater impact on attitude toward mobile advertising than any other variable, it can be concluded that, in mobile advertising, consumers' affective variables really make a difference. In this regard, the importance of mobile as a medium not only for conveying specific value propositions and offers, but also for carrying out other types of communication actions is clear. For instance, companies may find that mobile can be a very valuable medium when it comes to trying to build and even consolidate a certain brand image. In particular, they might try to achieve this by forging strong emotional bonds with consumers, which can then be transformed into feelings in the long term.

Finally, these findings are consistent with the field of social psychology, which holds that both cognitive and affective variables can influence attitude (e.g., Bagozzi and Burnkant, 1985; Chaiken and Strangor, 1987; Weiss and Cropanzano, 1996). More specifically, they are consistent with the influence of certain positive emotions on attitude toward mobile advertising reported elsewhere (e.g., Tsang et al., 2004; Ünal et al., 2011; Ruiz-Mafé et al., 2014). Tsang et al. (2004) also showed that the perceived irritation, along with the perceived entertainment, information, and credibility of mobile advertising, influenced attitude toward it.

#### REFERENCES


With regard to limitations pointing to new avenues of research, this paper only considered recipients of text-message mobile advertising. It would be interesting to replicate this study considering a larger number of formats and mobile advertising media due to the spread of 3G and 4G technologies (such as advertising on web pages optimized for mobile devices, mobile apps, etc.) and the development of wearables (watches, glasses, etc.). Finally, future research could aim to identify those sociodemographic variables of the target audience that help to explain differences between groups. For instance, does age affect the influence of the subjective norm? Or does the influence of emotions depend on gender?

#### AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication.

#### ACKNOWLEDGMENTS

This work was funded by the Spanish Ministry of Economy and Competitiveness under Research Project ECO2014-59688- R, "National Program for Research, Develop and Innovation Oriented toward Societal Challenges," within the context of the 2013-2016 National Scientific and Technical Research and Innovation Plan.


mobile phone," in Mobile Electronic Commerce: Foundations, Development, and Applications, 395. doi: 10.1201/b17686-26


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Martínez-Ruiz, Izquierdo-Yusta, Olarte-Pascual and Reinares-Lara. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

fpsyg-08-00800 May 23, 2017 Time: 15:38 # 1

# Music Audiences 3.0: Concert-Goers' Psychological Motivations at the Dawn of Virtual Reality

#### Jean-Philippe Charron\*

Departamento de Financiación e Investigación Comercial, Universidad Autónoma de Madrid, Madrid, Spain

Reviewing consumers' motivations to attend performances in a continuously evolving social and technological context is essential because live concerts generate an important and growing share of revenues for the music industry. Evolving fans' preferences and technological innovations constantly alter the way music is distributed and consumed. In a marketing 3.0 era, what consumers do with music is becoming more significant than simply owning or listening to a song. These changes are not only blurring the lines between production and consumption (i.e., co-creation), but also distorting the concept of live attendance altogether. Although mediated performances typically lack presence and authenticity, recent advances in immersive technologies, such as spherical videos and virtual reality goggles, could represent a new form of experiencing live music.

#### Edited by:

Ana I Jiménez-Zarco, Open University of Catalonia, Spain

#### Reviewed by:

Cristina Alcaide Muñoz, Universidad Pública de Navarra, Spain Silvia Sivera, Open University of Catalonia, Spain

#### \*Correspondence:

Jean-Philippe Charron jp.charron@uam.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 31 March 2017 Accepted: 02 May 2017 Published: 23 May 2017

#### Citation:

Charron J-P (2017) Music Audiences 3.0: Concert-Goers' Psychological Motivations at the Dawn of Virtual Reality. Front. Psychol. 8:800. doi: 10.3389/fpsyg.2017.00800 Keywords: virtual reality, concert attendance, marketing 3.0, psychological motivations, live music audience, mediated performance

### INTRODUCTION

Twenty year ago, Kotler and Scheff (1997) recommended cultural managers to give performing arts a marketing orientation, where target audiences' needs and preferences are central to all decision making. It appears that putting music fans at the epicenter of marketing strategies has become especially relevant in this digital day and age. In that sense, the evolution of music consumption models paired with recent technological advances now offers audiences new ways to participate in live music never thought of before, including virtually attending gigs. Therefore, knowledge about music fans' reasons and motivations to attend live concerts will extend our understanding of music audiences, a key aspect of arts marketing (Pitts and Spencer, 2008; Colbert and St-James, 2014).

Contextually, understanding concert-goers' motivations to participate in live performances is increasingly important since attendance represents an essential and growing share of revenues for the music industry and performers alike (Harbi et al., 2014). For example, the Spanish music industry's incomes plummeted because of digital piracy, the economic crisis, and a cultural sales tax of nearly twice the European average. Still, revenues from live music concerts grew a steady 10% over the last years to reach €194.5M in 2015 (APM and SAGE, 2016).

Today's ubiquitous access to digital content and social networks has modified consumers' relation to music in general and live concert participation in particular. This paper intends to review actual and future characteristics of concert attendance under the following triad of factors: concert-going psychological motivations, music-consumption models, and technological innovations.

### WHY GO TO CONCERTS?

fpsyg-08-00800 May 23, 2017 Time: 15:38 # 2

The way cultural goods are produced, distributed, and consumed is constantly evolving based on rapidly spreading technology and shifting consumers' preferences (Harbi et al., 2014). Accordingly, music consumers are moving away from a music acquisitionbased model to an access-based model (Wikström, 2012). Up to recently, the possession of music—even in its digital form stored on electronic devices—was valued by consumers and served as the extension of one's self (Belk, 2013). However, as Spotify, YouTube, and Deezer are becoming the dominant means of mass music consumption (Marshall, 2015), today's fans are not so much interested in "owning" a song, neither physically nor digitally<sup>1</sup> . Instead, music consumers now favor access to a large quantity of online content. As Belk (2013) noted, part of the value of digital goods came from the time and efforts required to obtain them. In that sense, accessing music on streaming platforms is easier than ever. Consequently, under the access-based model, a listener's music playlists may well be more valuable than the actual digital songs (i.e., the dematerialized possession) contained in them, as well as the ability to share such lists (i.e., one's musical tastes) online. These changing consumer preferences are eased by technological innovations including the digitalization of music, but also the pervasiveness of Internet, the multiplication of content-streaming platforms, and the emergence of social networks.

From a psychological standpoint, the digital revolution changed the importance traditionally given to at least two broadly recognized concert-attending motives (Trocchia et al., 2011; Kulczynski et al., 2016). For one, attending a concert to explore discover new artists has become almost irrelevant. Potential concert-goers are better informed than ever about touring bands, concerts dates, and selected venues. They rely on promotional information as well as expert opinions and word-of-mouth. By facilitating information and content, digital technologies encourage participants to try out new performers and reduces the inherent disappointment or financial risks of attending a concert (Burland and Pitts, 2010; Farrugia and Gobatto, 2010).

Conversely, social motives to attend leisure activities generally stem from two necessities: the desire for interpersonal relations and the need for esteem from others (Beard and Ragheb, 1983). The former mostly refers to the sense of community concert-goers experience when attending a live performance. But, as Bennett (2015) suggests, live concerts also complement participation in online communities such as music blogs and social networks. Regarding the later, achieving status and obtaining the respect and admiration of others drive individuals to attend live music performances (Kulczynski et al., 2016), what Holt (2010) refers to as "self-realization through cultural consumption." That is, concert-goers can earn bragging rights by showing off their participation in live concerts. Although "I was at Woodstock" has long been a status-enhancement statement, the phenomenon grew in magnitude alongside the rapid development of virtual social networks (Lingel and Naaman, 2012). Social media now allows concert-goers to share photos, videos, and comments with offsite but online friends, who in turn may reward concert-goers with instant feedback such as "likes" (Scott and Harmon, 2016). Accordingly, sharing attendance to live performances via social networks can help define one's identity by indicating a higher level of authenticity, where concert-going is generally perceived as the "real deal," a more authentic experience than streaming pre-recorded songs, for example (Cresswell-Jones and Bennett, 2015; Shuker, 2016). However, as Lingel and Naaman (2012) suggested, producing online content during a live performance means trading off present (personal) enjoyment for future (social) gains.

### MUSIC AUDIENCES 3.0

Holt (2010) argues that the value of music relates to factors beyond the music itself, and that "musical practices must be analyzed in the perspective of broader social and technological changes." In that context, the music industry will keep moving away from ownership and access-based models toward a contextbased model (Wikström, 2012). Because of recent advances in digital technologies, context-based models offer consumers the necessary tools to experience music rather than just listening to it. These innovations not only blur the boundary between production and consumption of music (i.e., co-creation), but also between live and mediated performances (i.e., liveness).

Under the service-dominant logic of marketing (Vargo and Lusch, 2004, 2008), co-creation implies that engaged consumers participate in creating and giving meaning to products, services, and experiences. In that sense, co-creation is an integral part of the artistic experience, where audiences engage in cognitive, emotional, and imaginal practices to make sense of the performance (Ramsey White et al., 2009; Colbert and St-James, 2014). Live concerts also symbolize a co-creation resulting from the interaction between performers and attendance, and where the end product is the concert experience (Minor et al., 2004; Holt, 2010). Other forms of co-creation also stem from live performances, including bootleg recordings of live concerts later distributed online via social networks (Farrugia and Gobatto, 2010; Lingel and Naaman, 2012).

Technological innovations also distort the frontier between live and mediated performances. Video-streaming platforms and other web-based applications offer users the possibility to attend "live" concerts online. The question of whether mediated artistic performances can procure a similar experience is causing much debate. On the one hand, some argue that digitally mediated concerts impede "the possibilities for the unexpected, iterative, and expansive experience" (Harper, 2015). In other words, even if digital mediation maintains the time dimension (now) of live performances, it ultimately loses its space dimension (here). In that sense, regardless of technological developments, live performances retain some elements of uniqueness that cannot be reproduced, such as being there (Holt, 2010; Harper, 2015).

<sup>1</sup>The music industry also supported the transition from an acquisition-based to an access-based model as a mean to fight back digital piracy and its negative effect on pre-recorded music sales. Paradoxically, illegal file sharing generated positive externalities for the music industry, such as growing audiences (Curien and Moreau, 2009; Mortimer et al., 2012) and surging ticket prices (Krueger, 2005).

On the other hand, fueled by the development of immersive technologies such as spherical videos and VR goggles, virtual concerts are rapidly growing in popularity online. Interestingly, the concept of presence is frequently associated with immersive technologies, where greater levels of immersive quality elicit higher levels of presence, in turn enhancing the effectiveness of a mediated experience (Jung et al., 2015; Cummings and Bailenson, 2016).

Whether live or mediated, the concepts of presence, of being part of something unique and special with likeminded, is manifestly a key component of the experiential nature of concertgoing (Brown and Knox, 2016). While greater engagement, participation, and involvement tends to improve customer experience (Ramsey White et al., 2009; Dobson, 2010; Chen et al., 2011; Kemp and White, 2013), such relationships in the virtual world remain largely unknown.

### CONCLUSION

Music consumption in the digital era is currently receiving a lot of attention. This paper contributes to the review of selected psychological motivations to attend live concerts in continuously evolving social and technological environments. Conceptually, the paper has addressed fans motives to participate in live music within two different Internet eras, often referred to as Marketing 2.0 and 3.0. The first section consisted of a reflection on the shifting importance of intellectual and social motivations

### REFERENCES


to attend concerts. The second section entailed the growing relation between music context and consumer experience. A brief reflection was also initiated on the emergence of virtual reality as novel form of mediated performances, and where immersive technologies can improve virtual participants' perception of being there, an important factor for concert-goers.

As recent technological advances now offer audiences new ways to participate in live music, including virtually attending gigs, knowledge about music fans' reasons and motivations to attend both live and mediated virtual concerts will extend our understanding of music audiences in general. A core question raised in this paper, and one that demands further research, is the extent to which technology can increase participants' engagement and improve the virtual concert experience in the physical absence of others, knowing that social and musical enjoyments often go together (Brown and Knox, 2016). A second question requiring more research is how future developments in immersive technology, although difficult to imagine for the moment, will affect the virtual concert experience and the demand for live performances. Potential changes in consumer preferences are important for the music industry because the economic potential of virtual concerts is almost infinite.

### AUTHOR CONTRIBUTIONS

The author confirms being the sole contributor of this work and approved it for publication.


fpsyg-08-00800 May 23, 2017 Time: 15:38 # 3

fpsyg-08-00800 May 23, 2017 Time: 15:38 # 4


Shuker, R. (2016). Understanding Popular Music Culture. Abingdon: Routledge.


**Conflict of Interest Statement:** The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer SS and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.

Copyright © 2017 Charron. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Differences of Perceived Image Generated through the Web Site: Empirical Evidence Obtained in Spanish Destinations

#### Juan J. Blazquez-Resino<sup>1</sup> \*, Ana I. Muro-Rodriguez <sup>2</sup> and Israel R. Perez-Jimenez <sup>2</sup>

<sup>1</sup> Business Administration Department, University of Castilla-La Mancha, Toledo, Spain, <sup>2</sup> Econometric Department, University of Castilla-La Mancha, Toledo, Spain

In this paper, a study of the perceived destination image created by promotional Web Pages is expounded in an attempt to identify their differences as generators of destination image in the consumers' mind. Specifically, it seeks to analyse whether the web sites of different Spanish regions improve the image that consumers have of the destination, identifying their main dimensions and analysing its effect on satisfaction and intentions of the future behavior of potential visitors. To achieve these objectives and verify the hypotheses, a laboratory experiment was performed, where it was determined what changes are produced in the tourist's previous image after browsing the tourist webs of three different regions. Moreover, it analyses the differences in the effect of the perceived image on satisfaction and potential visitors' future behavioral intentions. The results obtained enable us to identify differences in the composition of the perceived image according to the destination, while confirming the significant effect of different perceived image dimensions regarding satisfaction. The results allow managers to gain a better understanding of the effectiveness of their sites from a consumer perspective as well as suggestions to follow in order to achieve greater efficiency in their communication actions in order to improve the motivation of visitors to go to the destination.

#### Keywords: perceived image, information search behavior, Web pages, destinations, Spain

## INTRODUCTION

Over recent decades, the tourism industry has experienced continuous growth and has become one of the main economic sectors in the world. Different studies have indicated the strategic role of the tourism industry on the nation's economic growth (Sarıı¸sık et al., 2011) and as a principal driving force for new destinations development (Kim et al., 2013). In Spain, as one of the most important tourist destinations in the world, the tourism industry has had a considerable influence on the national economy, which has become a strategic sector.

However, tourism marketers must face an increasingly complex, competitive, and saturated marketplace overall (Echtner and Ritchie, 1993). Recently the effects of the Global Economic Crisis have been added. Although the tourism industry has also proven to be one of the strongest sectors during the Crisis, it has had a fundamental impact on tourist's demand. Therefore, this context creates the need to redesign the management strategies of the destinations (Sirgy and Su, 2000). In general, many destinations focus their strategies on the confidence that visitors will be attracted by

#### Edited by:

Ana I. Jiménez-Zarco, Open University of Catalonia, Spain

#### Reviewed by:

Isabel LLodrà-Riera, Illes Balears Innovacio Tecnologia, Spain Mauro Capestro, University of Salento, Italy

> \*Correspondence: Juan J. Blazquez-Resino juan.blazquez@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 26 September 2016 Accepted: 09 November 2016 Published: 24 November 2016

#### Citation:

Blazquez-Resino JJ, Muro-Rodriguez AI and Perez-Jimenez IR (2016) Differences of Perceived Image Generated through the Web Site: Empirical Evidence Obtained in Spanish Destinations. Front. Psychol. 7:1861. doi: 10.3389/fpsyg.2016.01861 the destination's tangible resources, forgetting that success is not in possession but in its effective deployment (Ritchie and Crouch, 2000). Thus, destinations with limited tangible resources, but managed efficiently, can make the visitor's perception better than destinations with more valuable resources but unable to achieve an effective value proposal (McCartney et al., 2008). Destination marketing organizations (DMOs) make substantial efforts to establish positive destination image because it is important in the process of attracting potential visitors (Fakeye and Crompton, 1991; Sirgy and Su, 2000). The assessment of the destination image can assist managers by identifying the strengths and weaknesses of their destination, helping predict tourists' behavioral intentions (Fakeye and Crompton, 1991; Bigne et al., 2001). In particular, potential visitors with limited knowledge of destinations depend on their perceived image of a destination when it comes to making choices (Um and Crompton, 1992; Beerli and Martín, 2004). The literature clearly shows that the creation and communication of an image for a destination represent a true competitive advantage and an effective strategy for competing in the market (Gallarza et al., 2002; San Martín and Del Bosque, 2008).

To create a favorable perceived image it is essential to analyse how the process of promotional destination information is developed by the visitor. Previous studies about consumer behavior have tried to understand how the operating the processing of information is, what variables can become crucial and which formats, means, or arguments are the most persuasive (Rodríguez-Molina et al., 2015). Girard and Gartner (1993)stated that best way to appreciate a destination is to visit it. However, when consumers have not previously visited the destination, they face a series of conditions of uncertainty whose main cause is the lack of objective criteria for evaluating the destination. To reduce this uncertainty, visitants used different information sources, both internal, and external, trying to acquire as much information as possible to form a perceived image about the destination prior to the visit.

In this sense, the Internet is becoming one of the most important ways of collecting tourism information and creating a perception of the destinations image. The development of the Internet as a means of communication has changed the behavior of external information search by potential visitors to the destination. As information it is meant that, "the Internet constitutes a communication channel that many traditional information sources leverage" (Llodra-Riera et al., 2015, p. 319). From the destinations perspective, the Internet offers numerous advantages over traditional communication sources. A Web page is a dynamic and interactive source of information, rich in content (Pan and Fesenmaier, 2006), which can generate virtual experiences through environmental simulations (e.g., Simulation of real visits). Thus, a destination web page becomes a fundamental instrument in promoting the destination and it leads to a strong feeling of destination image in the visitor (Cho et al., 2002).

Previous research has focused both on understanding how visitors use online communication channels to search for information, as in the way in which the DMOs can be used to improve their promotion strategies (e.g., Pan and Fesenmaier, 2006; Buhalis and Law, 2008; Papathanassis and Knolle, 2011; Ho et al., 2012). However, although a relative abundance of studies have focused on the effect of promotional materials on the destination image (e.g., Gartner, 1989; Sonmez and Sirakaya, 2002), few researchers have focused on the effect of the Internet. Most studies have focused on the analysis of the Internet and on evaluating the performance of websites in terms of content and accessibility, using content analysis of online platforms or analysing user generated content from the experiences of visitors (e.g., Leung et al., 2011; Llodra-Riera et al., 2015; Sun et al., 2015; Tseng et al., 2015). As sources of information, destination websites have reached an important effect on image formation processes (Choi et al., 2007), despite this, few researchers have empirically examined the role of website information on the potential visitant's destination image (e.g., Lepp et al., 2011; Jeong et al., 2012). This study aims to complement the previous research.

The objective of the research was to identify the effect of travel websites on potential tourists' images. Specifically, we seek to analyse whether the information provided by the web pages of different DMOs causes the previous destination image could be significantly different after exposure. In addition, it aims to determine the relationship of the image generated on satisfaction with the tourism destination and on future intentions to recommend or visit. To reach the research objectives the websites of different Spanish tourism destinations have been taken into consideration. Using the findings obtained, we will draw a set of academic and professional implications which enable the development of more efficient online communication strategies.

### LITERATURE REVIEW

#### Information Search Behaviour in Tourism

The way information is processed will influence the consumers purchase decision (Frias et al., 2008). For that reason, for a long time, research in consumer behavior has focused on analysing how the consumer process of information is. In an actual dynamic environment, the need to better understand how consumers acquire knowledge, and search information is important for marketing management decisions and service delivery (Gursoy and McCleary, 2004). Information search is defined as "the motivated activation of knowledge stored in memory or acquisition of information from the environment" (Engel et al., 2001, p. 106).

The main purpose of the information search is to support the process of decision making, reduce risk and uncertainty, and product choice. Hirschman and Wallenoorf (1982), stated that consumers engage in information search for these basic reasons: enhancing the knowledge about product or services and alternatives; and reducing the risk of incorrect choices and future purchase decisions. Risk reduction is considered particularly crucial in non-routinized and extensive decisions regarding acquirement of expensive and complex products, and when people often are strongly involved in decisionmaking processes. In this sense, most researches stated that tourists develop very extensive search of information due to the characteristics of tourism products, it cannot be tried before purchasing. Therefore, to reduce the perceived risk in consumption of unfamiliar tourist products, travelers often use multiple sources of information before making a final decision. Information search does not guarantee satisfaction in consumption experiences but help to reduce visitors perceived risks and therefore, optimize their decisions (Leung et al., 2011).

Fodness and Murray (1998) identify three dimensions for information search in any given purchase situation:


The traveler's search of information is one of the most frequently examined topics by tourism researchers (Schul and Crompton, 1983; Fodness and Murray, 1997, 1998; Vogt and Fesenmaier, 1998; Gursoy and McCleary, 2004; Bargeman and van der Poel, 2006; Hyde, 2008), and all decision-making models include prepurchase hunt for information as key components (e.g., Howard and Sheth, 1969; Schmidt and Spreng, 1996; Engel et al., 2001). For tourism destinations, information search is one of the first steps of the vacation decision-making process and has influence on travel behaviors, such as where to go, how long to stay and how much to spend (Romf et al., 2005).

Whenever a visitor realizes that they need to make a decision, initially an information search takes place internally as the basis for making a vacation decision. Internal sources include previous experiences, with the destination or similar, and the knowledge accumulated through an ongoing search process (Fodness and Murray, 1997; Vogt and Fesenmaier, 1998). However, if internal information proves inadequate or not up-to-date, travelers are likely to use additional information from external sources. In most travel decisions, the search is predominantly external, particularly for new destinations, representing a wide variety of sources of information, and considerable time (Fodness and Murray, 1997).

An important question of practical importance is where tourists obtain external travel-related information. External search consists not only in collecting information from the marketplace but also from a variety of more or less independent or unbiased sources such as news media, guidebooks, and acquaintances. Visitors tend to use a broad combination of external information sources as their search strategies. Different researchers (e.g., Fodness and Murray, 1997; Vogt and Fesenmaier, 1998; Gursoy and Umbreit, 2004) have categorized external information sources as: (1) social, personal, marketing, and editorial; (2) commercial and non-commercial; (3) marketer controlled, reseller information, thirdparty independent organizations, interpersonal sources, and direct inspection; and (4) consumer dominated, marketer dominated, and neutral sources. Travelers rely on both marketing-dominated (mass media, travel brochures, guidebooks) and non-marketingdominated (includes friends, relatives, and personal experiences) sources of information for finding information related to travel and plan the trips.

The visitor's search of information will be as varied and long as the benefits of acquiring information is higher than the costs (Gursoy and McCleary, 2004). Not only monetary costs but also the time spent can influence on the external search. In this sense, the Internet becomes the indispensable channel for people seeking to use tourism information, also in planning and purchasing a travel (Buhalis and Law, 2008). The advantages of Internet as an information source include, first of all, interactivity, but also customized information, low cost, wide coverage, and comprehensive functions (Ho et al., 2012). On the other hand, with the huge amount of information available to travelers, the Internet constitutes an important platform for information exchange between consumer and industry suppliers.

From a consumer behavior perspective, the Internet has gained considerable importance as a communicative means of sharing and disseminating information, different from mass media (television, radio, newspaper, or magazine), becoming one of the main sources of tourist information (Li et al., 2009). The unique characteristics are affecting the consumer behavior (Dholakia and Bagozzi, 2001). The Internet is partly used for practical pre-departure purposes such as travel planning, booking, and payment of tourism products (Hyde, 2008). The Internet offers a rich environment for the information needed by potential travelers who want to gain familiarity with a destination and to locate something of interest to them (Ho et al., 2012).

The research related to the Internet tourism information search has attracted much attention. Earlier studies focused on information search behavior, that utilizes the Internet and what kind of tourism information they are looking for (Pan and Fesenmaier, 2006; Xiang et al., 2008). However, recent researches have focused on the progress of tourism information search behavior caused by the changes of information technology, as social media or blogs, more related with informal information, such as the travel experiences and recommendations of travelers (Xiang and Gretzel, 2010; Sun et al., 2015).

In this way, the Internet constitutes the most powerful communication tool for reaching and attracting more tourists due to its interactive capacity (Beldona and Cai, 2006). For this reason, many tourist destinations have decided to use Web pages as a means of promotion. The Website of a tourist destination is the center from which its attractions spread out to net browsers and other members of the value chain, such as hotels and restaurants. Despite this, given the competition, it is more and more difficult to stand out in this cyber space. Survival depends on different factors related to the design of the Web (Mayordomo, 2003), such as the functionality (content) and usability (ease of use) of the pages. The majority of the Webs of tourist destinations focus on usability and on providing information, since they consider it sufficient to attract potential tourists (Zach et al., 2007). However, a Web page must be a balance between visual and graphic design, business logic, and practical utility. The destination Websites serve as a calling card and they have to be sufficiently attractive for the tourist to decide to visit that destination. The content of destination web pages is especially important because it directly affects the perceived image of the destination, creating virtual experiences that increase the more interactivity is present (Doolin et al., 2002 p. 557). Therefore, the structure and content of the Web must take into account the following questions: what destination image are we seeking to transmit? How can added value be provided to the potential tourist in order to improve the overall image? (Mayordomo, 2003).

#### Perceived Image of Tourism Destination

Perceived destination image has been widely studied in literature. Nowadays, there is a general consensus that the destination image has a key influence on the visitor's travel decision, consumer's satisfaction and destination evaluation (Bigne et al., 2001; Gallarza et al., 2002; Beerli and Martín, 2004; Chen and Tsai, 2007). In general, visitors have a limited knowledge of the destinations which they have yet to visit, thus bestowing an important role on image when it comes to attracting visitors (Huang and Gross, 2010). Destinations with strong, positive and recognizable images will become more probable to be included in the visitor's process of decision-making (Echtner and Ritchie, 1993; Beerli and Martín, 2004). Baloglu and McCleary (1999) postulated that perceived image is a key indicator of destination performance and of the visitor's satisfaction, hence, influencing travel behavior, potential travel intention, and consumption patterns. Thus, the perceived image is one of the most important aspects in the positioning of a destination (Echtner and Ritchie, 1993) since it contributes to creating factors that distinguish it from the competition (Li and Vogelsong, 2006).

Tourism scholars have come up with numerous definitions of destination image, nearly as many definitions of the image as researchers studying it (Gallarza et al., 2002). Although some authors identify it as the perception or set of impressions regarding the place (Phelps, 1986; Fakeye and Crompton, 1991), most consider that the perceived image is the mental representation of a destination (Alhemoud and Armstrong, 1996) or an attitudinal concept consisting on the sum of beliefs, ideas and sensations which individuals hold of a place or destination (Crompton, 1979; Baloglu and McCleary, 1999; Bigne et al., 2001; Li et al., 2009).

Recently, Jani (2016) indicated there is not any accurate definition of perceived destination image due to the presence of different destination image components. However, in general, the destination image is composed of two main dimensions: cognitive and affective (Gartner, 1993; Baloglu and McCleary, 1999; Beerli and Martín, 2004). For instance, Echtner and Ritchie (1993) stated that the destination image vary on a continuum from functional destination attributes, structural elements, or physical characteristics of destination easily observable or measurable, to psychological characteristics, abstract and not easily observed. While cognitive image is based on the perceptions of structural or tangible physical elements, the affective dimension reflects a psychological response, the tourists' emotion or feeling about the destination (Echtner and Ritchie, 1993; Baloglu and McCleary, 1999; Beerli and Martín, 2004). Cognitive components allow us to understand the process of choosing a tourist destination (e.g., Chen and Hsu, 2000). In this sense, MacKay and Fesenmaier (1997:539) claim that "destination image is composed of different products or attractions and attributes which create an overall impression." This means that destination image is determined by the notion that the tourist has regarding the attributes which comprise it and whose presence or absence determines the tourist's perception. Different research has found a cognitive dimension as ideal destination images (Chen and Hsu, 2000; Sonmez and Sirakaya, 2002; Bonn et al., 2005; Barroso et al., 2007). However, affective image also influences on the evaluation of destination image (see Nghiêm-Phú, 2014 for review) and therefore influences the decisionmaking and the desire to visit a destination.

On the other hand, the different categorization of destination images implies that the concept is liable to be defined in various ways using different stages of travel and different sources of travel information (Gartner, 1993; Beerli and Martín, 2004). Thus, the Gunn (1988) stated perceived image can be modified following a sequence: accumulation of mental images of destination, modification of the images through information, and modification of the destination image after experiencing the destination. Then, the perceived destination image includes organic, induced, and complex or modified image (Jeong et al., 2012). The information originates in numerous and diverse sources (Gartner, 1993). Firstly, an organic image is created through general life experiences and non-commercial accumulated information sources such as movies, newspapers, periodicals, books, and personal sources, while induced images are created by commercial travel information sources such as the tourism promotion literature, including magazine articles, guidebooks, Web pages, and TV promotions (Echtner and Ritchie, 1993).

Various researchers have further developed Gunn (1988) concept of image change (Fakeye and Crompton, 1991; Gartner, 1993; Baloglu and McCleary, 1999; Beerli and Martín, 2004; Yuksel and Akgul, 2006). Based on the effect of travel information search, Lin et al. (2007) categorized destination image into baseline image, held before collecting travel information, and enhanced image after having collected travel information. Phelps (1986) categorized destination images into primary and secondary depending on the information sources used. While primary images are formed through internal information such as past experiences, secondary images are influenced by information received from some external sources. Secondary sources fulfill three basic functions in destination choice: to minimize the risk of making a wrong decision, to create destinations image, and to justify a subsequent choice (Mansfeld, 1992). Fakeye and Crompton (1991) used the term "evolution image" to explain the process of image change, adding the additional step, the complex image that is generated from the actual visitation of a destination. In this sense, from organic to induced reflect pre-visit image and complex image represent the post-visit image, after experiencing the destination.

The information sources used by the future visitor are one of the factors widely regarded for its influence on the generation of the pre-visit destination image (Frias et al., 2008). In this sense, it becomes very important to study whether the image projected by the destination promotional materials correspond to those held by visitors (Stabler, 1987). Tasci and Gartner (2007, p.403) claim that to achieve success of a tourist destination is very important to have an adequate image development. Although it is the subjective assessment of external stimuli that forms the image of the destination (Gartner, 1993), it becomes very important to consider what stimuli wants to be present in the DMOs, because sometimes the projected image might not be the same as the received one. In fact, the information transmitted between suppliers, intermediaries and recipients has become more complicated with the arrival of the Internet (Choi et al., 2007).

With the increasing popularity of the Internet, DMOs have used official travel websites as main communication channels (So and Morrison, 2003; Kaplanidou and Vogt, 2006). Due to the multiple dimensions of the destination image and the complexity that the Internet has brought, it is important to examine the provided information to understand the process of image formation in the online context. The intangible characteristics of destination product means that the transmission of the most important images of the destination can only be carried out through their representation in the graphic media and/or audiovisual means. The Internet has great potential to influence on consumers' perceived images it allows to generate virtual experiences (Gretzel et al., 2000) and present a virtual image of the destination that might reduce the perceived risk of a wrong choice.

Previous online information studies have shown two major trends. Some researchers about the travel websites have focused on analysing the operation of the website in terms of accessibility and content (e.g., Bai et al., 2008; Tang and Jang, 2008; Loda et al., 2009; Woodside et al., 2011; Rodríguez-Molina et al., 2015). Their results have shown that website design and Internet marketing features contribute to an effective delivery of messages, quality of products and services, and brand image. Other studies have focused on how the websites influence on the formation of the pre-visit destination image travel intentions. Nevertheless, there are only a few studies (e.g., Frias et al., 2008; Lepp et al., 2011; Jeong et al., 2012; Rodríguez-Molina et al., 2015) and the results are inconclusive. For example, Lepp et al. (2011) found that exposure to the website improves the image and reduces the perceived risk of the destination of Uganda. Jeong et al. (2012) revealed that exposure to a travel website significantly affected most cognitive and overall destination images. Other studies demonstrate that the vast quantity of information available on the Internet leads to information overload and disorientation among its users (Frias et al., 2008; Rodríguez-Molina et al., 2015).

### PROCEDURE

It is considered that the perceived destination image is multiple, dynamic and complex (Gallarza et al., 2002). Previous literature asserts that image-building is related with the unique identity of a destination (Park and Petrick, 2006). Given that not all the destinations possess the same resources nor combine the same characteristics, the final configuration of the image will be conditioned by the destination itself. Furthermore, not all tourists will identify the resources of the destination in the same way. Therefore, as the first objective of the research, we propose to delimit the main aspects which determine destination image. To achieve this objective, we posit the following hypothesis:

#### **H1. Configuration of the perceived image varies according to the tourist destination.**

In literature it is possible to distinguish different main conceptual trends about the effect of information sources on the creation of image (Baloglu and McCleary, 1999; Beerli and Martín, 2004; Huang and Gross, 2010). The cognitive-affective model is the one most used when it comes to studying the relationship between destination image and other constructs of interest. Crompton (1979) was the first author to speak of the cognitive image which reflects the idea that a person has regarding the physical properties of a place. Some authors (Echtner and Ritchie, 1993; Walmsley and Young, 1998) claim that image is not only cognitive (beliefs), but that it also possesses an affective component (feelings), which comprises the overall impression, called total image (Baloglu and McCleary, 1999), which produces the positive or negative evaluation of the destination (Beerli and Martín, 2004).

In this regard, some works have identified the significant influence of information sources on the cognitive image, but not on the affective one (Baloglu and McCleary, 1999). However, the printed information media, such as brochures, are limited when it comes to sending messages which transmit emotions, unlike the online media whose dynamic nature, and interactivity enable them to offer information related to affective aspects more effectively (Kim and Fesenmaier, 2008). In a more recent study, Li et al. (2009) conclude that the affective and overall image change after browsing the Web, while the cognitive one remains stable, greater stimuli being necessary to generate significant changes in the cognitive image. Following the latter observation, it is important to carry out new studies in order to determine the effect of tourist Web Pages on the image of the tourist by studying whether the previous image, both cognitive and affective, differs from that shown after browsing the Web. Regarding this aspect, we posit the following research hypotheses:

#### **H2. Tourist Web Pages modify the tourists' total image regarding to the destination.**

**H2a. Tourist Web Pages modify the tourists' cognitive image regarding the destination. H2b. Tourist Web Pages modify the tourists' affective image regarding the destination.**

On the other hand, destination image not only affects the destination choice but also its assessment and the future behavior of the tourist (Bigne et al., 2001; Chen and Tsai, 2007; Chi and Qu, 2008). Satisfaction is a key factor in the success of tourist consideration, given that it is linked to the choice of destination, the consumption of products and services and also the decision to return to the destination. The previous image significantly determines the expectations of the tourist toward the destination, generated by the information search processes, something which will affect the satisfaction attained by the tourist. The analysis of the effect of image on satisfaction with the destination enables us to identify the key attributes which ensure that the destination can reach or surpass the expectations of the customer and, therefore, ensure a return to the destination. Bearing in mind this approach, and considering tourist satisfaction to reflect the perceived image, in the present paper we consider that the relationship between destination image and the behavioral intention of the tourist is intermediated by satisfaction with the destination after the information search. In consequence, we posit the following hypotheses:

#### **H3. Perceived image positively affects tourist satisfaction with the search for online information.**

#### **H4. Satisfaction with the search for online information regarding a tourist destination positively affects future behavioral intentions.**

To attain the objectives proposed and verify the hypotheses posited, a procedure is designed for collecting the information based on a laboratory study, due to the need of establishing an environment in which the researcher can control other external variables which might affect the sample units. The main features of the research are presented in **Table 1**. To best knowledge, only few studies used experimental design to analysis the effect of internet-based information on the tourist image change (e.g., Lepp et al., 2011; Jeong et al., 2012).

To obtain the necessary information, a personal structured questionnaire is given to each sample unit with instructions for its correct completion. In the first place, those participating must respond to a set of questions which seek to find out the prior image they have of a specific destination. Subsequently, they must work out an information search process for a future trip in the official web page of the same destination, carrying out a totally free browse. Once the search is performed, the participants in the survey must respond again to questions regarding the destination image after exposure to the website information, as well as their satisfaction and future intentions to make a visit. Each group is exposed to one of the tourist web pages included in the study.

To perform the analysis of the differences in the composition of the destination image, it was necessary to select the webs of different tourist destinations. Due to the fact that the final delimitation of the population of the study was national tourists, the choice of destinations is performed by using the data obtained by the survey: Spanish Tourist Movements—Familitur. The choice was based on criteria related to the total number of visits and differences in the tourist products and services offered. The


final choice corresponds to the Autonomous Communities of Castilla-La Mancha, Castilla-Leon and Andalusia.

In the development of the measurements of each of the variables included in the study, we have followed the recommendations established in the literature regarding the determination of the nature of the constructs to avoid committing an incorrect specification (MacKenzie et al., 2005). Most studies about perceived image have used a structured approach, employing Likert scale and/or semantic differential to measure the cognitive and affective dimensions (e.g., Chen and Hsu, 2000). Typically, researchers grouped a set of predetermined image attributes, based on literature review and study context, into several dimensions by data reduction techniques (Bigne et al., 2009). Although Gallarza et al. (2002) showed a list of the most used elements, there is not a battery of image items widely accepted and applicable to different types of destinations.

In the development of the marketing measurements, in general, the "almost automatic acceptance" of the models of latent and reflective constructs has occurred (Diamantopoulos and Winklhofer, 2001: p. 274), which establish that the causality flows from the latent construct toward the observable measurements, which are manifestations of the construct. However, some authors claim that in certain cases the relationship between the latent variable and the observable variables is the opposite (Edwards and Bagozzi, 2000), giving rise to models of formative constructs, which mean that the measurements form the construct. In this respect, the destination image is identified as a formative construct, where the image of each of the attributes of the destination forms the overall image. Currently, other constructs related to attitude are being analyzed using formative construct models (Zabkar et al., 2010; Blazquez-Resino et al., 2015). Even though the literature has paid less attention to the development of formative measurements, some previous studies (MacKenzie et al., 2005; Petter et al., 2007) have enabled the identification of a series of stages to be followed in the development of formative measurements (See **Table 2**).

The final scale is composed of 21 variables which measure the cognitive dimension, grouped into 8 dimensions 6 variables which are grouped into the affective dimension. Respondents must "give the opinion about the following tourism destination characteristics," based on the 29 variables included in the analysis. A seven-point Likert scale (from: Totally Disagree to Totally Agree) was used for the cognitive dimension, while a seven-point Semantic Differential Scale for the affective one.

For the development of the measurements of satisfaction and behavioral intentions the classical model is followed (Churchill, 1979), carrying out a broad review of the literature related to each of the constructs with the aim of recognizing those scales and measurement variables which have been verified and validated in previous studies. Satisfaction is measured by means of three questions ("It's a worthy destination for visiting," "I like as a tourist destination" and "My overall perception of the destination is very good") with which it is sought to estimate the overall destination evaluation obtained from the information search in the web site. Future behavioral intentions are studied by means of the intention to visit the destination in the future ("I will try to go to the destination in the coming years" and "I think that I will visit the destination in the future") and intention to recommend with


two questions ("I will recommend the destination," "Encourage family and friends to visit the destination"). For these variables, a seven-point Likert scale is used, from "Totally Disagree" to "Totally Agree."

### RESULTS AND DISCUSSION

First, the descriptive results show how the information obtained through the website of the destination influences the previous ratings (see **Table 3**). Although in most of the analyzed variables exceeds the value given after browsing the website, for some variables and destinations the effect is negative.

In this sense, for the contrast of the hypothesis, the structural equation modeling method (SEM) is used. It is possible to find two different types of modeling: SEM based on co-variances and SEM based on components, also called Partial Least Squares Path Modelling (PLS PM). Although the first is the most used, PLS PM is an especially useful technique for incorporating formative constructs into the structural model (Petter et al., 2007). Therefore, PLS PM is considered to adapt to the proposal of the current research, performing the analyses by means of the SmartPLS 2.0 statistical pack.

PLS PM, as an SEM analysis model, is determined by two fundamental elements: estimation of the measurement model, where the relationship between the indicators and the latent construct is determined, and an estimation of the structural model, where the relationships between the constructs are evaluated by means of the path coefficients and their significance. Therefore, as a prior step to verifying the hypotheses posited, it is necessary to analyse the psychometric properties of the measurement model, distinguishing between the reflective and formative measurements, given that the estimation and validation framework is different.

To analyse the reflective measurement in PLS, the study of their reliability, convergent validity and discriminant validity is carried out (Gefen and Straub, 2005). The reliability of the constructs is examined by means of Cronbach's alpha and the compound reliability, automatically provided by the SmartPLS 2.0 program. The results obtained in both tests show values which exceed the recommended threshold of 0.7 (Churchill, 1979) and even the stricter one of 0.8 (Nunnally, 1978), which confirms the consistency and reliability of the constructs used. Convergent validity is analyzed by means of the study of the sizes of the factorial loads and the Average Variance Extracted, which indicates the variance captured by a factor with respect to the variance due to the measurement error. In the first case, the results show that the factorial loads of the measurement variables on their respective constructs exceed the minimum value of 0.7 (Chin, 1998), all the values being significant. On the other hand, the convergent validity is adjusted when the value of the Variance Extracted is >0.5, a value which is considerably exceeded given that all the values are close to or >0.8.

Finally, the discriminant validity is analyzed by means of the cross-loading analysis between the indicators and the constructs, where all the loads of the indicators on their latent variable must be greater than the loads on the rest of the factors. Secondly, the Fornell and Larcker (1981) is applied, which establishes that the variance shared between two constructs, measured by the square of their cross loading, must be lower than the Average Variance Extracted of any of the constructs. The results of both tests enable us to confirm the discriminant validity of the reflective constructs.

On the other hand, analysis of the reliability and validity of the formative indicators is carried out by means of the analysis of the weights of the elements on their corresponding formative constructs (Chin, 1998) and their respective significance (Petter et al., 2007). The weights reflect the contribution of the individual indicator on the construct. In PLS, the significance can only be estimated by the re-sampling method using bootstrapping

#### TABLE 3 | Previous vs. after evaluation—means values.


techniques<sup>1</sup> . Following the recommendation of Brown and Chin (2004), 500 subsamples are generated of the same size as the original sample where it is necessary for the loads to be significant to at least 0.05 (Gefen and Straub, 2005).

The results obtained (See **Table 4**) allow us to accept Hypothesis 1, which establishes that the final configuration of the image depends on the specific destination analyzed. In the present study it can be observed how each of the dimensions of the image is determined by a number of different variables according to the tourist destination.

For the verification of the rest of the hypotheses, it is necessary to perform the analysis of the structural model, which will enable an estimation of the path coefficients and their significance. In **Table 4**, we can observe the results of the structural analysis by means of the PLS, showing the structural coefficients, which indicate the strength of the relationships between the variables. For the analysis of the stability and significance of the parameters, which allows us to support the relationships established in the hypotheses, the bootstrapping technique generating 500 subsamples with the same size as the original sample was used. In the same way, the validity of the model is verified by means of analysis of the R 2 value, which measures the predictive power, following the criteria proposed by Falk and Miller (1992) that the R <sup>2</sup> of each of the dependant constructs must exceed the value of 0.1.

The third hypothesis establishes a positive effect of the dimensions which determine the image of a destination through online promotion on tourist satisfaction. From the results obtained (See **Table 5**), it is not possible to reject the hypothesis completely, given that it can be observed that the influence of the image generated on tourist satisfaction is different per the tourist destination analyzed. For the three destinations analyzed, the affective image exercises a significant influence on satisfaction. However, the dimensions of the cognitive image show distinct effects according to the destination. Specifically, in the Castilla-La Mancha Web, satisfaction is significantly influenced by the Cultural, Historical and Artistic dimensions

<sup>1</sup>Bootstrapping (or bootstrap) (Chin, 1998) involves the generation of a certain number of samples with the same size as the original simple, by means of the substitution of values, obtaining the distribution of values for the standard error.

#### TABLE 4 | Analysis of the validity of the image measurements.


\*\*\*p < 0.01; \*\*p < 0.05; \*p < 0.10.

NAT.RES, Natural Resources; INFR, Infrastructures; HOT.SER, Hotel services; LEIS, Leisure; CULT; Culture, history and art; POL.ECO, Political-economic factors; ENVI, Environment; SOCI, Social aspects; AFFE, Affective dimension.

(π = 0.4483, p < 0.01) and Leisure (π = 0.219, p < 0.05). For Andalusia satisfaction is determined by the Natural Resources (π = 0.0994, p < 0.05), Hotel Services (π = 0.0666, p < 0.10) and the Culture, History and Art (π = 0.219, p < 0.05), while the Natural Environment (π = 0.2267, p < 0.05) is the only cognitive dimension which attains a significant value over satisfaction in Castilla-Leon.

With respect to the relationship between satisfaction and future behavioral intentions, the results show a significant effect of 99% (See **Table 6**), which allows us to support Hypothesis 4: i.e., the satisfaction derived from browsing the Webs positively affects behavioral intention.

Finally, in order to determine whether the online information search affects the tourist's previous image, a Parametric Analysis model is developed, which allows us to determine whether significant differences exist in the strength of the relationships which might show the presence of any moderating effect (Tsang, 2002). The test makes use of the path coefficients obtained in the structural model and the standard errors of each sample in order to determine the existence of significance in the differences between said parameters, according to the different groups considered. The moderating effect which belonging to a specific group is analyzed using the t-test. In this regard, the results obtained (See **Table 7**) enable us to observe certain differences between the structural coefficients prior to browsing the tourist Web pages and those determined after the information search. Specifically, the affective image shows significant differences in the case of Castilla-La Mancha and Andalusia, while only the Web page of Castilla-La Mancha displays differences in the dimensions of the cognitive image. This allows us to confirm only partially the hypotheses H2, H2a, and H2b.

### CONCLUSIONS AND MANAGEMENT RECOMMENDATIONS

Information, communication technologies, and tourism comprise the services which will orientate the economy of the twenty first century. These three areas are the key to the


#### TABLE 5 | Structural relationships between image and satisfaction.

\*\*\*p < 0.01; \*\*p < 0.05; \*p < 0.10.

#### TABLE 6 | Structural relationship between satisfaction and behavioral intentions.


\*\*\*p < 0.01.

revitalization and innovation as well as enabling the growth of Autonomous Communities in order to communicate with each other and interact with the environment. Likewise, they are tools that offer strategic opportunities and economic growth. Nevertheless, they also represent new challenges and threats for those agents, which lack the development necessary to adapt to the new advances. Information and communication technologies encourage globalization and the spread of tourism, but in turn, a demand for goods and services is generated which must be satisfied through technological progress.

Specifically, Internet has brought about changes in the direction of traditional marketing activities, given that it constitutes the ideal communication platform for a destination to offer what the different customers want and to communicate with them in different areas. Understanding how tourists acquire knowledge is important for marketing management decisions, designing effective communication campaigns, and service delivery (Srinivasan, 1990). It is during information acquisition that marketers can influence consumers' buying decisions (Schmidt and Spreng, 1996). The Internet facilitates interactive, one-to-one communication, something impossible using more traditional channels. With regard to this, in the present study an analysis has been developed regarding the effect of web pages as a communication instrument on the previous image which tourists' have of a specific destination. The assessment of destination image can assist managers by identifying the strengths and weaknesses of their destination, providing critical insights for managing and developing tourist destinations (Fakeye and Crompton, 1991; Bigne et al., 2001).

In general, the results obtained allow us to identify a set of discrepancies in the influence of promotional tourist web pages on the previous tourist image, recognizing the modifying effect of the type of destination. In the first place, differences have been identified in the final composition of the image according to the web page analyzed. The findings show that tourists determine their image of the destination based on the set of resources and attractions of the destination, which may be different to other similar destinations. These differences also mean unequal influences on the satisfaction of the tourist. Although the affective image shows a significant influence on the three destinations analyzed, only some of the dimensions of the cognitive image attain significant values in their influence on satisfaction. As in previous studies, a significant effect in the relationship between satisfaction and future behavioral intentions has been found for the three webs analyzed, which means that a tourist satisfied with

#### TABLE 7 | Valuation of differences.


the online information search has a greater likelihood of visiting the destination and recommending it to other people.

Finally, the differences have been analyzed between the tourist's previous image and the image subsequent to browsing the web page of the tourist destination. The results obtained show that only certain dimensions of the image, and only for some destinations, will the online information search modify the tourist's previous image. The reasons for the lack of significance in the effect of the web on the image may be various, such as the fact that individuals have an image of the destination so fixed that new information does not give them new data or because the web fails when it comes to faithfully and attractively transmitting the resources and capacities which characterize it. In general, the results obtained enable us to partially confirm the claims of Li et al. (2009), which conclude that the affective image is modified after browsing the Net, with the cognitive image remaining stable. In the present study it has been possible to build on these results and conclude that in the study of the relationship between the information search in the web and the modification of the image the type of destination must be considered, given that the for some destinations the information provided in their webs may mean that the affective image remains invariable, while the cognitive image is modified.

In a recent study, Fernández-Cavia et al. (2016) showed that online communications of tourist destinations are not fully professionalized and standardized, and that DMOs do not use online tools strategically, but tactically. Thus, based on this study and the results obtained, we would emphasize a series of implications which would allow destination managers to achieve better knowledge of the tourists and to develop more efficient communication strategies and promotion. In the first place, to achieve a competitive advantage through the Internet, the destinations have to ensure that their tourist resources are presented in their webs attractively and accurately, focusing on promoting those resources which allow them to improve the destination image and help it to present a better tourist experience.

Given that image depends on the destination itself, it is important to identify the tourist's previous image as well as their needs and desires, in order to develop more efficient marketing strategies. On the other hand, the destinations must be aware of the power of Internet and its potential as a communication tool. Consumers are more and more moving away from the more traditional channels, at the same as they demand a greater control over the sources of information (Vollmer and Precourt, 2008). In the paradigm of the new communications, the nature and power of the social networks must be recognized, given that consumers show greater trust in these sources of social and interactive information (Foux, 2006) and, therefore, they have a greater effect on all aspects of consumer behavior. Thus, a change of attitude in the destination managers is required who have to accept the reality of the transmission of information between consumers and, instead of speaking to them, must learn to converse with their consumers and in this way influence the discussions which take place in virtual spaces.

The research carried out here also presents some limitations. On the one hand, selection of the image measurements has been determined by the inclusion of aspects common to any destination considered. However, each destination possesses a set of particular characteristics and resources which in many cases may have a decisive influence on overall perception. Likewise, the selection of the web pages has followed criteria of variety as regards what the tourist is offered, although it is possible to identify other destinations that possess resources and capacities which generate a different image in the mind of the tourists. These limitations prompt us to recommend future studies such as the application of the model to other domestic destinations, including those resources and attractions typical of the destination. Second, although the relationships posited are supported by the literature, some relationships have not been included, such as the direct effect of the image on behavioral intentions, or variables which may have an influence on the attitude and behavior of the tourist, such as motivation.

REFERENCES


Therefore, we propose the inclusion of new variables and new relationships between the constructs. Finally, the selection of the sample, due to reasons of convenience, is characterized by a spatial limitation, and we would propose, as a future line of research, the enlargement of the field of investigation by using online surveys.

#### ETHICS STATEMENT

As far as ethical protocol, (1) we gave participants detailed and written information about the study and the procedure; (2) we refrained from collecting data related (direct or indirectly) to the subject's health and did not refer to the Declaration of Helsinki when informing the subjects; and (3) we guaranteed the anonymity of the data collected. We did not obtain prior approval from a board or committee. Because the questionnaire was completely voluntary, we took its completion as participants' consent to their data being used in our research.

#### AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication.

#### FUNDING

This work has been funded by The Ministry of Economy and Competitivity (Spain), Resarch Project with reference: ECO2014- 59688-R, Programa Estatal de Investigación, Desarrollo e Innovación Orientada a los Retos de la Sociedad, Plan Estatal de Investigación Científica y Técnica y de Innovación 2013-2016, and Ayudas a Grupos de Investigación de la Universidad de Castilla-La Mancha, Orgánica 01110G6032-GI20163426.


and some recommended solutions. J. Appl. Psychol. 90, 710–730. doi: 10.1037/ 0021-9010.90.4.710


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Blazquez-Resino, Muro-Rodriguez and Perez-Jimenez. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Consumer Expectations of Online Services in the Insurance Industry: An Exploratory Study of Drivers and Outcomes

#### M. Dolores Méndez-Aparicio<sup>1</sup> \*, Alicia Izquierdo-Yusta<sup>1</sup> and Ana I. Jiménez-Zarco<sup>2</sup>

<sup>1</sup> Business Administration, University of Burgos, Burgos, Spain, <sup>2</sup> Faculty of Economics and Business, Open University of Catalonia, Barcelona, Spain

Today, the customer-brand relationship is fundamental to a company's bottom line, especially in the service sector and with services offered via online channels. In order to maximize its effects, organizations need (1) to know which factors influence the formation of an individual's service expectations in an online environment; and (2) to establish the influence of these expectations on customers' likelihood of recommending a service before they have even used it. In accordance with the TAM model (Davis, 1989; Davis et al., 1992), the TRA model (Fishbein and Ajzen, 1975), the extended UTAUT model (Venkatesh et al., 2012), and the approach described by Alloza (2011), this work proposes a theoretical model of the antecedents and consequences of consumer expectations of online services. In order to validate the proposed theoretical model, a sample of individual insurance company customers was analyzed. The results showed, first, the importance of customers' expectations with regard to the intention to recommend the "private area" of the company's website to other customers prior to using it themselves. They also revealed the importance to expectations of the antecedents perceived usefulness, ease of use, frequency of use, reputation, and subjective norm.

#### Edited by:

Gabriela Topa, Universidad Nacional de Educación a Distancia (UNED), Spain

#### Reviewed by:

Juan Jose Blazquez-Resino, Universidad de Castilla-La Mancha, Spain Giovanni Pino, University of Salento, Italy

\*Correspondence:

M. Dolores Méndez-Aparicio mma0174@alu.ubu.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 21 March 2017 Accepted: 10 July 2017 Published: 27 July 2017

#### Citation:

Méndez-Aparicio MD, Izquierdo-Yusta A and Jiménez-Zarco AI (2017) Consumer Expectations of Online Services in the Insurance Industry: An Exploratory Study of Drivers and Outcomes. Front. Psychol. 8:1254. doi: 10.3389/fpsyg.2017.01254 Keywords: expectations, reputation, perceived usefulness, subjective norm, prior recommendation

### INTRODUCTION

Since the start of the current decade, an extensive body of literature has been published on the factors influencing the formation of customer service expectations, in both online and offline environments. This issue is especially important for companies such as insurance companies, which base their innovation strategies on the use of the Internet as a channel for customer relations and dialog (PWC, 2017).

Obviously, what a consumer expects to receive upon using a service is influenced by his or her perception of the brand behind it, as well as by that brand's ability to deliver on its promises (Lassoued and Hobbs, 2015). Likewise, the pressure exerted by a subject's immediate environment influences service expectations (De Leeuw et al., 2015; Bilgihan et al., 2016). Thus, third-party opinions influence the outcomes individuals expect to receive after consuming or using a brand, but they can also affect consumers' feelings about the brand, such as brand love, even in the absence of experience-based antecedents (Roy et al., 2013)

Tornatzky and Fleischer's (1990) methodological proposal added another factor to the model-building process. In addition to the aforementioned factors, these authors held that it was essential to consider other factors inherent to the technology used to support the service offering. This is because the characteristics that subjects perceive of the technological tools used to deliver a service exert a considerable influence on the expected outcome in terms of both: (1) the company's ability to meet their needs (Venkatesh et al., 2003), and (2) the relationship the subject establishes with it (Wang et al., 2016).

Because of the instrumental nature of technology, the research conducted in this field has focused on how technology influences the process of purchasing, obtaining, and using a service (Thakur and Srivastava, 2014). In this regard, Yucel and Gulbahar (2013) and McKechnie et al. (2006) found that the degree to which individuals perceive that using a given technology will make it easier to obtain a service and the degree to which they believe it will be useful for obtaining certain benefits are key factors in the formation of individual expectations.

Analyzing the antecedents of individual expectations is essential to understanding what leads a person to recommend the use of a company's service (López and Sicilia, 2013). From a marketing perspective, recommendation is vital to future success (Hui et al., 2007), as it exerts a considerable influence on the behavior of prospective customers. According to Nielsen (2015), in 92% of purchases, consumers base their decision on the recommendation they receive from friends and acquaintances as opposed to other sources. This is especially important in the online context, where WOM has become one of the most reliable and credible sources (Chari et al., 2016). As noted by Wang and Benbasat (2007), the potential user assumes that the person recommending the brand is trustworthy and familiar with it. The recommendation thus enhances the brand's credibility and new users become more likely to try the product or service (Hsu et al., 2013; Lu et al., 2014; Hanssens et al., 2015).

In this line, a StrongView (2014) report by StrongView and Edison Research revealed two key figures: (1) 77% of people make their purchases based on the recommendation received; and (2) people recommend a brand when it greatly exceeds their expectations. Ladhari et al. (2011) and Meleddu et al. (2015), among others, have argued that recommendation is a behavior typical of the post-purchase and use stage and that the individuals who engage in it are the people with high levels of brand satisfaction, loyalty, and engagement. However, sometimes the recommendation is made early on in the purchasing process, i.e., in the pre-purchase stage (Roy et al., 2013; Ruiz et al., 2014). In this stage, the individual does not yet have experience using or consuming the service; therefore, the only driver for recommendation would seem to be the existence of high and positive expectations for the outcome (Roy et al., 2013).

Knowing what elements influence consumers' expectations, as well as the implications – or consequences – thereof as concerns recommendation, is key in strategic terms for service companies. Among service companies, the insurance industry in particular should be given special attention, due to the important role that insurance plays in the economy by enabling the assumption of risks and mobilizing savings. When it works well, it contributes to economic growth and financial stability. With assets worth two thirds of EU GDP, the EU insurance industry is a significant player in the financial sector. In some countries, such as Spain, the industry accounts for as much as 5.5% of GDP (González and Marques, 2014).

In the insurance industry, knowing the antecedents and consequences of expectations will: (1) make it possible to tailor the service offering to customers' needs; (2) increase credibility and boost consumers' trust in the company; (3) increase customer engagement; and (4) facilitate the attraction of new customers. According to a recent PWC (2017) report, today digitalization is a reality for insurance companies. The challenge, however, is not only to increase digital revenue, but also to consolidate the adoption and dissemination of the Web channel among customers (Nicoletti, 2016). Expectation formation clearly plays a fundamental role in this process, both for customers trying the Web channel for the first time and for those who already have experience using it. The intangibility, inseparability, and heterogeneity of services lead even customers with use experience to have different expectations with regard to the outcome (Qureshi and Bhatt, 2015).

Given the virtual non-existence of literature on this line of research, and drawing on the above ideas, the present paper proposes a theoretical model of the antecedents of individual online insurance service expectations, as well as their consequences in terms of pre-purchase recommendation. To this end, the remainder of this paper is organized as follows. First, a theoretical framework is developed and the hypotheses to be tested are proposed. Next, the fieldwork and results are described. Finally, the implications for business are discussed.

### THEORETICAL FRAMEWORK AND HYPOTHESES

### Individual Expectations

The literature offers various definitions of this concept, following two main proposals, those of Oliver (1980) and Parasuraman et al. (1988). In the early 1980s, Oliver defined the expectation-disconfirmation paradigm, stating that "expectations are consumer-defined probabilities of the occurrence of positive and negative events if the consumer engages in some behavior" (Oliver, 1981). In contrast, the gap-based service-quality model developed by Parasuraman et al. (1988) defines expectations in terms of what customers feel they should be offered. The former has been restated as "predictive expectations" and the latter as "desired expectations" (Yi, 1990).

Today, conceptualizations of expectations generally fall within one or the other line of thinking. However, the lines share certain features: (1) they recognize that expectations are the result of an individual, and (2) they entail awaiting a given result. In this regard, Zeithaml et al. (1993) found that consumer expectations are pretrial beliefs about a given product that serve as a standard or reference point against which to judge its performance. Spreng et al. (1996) defined them as "beliefs about a product's attributes or performance at some time in the future." In contrast, for Wu et al. (2014) they are "generalized beliefs that individuals have

about a social object." Finally, Kujala et al. (2017) have noted that expectations determine what an individual expects to receive from a service and that they, in turn, are conditioned by the desires and level of abstraction the individual achieves during the evaluation process.

In the online environment, the above definitions are fully valid. However, in addition to considering the psychological nature of expectations, the influence of technology on the quality, performance, or experiences individuals receive both while consuming the service and during the purchasing process must also be considered (De Keyser and Lariviere, 2014).

Traditionally, expectation research in the field of services has consistently referred to quality, since customers' main objective is for the service they intend to purchase to meet certain quality standards (Parasuraman et al., 1988; Alén and Rodríguez, 2004). With online services, technology is key to quality; therefore, characteristics such as ease of use, the quantity and timeliness of the information provided, or the speed, precision and security with which it enables completion of the process, among others, have all been identified as dimensions of service quality (Zavareh et al., 2012). However, expectations may also be related to how or what an individual might feel when consuming the service (Koenig-Lewis and Palmer, 2014). While many of these feelings and emotions develop from the purchasing and consumption process itself, others are the result of the characteristics of the environment in which these processes take place (Nakamura and Csikszentmihalyi, 2014).

In this regard, technology once again plays an essential role, since it can generate flow experiences in subjects that leave them both deeply involved in an enjoyable activity and emotionally absorbed (Ozkara et al., 2016). However, although technology influences the expectation formation process, it is the individual who ultimately carries the process out. Hence, internal factors related to how individuals perceive the brand, their experience with it, their personal characteristics, and the third-party recommendations they receive directly influence the expected future service outcome (Oliver, 1977, 1980; Cadotte et al., 1987; Oliver and Burke, 1999; Andreassen, 2000; Torres Moraga, 2010; Duque-Oliva and Mercado-Barboza, 2011; all cited in Pelegrín-Borondo et al., 2016).

### Characteristics of the Technology: Service Adoption Models

Understanding what drives an individual to use technology has been one of the chief concerns among management scholars and professionals. A great deal of the literature has focused on comparing the predictive capacity of the different theories on technology adoption and use. Many of these studies, such as Gounaris and Koritos's (2008) study on the adoption of online banking, have shown that incorporating concepts from various theoretical frameworks increases a model's explanatory power.

Most of the theoretical developments have been based on the Theory of Reasoned Action (TRA), formulated by Ajzen and Fishbein between 1975 and 1980. This theory holds that individual behavior depends on intentions, which, in turn, depend on attitude and social pressure (subjective norm) to engage in the behavior (Montano and Kasprzyk, 2015). Thus, any other factor that might influence behavior does so only indirectly. In the 1980s, the TRA was widely used. Its great explanatory power and consistency with other studies made it suitable to predict a broad set of behaviors (Davis et al., 1989). Indeed, such was the TRA model's importance that it was the basis for two subsequent lines of work: the Theory of Planned Behavior (TPB) and the Technology Acceptance Model (TAM). Both lines corroborate the role of individual intentions as triggers of individual behavior.

Sampedro et al. (2014) have argued that the TPB offers a general model that explains individual conduct based on the beliefs-attitude-intention-behavior relationship. The intention to act is considered the best indicator of the behavior, since it is indicative of the effort the individual is willing to make to perform a given action. The model is completed with the inclusion of three exogenous variables that explain the behavioral intention: attitude toward the behavior, subjective norm, and perceived control in the behavior.

Meanwhile, the TAM is perhaps the model to receive the most attention in academia. Developed by Davis (1989) and Davis et al. (1992), the TAM is an adaptation of the TRA model that focuses on the behavior of new technology use.

Analytically simplifying the previous models, the TAM places special emphasis on analyzing the factors affecting individual attitudes and intentions. Hence, it proposes that the decision to use a technology is based on its degree of functionality and the characteristics of the interface (Yucel and Gulbahar, 2013). In particular, the TAM predicts that the use of ICT is conditioned by two specific individual beliefs about technology: (1) perceived usefulness (PU), and (2) perceived ease of use (PEOU).

Perceived usefulness was first introduced as a factor in the TAM model. Subsequently, Venkatesh et al. (2003) included it as part of the concept of performance expectancy. Davis (1989) defined it as "the degree to which a person believes that using a particular system would enhance his or her job performance." Individuals thus generate this perception by assessing whether a new system offering additional features would enhance their performance compared to that achieved with a system previously used to carry out the same function or even before using any system at all. Perceived usefulness can thus be considered extrinsic to the technology itself, related instead to efficiency and effectiveness in the performance of an activity and to expectation formation.

#### H1. The perceived usefulness of a technological application positively influences a user's service outcome expectations.

Perceived ease of use is also part of the TAM and refers to "the degree to which a person believes that using a particular system would be free of effort" (Davis, 1989). This concept was introduced in the Innovation Diffusion Theory (IDT), but in the opposite sense, in terms of complexity (Rogers, 1962; Moore and Benbasat, 1996). Subsequently, Venkatesh et al. (2003) conceptualized the idea as the construct effort expectancy.

This factor is relevant to the adoption of a system, because systems that are easy to use require a smaller effort on the part of

the user, thereby allowing him or her to allocate more resources to other activities, since effort is a finite resource (Radner and Rothschild, 1975). In this regard, Davis (1989) holds that a system that is easy to use will generate a more positive attitude in the user toward using it. Furthermore, like perceived usefulness, PEOU is subjective in nature and, thus, can vary from one individual to the next in relation to the same system. The perception will be different depending on the user's knowledge or prior experience with similar systems. Finally, PEOU can also vary over time. As users become more proficient in the use of a system, they may begin to perceive it as increasingly easy to use (Martins et al., 2014). In light of these considerations, the following hypothesis was formulated:

#### H2. The PEOU of a technological application has a dual effect on the user's service outcome expectations.

In general, the TAM is the most widely applied theoretical system in the field of information systems. Because of this widespread use, it is considered a well-established and robust theory (Svendsen et al., 2013; Yucel and Gulbahar, 2013).

However, over time, certain aspects of the TAM, related to its definition, design, and direction of implementation, have proven limited. Baron and Ensley (2006) identified important problems in the model, related to the definition of its key constructs. Similarly, Venkatesh et al. (2003, 2012) and Bagozzi (2007) have highlighted the need to increase the model's explanatory power by including additional variables and even the need for a paradigm shift in this regard. In other fields, authors such as Goh et al. (2016) have also underscored the need to increase the model's explanatory power through the inclusion of additional variables.

### Characteristics of the Service User

In addition to the characteristics of the technology used to support the service, online service expectations also depend on the individual, on his or her perception of and relationship with the brand, and on the influence exerted on him or her by the environment.

#### The Influence of External Recommendations

One of the main contributions of the TRA (Fishbein and Ajzen, 1975) lies in its recognition of the influence of social environment on individual behavior. Referring to the subjective norm, these authors asserted that individual behavior is strongly shaped by the influence on the individual of certain important groups or actors.

The most common conceptualization of this phenomenon in the literature on the influence of social factors is that of the subjective norm imposed by reference groups. The subjective norm is an external driver related to the beliefs of others – people whom the individual regards as important – who provide the individual with information that he or she considers credible and relevant and, thus, ultimately strongly influence his or her behavior (Venkatesh et al., 2012). Bagozzi and Dholakia (2002) noted that individuals incorporate the opinions of a referent – an individual or group – that is important to them as part of their own belief structure, making them their own. Thus, third-party opinions affect the attitudes, motivations, and expectations the individual shows with regard to performing a behavior (Rivis and Sheeran, 2003; Hsu and Lu, 2004).

For the subjective norm to be positive, the individual has to perceive that his or her reference group approves of the behavior (Bhatti, 2007). The basis for this norm is twofold: individual motivation and the beliefs of the people the individual regards as referents. The motivation to perform the behavior lies in the individual's need to be accepted by the influencing group, while beliefs refer to the individual's conviction that the group's opinion is the best and most favorable for him or her (Küster and Hernández, 2013).

Other times, the influence is exerted by some other agent or institution, such as the manufacturer, retailer, or brand. With brands, the influence exerted on an individual's motivations, expectations, and behavior is determined by the credibility this source has for the individual (Westerman et al., 2014). Perceived source credibility has been defined as "judgments made by a perceiver concerning the believability of a communicator" (O'Keefe, 1990). While the precise factor structure of source credibility is still being debated (Cronkhite and Liska, 1976), one of the most common factor structures includes three dimensions: expertise or competence (i.e., the degree to which the perceiver believes the source knows the truth), trustworthiness (i.e., the degree to which the perceiver believes the source can be trusted to tell the truth to the best of his or her knowledge), and goodwill (i.e., the degree to which the perceiver believes the source has the perceiver's best interests at heart.)

The subjective norm, whether exerted by friends, family members, and peers or by any other agent, plays a decisive role in the performance of certain social behaviors, including purchasing and consumption behavior. Social influence is likewise decisive in the purchase of certain products or brands and in their use in certain contexts or situations (López de Ayala López et al., 2015).

Nevertheless, some studies do not include the social norm as an antecedent of individual expectations and behavior, citing the lack of empirical evidence supporting the relationship between social influence and behavioral intention (Davis, 1989). Mathieson (1991) added that this null influence of the reference groups could be due to methodological aspects, such as the choice of the behavior to be studied, the sample composition, or the scope of the research. Based on the above ideas, the following hypothesis was formulated:

H3. External recommendations positively influence users' service outcome expectations.

#### Influence of Corporate Reputation

The literature offers various definitions of this concept. Alloza et al. (2013) define corporate reputation as the set of collective evaluations that a company's behavior evokes in different audiences, motivating their conducts of support or opposition. Similarly, Gotsi and Wilson (2001) and Schwaiger et al. (2009) argue that reputation is a general evaluation of a company over time. According to these authors, this evaluation is made based on: (a) the stakeholder's direct experience with the company, or (b) any form of symbolism that provides information about the

company's actions and/or offers a comparison with the actions of its leading rivals.

Despite the diversity of definitions, all share certain common features. The first is that reputation has a temporal dimension, which distinguishes it from image or identity (Cravens and Oliver, 2006). Additionally, reputation arises as a result of interactions between stakeholders and the organization over time (Argenti and Druckenmiller, 2004). Consequently, an organization does not have just one reputation, but rather as many reputations as there are groups with which it interacts. Finally, reputation is the result of the evaluation individuals make regarding the behavior of the organization or brand.

Helm (2007) has noted that there is considerable agreement on the positive effects of having a good reputation. Gotsi and Wilson (2001) highlighted the relationship between reputation and individuals' perceptions of an organization, although here it must be added that the evaluation made directly influences the individuals' future expectations of the company or brand's behavior (Walker, 2010). A good reputation leads to high expectations, as people consider there is no dissonance between what the company promises and what it delivers (Walsh et al., 2016). Fombrun (2011) argued that reputation is the basis for trust. This feeling is of enormous importance, as it is the trigger for favorable attitudes and behaviors toward a brand or organization. A good reputation is based on good work, on fulfilling the promises an organization makes in response to its stakeholders' expectations. Accordingly, Alloza (2011) has suggested that a good reputation generates confidence, while at the same time increasing stakeholder retention and satisfaction. Roberts and Dowling (2002) indicate that a good reputation generates positive outcomes for an organization, influencing its survival and financial performance. According to Firestein (2006), reputation is the strongest determinant of an organization's sustainability. While strategies can always be changed, once a reputation has been seriously harmed, it is difficult for an organization to recover. Based on the above ideas, the following hypothesis was formulated:

H4. Reputation positively influences the user's service outcome expectations.

#### Use Experience

Prior experience using and consuming a product has traditionally been considered one of the most important variables moderating the variation in attitudes toward a brand or company (Lee and Ma, 2012). This experience is the result of direct or indirect contact with a company. Direct contact generally occurs during the purchasing process, or during the use of a product or service, and is usually initiated by the customer. In contrast, indirect contact usually consists of unplanned encounters with representations of a company's products, services, or brands in the form of recommendations or verbal critiques by other customers, advertising, news reports, reviews, etc. As a result, individuals acquire knowledge of the brand and its attributes, leaving them in a position to compare the outcomes obtained with their initial expectations.

In the early 1990s, Johnson and Fornell (1991) and Zeithaml et al. (1993) found that the consumption experience is the most important factor for the formation of future expectations. These expectations then become the basis for evaluating the service outcome and making judgments about satisfaction (Oliver, 1980) and quality (Parasuraman et al., 1991). Additionally, Karapanos (2013) found that prior experience is the most important source provided by a product for acquiring knowledge about the brand.

Experience accumulates over time. The more experience an individual has with a brand or product, the greater his or her degree of familiarity with it becomes. Consequently, the product's perceived quality is likely to change over time, as is the relative importance of certain qualities. For instance, while learnability and novelty may play a vital role at the start, other aspects, such as usefulness or social capital, may ultimately drive prolonged use (Hassenzahl, 2004). One would thus expect different levels of experience to lead to changes in consumer attitudes. However, the literature reveals numerous contradictions in this regard. For example, the theory of assimilation (Hovland et al., 1957) and the theory of contrast (Singer et al., 1972) offer a different perspective on the effect of expectations on satisfaction. In light of these considerations, the following hypothesis was formulated:

H5. Prior experience with the service positively influences the user's service outcome expectations.

### The Effects of Expectations: Recommendation

One of companies' main objectives is to get customers to recommend the use of their service. In the marketing area, recommending means suggesting or advising that a certain product, service, or brand is the best in its category to meet a given need (Skålén et al., 2015). Research on marketing has highlighted the key role that expectations play in service recommendation (Parasuraman et al., 1988). Consumers define their expectations based on different service characteristics and outcomes (Oliver, 1980; Alén and Rodríguez, 2004). At the same time, individuals will recommend a product or service: (1) when, based on their use experience, they are able to confirm that the product or service has met their expectations (Gupta and Harris, 2010); or (2) when, despite not having use experience, they have a high degree of trust in the brand or company, and, therefore, in its ability to fulfill its promises with regard to the product or service, based on which they formed their expectations (Esteban et al., 2014).

In the first case, consumers make the recommendation once they have used the service, upon determining that the outcome was satisfactory (Flavian et al., 2014). In other words, once they have used the service, consumers' own experience allows them to evaluate whether the final outcome exceeds, falls short of, or matches their initial expectations. In online environments, confirmation of customers' prior expectations positively influences their satisfaction level (Kim, 2012). Thus, prior customers form their expectations based on their perception of certain characteristics of the seller, such as reputation, size, the information available in the media, etc. Customer satisfaction is the result of a post-purchase and

post-service-use evaluation and comparison process that affects customers' intention both to recommend the service to other customers and to re-use the company's services themselves (i.e., user loyalty) (Yi, 1990; Anderson et al., 1994). Consequently, customers' willingness to recommend a service is often used in marketing literature to analyze the relationship between customer satisfaction and customer loyalty (Rodríguez del Bosque et al., 2006; Lin and Lekhawipat, 2014).

Notwithstanding the above, with online services, each use situation can be considered a new experience (Blut et al., 2014). The evaluation of the outcome and its comparison with prior expectations are complex. The intangible, inseparable, and heterogeneous nature of services means that each use situation can result in a different outcome. Consequently, even if prior experience with a service influences expectations, there may still be a certain degree of uncertainty with regard to the expected outcome.

According to Chaudhuri and Holbrook (2001), brand loyalty has been defined as a deeply rooted emotional commitment to the brand that leads the customer to consistently engage in certain repetitive behaviors in the future, such as: (a) rebuying or repatronizing a preferred product or service, or (b) recommending it to others. These behaviors result in repeated recommendation or purchasing of the same brand despite situational influences and marketing efforts that could potentially cause the customer to switch to a different brand.

As this description shows, even in the face of negative external influences, the customer still feels compelled to repurchase, recommend, and commit to the brand over others. The idea of brand loyalty assumes: first, the existence of an actor with free will; second, that this actor is focusing his or her free will on an object; and, third, that brand loyalty is built over time (Pearson, 2016). This long-term brandloyalty relationship can moreover be developed in two different dimensions: a behavioral dimension and an attitudinal one (Söderlund, 2006). Bandyopadhyay and Martell (2007) and Veloutsou (2015) have discussed whether both dimensions are needed to achieve a higher level of brand loyalty, known as "true brand loyalty".

The behavioral dimension of brand loyalty is identified based on customers' observable and repetitive behavioral patterns at a given time. Such patterns would include re-purchasing a preferred item or recommending the brand to others. In contrast, the attitudinal dimension consists of customers' non-observable patterns and is thus built on attitudes, intentions, and the strength of the customer's relationship with the brand (Dick and Basu, 1994; Bandyopadhyay and Martell, 2007).

The online context has underscored the strategic importance of recommendation. Marketing literature recognizes recommendation as a direct effect of satisfaction, such that satisfaction increases the likelihood that consumers will convey positive information about brands or products. Known as word-of-mouth or WOM, this post-consumption behavior is considered to positively affect both the brand's reputation and prospective customers' purchasing decisions (López and Sicilia, 2013). Satisfied customers whose prior expectations are confirmed when they use a company's online process are more likely to recommend it, making them (the customers) promoters of the company. In an online context, the effect of WOM is further amplified, as a result of the Internet's global reach (Yoo et al., 2013; López and Sicilia, 2014).

In the second case, despite not having experience using the service (Roy et al., 2013; Esteban et al., 2014), the customer exhibits a high degree of trust in the brand or organization and, thus, in its ability to fulfill its promises regarding the service on which the customer's expectations are based (Esteban et al., 2014). In other words, the lack of use experience suggests that it is individuals' brand expectations that lead them to make the recommendation.

Wilson et al. (2017) found that some individuals develop a strong emotional link with the brand, known as "brand love," which increases the levels of brand loyalty and positive recommendation. Similarly, Esteban et al. (2014) noted that the personal link between consumer and brand has become a strategic issue for companies, as its main consequences include, among other things: brand trust (Albert and Merunka, 2013); willingness to pay a brand premium (Thomson et al., 2005); and willingness to forgive brand failures (Batra et al., 2012).

According to Esteban et al. (2014), brand love can be defined as an intimate and emotional tie between an individual and a brand, characterized by a set of different beliefs (e.g., personal integration with the brand), feelings (emotional connection, affiliation), and behaviors (desire to use it, willingness to recommend it, etc.).

Roy et al. (2013) showed that brand love can develop either based on experience using the brand – or the online service offered – or through managed (firm-sponsored) or unmanaged (word-of-mouth) communication about the brand. The first kind of antecedent is related to a satisfying brand experience over time, while the second is related to the non-experiential antecedent of brand love. In this situation, the origin of this emotional bond lies in the brand's reputation and the external influence – subjective norm – exerted on the individual by his or her environment. Thus, consumer beliefs about the brand and the services it can offer, as well as such external communication, act as cues for consumers, leading them to develop certain perceptions and attach meaning to brands even when they have no direct experience with them.

Bergkvist and Bech-Larsen (2010) identified two other factors that can promote brand love based on experience using a brand – or online service – namely, consumer-brand identification and the sense of community of the brand's users. With regard to brand identification, Ahuvia (2005) found that love objects are central to how people identify. More recently, Belk (2013) and Sheth and Solomon (2014) showed that this feeling is also found in the digital world. Similarly, Bagozzi and Dholakia (2006) proposed the concept of brand identification as the "extent to which the consumer sees his or her self-image as overlapping the brand's image." Bergkvist and Bech-Larsen (2010) showed that identification means "self-image congruence" and "selfconnection" and also noted that previous studies, such as Fournier (1998) and Kressmann et al. (2006), had found a positive relationship between brand identification and brand passion and love.

The desire of be part of a community – sense of community – can also promote love of a brand and its services even among consumers with no experience using it. In this regard, Bagozzi and Dholakia (2006) applied the concept of social identity in the context of brand community. In particular, they noted that social identity is positively related to brand identification, because increased identification with the brand community leads to greater engagement with the brand, which, in turn, leads to an assimilation of the brand's identity into one's own. Bergkvist and Bech-Larsen (2010) defined "sense of community" as the kinship or affiliation a consumer feels with other people associated with the brand and showed that sometimes this feeling becomes a need for the individual, thereby strengthening his or her bond with the brand, as well as his or her desire to belong to the community.

All of these elements increase the individual's expectations with regard to the service's expected outcome. Similarly, Hegner et al. (2017) found that these factors are conducive to this behavior, insofar as they raise the individual's outcome expectations in terms of both the Web channel and the service. In light of these considerations, the following hypothesis was proposed:

H6. The user's service outcome expectations (positive/negative) influence recommendation (positive/negative) of the service.

#### RESEARCH METHODOLOGY

In order to ensure homogeneous data and enable the comparison of user/consumer assessments, a specific service was chosen, namely, an insurance company. This choice made it possible to compare the results obtained by variable (e.g., reputation or expectations). To this end, an insurance company was selected that provided coverage throughout Spain and had a high number of insured. The selected insurance company ranks fourth in terms of market share (7.44%) out of a total of 100 insurance companies operating in Spain (Investigación Cooperativa entre Entidades Aseguradoras y Fondos de Pensiones [ICEA], 2016). Of the entire insured population, those individuals listed as registered in the private area for clients were chosen. This condition was predetermined by the research objective. Specifically, the target customers were those customers who had used the company's online services in the last month (23,223 customers in total). Once the users/customers had been selected, they were e-mailed an invitation to access a questionnaire. Access was voluntary, but as an incentive to complete the questionnaire, they were offered a small reward: a discount coupon for the purchase of fuel. The technical details of the research are shown in **Table 1**.

Between June 9 and 26, 3 waves of e-mails were sent out to these users with a link to the online survey. A total of 4,178 surveys were completed. To obtain the data, a structured questionnaire was used, with closed, single-response questions. By age group, 30% of the respondents were between the ages of 36 and 45; 25% were between the ages of 26 and 35; 25% were between the ages of 46 and 55; 14% were between the ages of 56 and 65, 4% were over the age of 65; and only 2% were TABLE 1 | Technical details of the research.


under the age of 25. Of the total number of respondents, 85% had prior experience using the private area for clients; of these, the experience of 81.6% was with this company.

To test the proposed hypotheses, scales used in previous studies were selected based on the literature review. Ten-point Likert scales were used to measure the variables, as they are the most suitable for this type of research. Specifically: (a) Likert scales are the most suitable for measuring individual attitudes and perceptions (Likert, 1932; Krieg, 1999); (b) they are suitable for obtaining information through the design of online surveys (Mathieson and Doane, 2005); and (c) they are suitable for use in causal models (Krieg, 1999). The scales used and the description and codification of the items are shown in Annex I.

The structural model used (**Figure 1**) was validated using the partial least squares (PLS) regression technique. The model was estimated using SmartPLS 2.0 software, and the significance of the parameters was established through bootstrapping with 4,178 subsamples the same size as the original. To ensure convergent validity, all indicators with a factor loading that was not significant or was less than 0.7 were eliminated. The resulting model thus had no reliability problems, according to the established criteria (Cronbach's alpha, composite reliability, average variance extracted) (**Table 2**). As can be seen in **Table 2**, two of the scales had a Cronbach's alpha less than 0.7. In this regard, Loewenthal (1996) suggested that a reliability score of 0.6 can be considered acceptable for scales with fewer than 10 items. Likewise, Nunnally (1978), Cronbach and Shavelson (2004), Huh et al. (2006), and Malhotra (2008) have suggested that scores greater than 0.6 can be considered acceptable.

To assess the discriminant validity, the average variance extracted for each factor was used, taking into account that it should be greater than the square of the correlation between each factor pair (Fornell and Larcker, 1981), as indicated in **Table 3**.

Once the measurement instrument's psychometric properties had been evaluated, the structural model synthesizing the proposed hypotheses (**Figure 1**) was estimated using PLS, and the same criteria were used to determine the significance of the parameters (bootstrapping with 4,178 subsamples, the same size as the original sample) (Efron, 1982; Efron and Gong, 1983; Efron and Tibshirani, 1993; Chin, 1998a,b, p. 320; Streukens and Leroi-Werelds, 2016).

In order to evaluate the structural model's predictive capacity, the criterion proposed by Falk and Miller (1992) was used, whereby the R<sup>2</sup> of each dependent construct must be greater than 0.1. Lower values, even if significant, should not be accepted. This made it possible to determine whether there was support for the proposed hypotheses based on the significance of the standardized estimated regression coefficients (**Table 4**).

#### RESULTS

The results obtained in the model showed, first, that there was support for all the proposed hypotheses. Second, they showed that the most important effects were the effects of expectations on the intention to recommend using the company's website to perform transactions and process claims (β = 0.384; p < 0.01; H6), followed by the influence of perceived usefulness on expectations (β = 0.293; p < 0.01; H1). The effects of reputation and PEOU on expectations came third, with nearly identical values (β = 0.197, p < 0.01 for H4 vs. β = 0.184, p < 0.01 for H2). Last but not least was the influence of the subjective norm (β = 0.053; p < 0.01; H3) and of frequency of use (β = 0.034; p < 0.01; H5) on expectation formation.

This research verified that expectations strongly influence a customer's intention to recommend using the private area for clients to other customers to process their claims. These results were consistent with previous studies, which have highlighted the relationship between expectations, satisfaction, and loyalty (Lin and Lekhawipat, 2014; Kanthachai, 2015), as well as the intention to purchase and/or use a product or service. This finding would also be in keeping with those established by the TPB and TAM models used in other contexts (e.g., online banking or tourism) (Martins et al., 2014). With regard to the drivers of expectations, the results show the importance of the drivers of attitude – perceived usefulness and PEOU – from the TAM (Davis, 1986, 1989) and TRA (subjective norm) (Fishbein and Ajzen, 1975), both of which hold that for a technology to be used, individuals must perceive that the benefit of using it outweighs that which they would obtain if they did not use it. Finally, attention should be drawn to the importance of the variable reputation in the formation of expectations, as a mitigator of perceived risk and a quality signal emitted by the company (Kirmani and Rao, 2000).

#### CONCLUSION

The results of this research have partially remedied the lack of previous research (a) using reputation, perceived used, ease of use, subjective norm, and frequency of use as joint antecedents of expectations, and (b) on recommendation to use a private area for

#### TABLE 2 | Reliability and convergent validity of the model.


NA, Not applicable; ∗∗∗p < 0.01.

#### TABLE 3 | Discriminant validity.


Off-diagonal elements are the estimations of the correlations among the factors. The diagonal elements (in bold) are the square root of the average variance extracted.

clients before the recommending customer has used the feature him or herself. In other words, this paper sought to assess aspects that influence the likelihood of recommendation of an online service prior to the process of purchasing or using it.

One key finding was that one way to ensure use of a private area for clients (available exclusively online and to the users of a service) is to encourage the users/customers by emphasizing the important benefits to be obtained from using it, as proposed by McKechnie et al. (2006) and Yucel and Gulbahar (2013). Another key finding was that, despite the high rate of Internet use and the various omnichannel strategies on the market, customers are reluctant to use the private area of a website for their transactions with a company. Instead, they prefer to perform transactions through the traditional channel, because of the personalized nature of the service, using the rest of the channels for less risky procedures (to check information, look for sales, etc.).

#### TABLE 4 | Testing of hypotheses.


R <sup>2</sup> Expectations = 0.311; R<sup>2</sup> Intention to recommend = 0.148. ∗∗∗p < 0.001; ∗∗p < 0.05.

In light of these findings, first, companies should look for ways to encourage the users of the private area to act as vehicles for disseminating the benefits of its use. To this end, they could use strategies based on customizing advertising messages to feature

actual customers and creating discussion forums to generate buzz about the benefits (eWOM). WOM is both one of the main information sources used by users and the most reliable (Chari et al., 2016). For instance, a recent Nielsen (2015) study found that 95% of consumer purchase decisions are based on recommendations received from friends and acquaintances as opposed to other sources.

Second, this research has shown that the drivers of the formation of customer expectations are determined by such important variables as perceived usefulness, reputation, subjective norm, and PEOU. The study sought to determine how consumer/user expectations are formed with regard to the use of an online service as opposed to the provision of the same service offline or through traditional channels. In this regard, many of the studies conducted on the TAM have demonstrated the importance of perceived usefulness. The present findings are consistent with previous studies (e.g., Wu and Wang, 2005; Kim et al., 2008; Hess et al., 2014; Izquierdo-Yusta et al., 2015). By way of example, attention should be drawn to the perceived value model proposed by Zeithaml (1998), which served as the inspiration for several subsequent studies highlighting the influence of perceived value on consumers' behavioral intentions (Dodds et al., 1991; Grewal et al., 1998). From the perspective of the insured, use of the private area of the website should facilitate the entire process, allowing the insured to carry out all processes easily and entailing cost savings (whether in terms of waiting times, travel, opportunity costs, etc.). Therefore, companies should design their websites to be accessible through a small number of clicks, provide fast downloads, be easy to navigate, etc.

Finally, a company's reputation is more important for companies that do not yet have a firmly established online presence than for those that use multiple channels, since in omnichannel environments, consumers/customers form their expectations based on their experiences in offline environments or on knowledge acquired in traditional channels. In this regard, a strong reputation can positively influence customer attitudes (determined by perceived usefulness and PEOU) toward a company's products or services or the channels through which it operates (Erdem and Swait, 2004), as well as the formation of their expectations and the subsequent intention to purchase/use the company's product or service or recommend it. Accordingly, reputation serves as a credible signal that companies send to the market and that consumers use to deal with information asymmetries (Fombrun and Shanley, 1990; Kirmani

### REFERENCES


and Rao, 2000). Consequently, companies should seek to create and/or cultivate or invest in their reputation, as opposed to acting opportunistically (Kirmani and Rao, 2000), making it a source of competitive advantage and, thus, of transaction cost savings.

### FUTURE LINES OF RESEARCH AND LIMITATIONS

This exploratory study enabled the assessment of consumer behavior in online environments. Specifically, it looked at the intention to recommend use of an online service from the perspective of the expectations individuals form prior to using the service themselves. Future lines of research should seek to assess the outcome of this process, i.e., satisfaction, trust, and loyalty. Also, it should be interesting to compare between registered and unregistered users.

The main limitation of the present study lies in the selected sample, i.e., insurance company users. Subsequent research should seek to test the model in another industry with several different products and/or services.

### AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

#### ACKNOWLEDGMENTS

This research was funded by the Spanish Ministry of Economy and Competitiveness under Research Project ECO2014-59688- R, "National Program for Research, Develop, and Innovation Oriented toward Societal Challenges," within the context of the 2013–2016 National Scientific and Technical Research and Innovation Plan.

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fpsyg. 2017.01254/full#supplementary-material





O'Keefe, D. J. (1990). Persuasion: Theory & Research. Thousand Oaks, CA: Sage.


indirect assessment of attitude. J. Pers. Soc. Psychol. 21, 281–295. doi: 10.1037/ h0032322



Zeithaml, V. A., Berry, L. L., and Parasuraman, A. (1993). The nature and determinants of customer expectations of service. J. Acad. Mark. Sci. 21, 1–12. doi: 10.1177/0092070393211001

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Méndez-Aparicio, Izquierdo-Yusta and Jiménez-Zarco. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# A Literature Review of Word of Mouth and Electronic Word of Mouth: Implications for Consumer Behavior

#### Nuria Huete-Alcocer\*

Economía Española e Internacional, Econometría e Historia e Instituciones Económicas, University of Castilla-La Mancha, Albacete, Spain

The rise and spread of the Internet has led to the emergence of a new form of word of mouth (WOM): electronic word of mouth (eWOM), considered one of the most influential informal media among consumers, businesses, and the population at large. Drawing on these ideas, this paper reviews the relevant literature, analyzing the impact of traditional WOM and eWOM in the field of consumer behavior and highlighting the main differences between the two types of recommendations, with a view to contributing to a better understanding of the potential of both.

#### Keywords: businesses, consumers, WOM, eWOM, communication

#### Edited by:

Alicia Izquierdo-Yusta, University of Burgos, Spain

#### Reviewed by:

Doreen Pick, Merseburg University of Applied Sciences, Germany M. Teresa Anguera, University of Barcelona, Spain

> \*Correspondence: Nuria Huete-Alcocer Nuria.Huete@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 31 March 2017 Accepted: 10 July 2017 Published: 25 July 2017

#### Citation:

Huete-Alcocer N (2017) A Literature Review of Word of Mouth and Electronic Word of Mouth: Implications for Consumer Behavior. Front. Psychol. 8:1256. doi: 10.3389/fpsyg.2017.01256

## INTRODUCTION

Consumers increasingly use online tools (e.g., social media, blogs, etc.) to share their opinions about the products and services they consume (Gupta and Harris, 2010; Lee et al., 2011) and to research the companies that sell them. These tools are significantly changing everyday life and the relationship between customers and businesses (Lee et al., 2011).

The rapid growth of online communication through social media, websites, blogs, etc., has increased academic interest in word of mouth (WOM) and electronic word of mouth (eWOM) (e.g., Hennig-Thurau et al., 2004; Brown et al., 2007; Cheung and Thadani, 2012; Hussain et al., 2017; Yang, 2017). Specifically, the present paper will review the literature on how these two media have evolved, the main differences between them, and the degree to which they influence both businesses and consumers, now that they have become some of the most influential information sources for decision-making.

### BACKGROUND

Word of mouth is one of the oldest ways of conveying information (Dellarocas, 2003), and it has been defined in many ways. One of the earliest definitions was that put forward by Katz and Lazarsfeld (1966), who described it as the exchanging of marketing information between consumers in such a way that it plays a fundamental role in shaping their behavior and in changing attitudes toward products and services. Other authors (e.g., Arndt, 1967) have suggested that WOM is a person-to-person communication tool, between a communicator and a receiver, who perceives the information received about a brand, product, or service as non-commercial. Likewise, WOM has been defined as communication between consumers about a product, service, or company in which the sources are considered independent of commercial influence

(Litvin et al., 2008). These interpersonal exchanges provide access to information related to the consumption of that product or service over and above formal advertising, i.e., that goes beyond the messages provided by the companies and involuntarily influences the individual's decision-making (Brown et al., 2007). WOM is widely regarded as one of the most influential factors affecting consumer behavior (Daugherty and Hoffman, 2014). This influence is especially important with intangible products that are difficult to evaluate prior to consumption, such as tourism or hospitality. Consequently, WOM is considered the most important information source in consumers' buying decisions (Litvin et al., 2008; Jalilvand and Samiei, 2012) and intended behavior. For example, tourist satisfaction is of utmost importance because of its influence on behavioral intentions, WOM and purchasing decisions. In other words, overall satisfaction leads to the possibility of revisiting and recommending the destination (Sotiriadis and Van Zyl, 2013).

Similarly, previous research indicates that consumers regard WOM as a much more reliable medium than traditional media (e.g., television, radio, print advertisements, etc.) (Cheung and Thadani, 2012). It is thus considered one of the most influential sources of information about products and services (Lee and Youn, 2009). Users generally trust other consumers more than sellers (Nieto et al., 2014). As a result, WOM can influence many receivers (Lau and Ng, 2001) and is viewed as a consumer-dominated marketing channel in which the senders are independent of the market, which lends them credibility (Brown et al., 2007). This independence makes WOM a more reliable and credible medium (Arndt, 1967; Lee and Youn, 2009).

Today's new form of online WOM communication is known as electronic word-of-mouth or eWOM (Yang, 2017). This form of communication has taken on special importance with the emergence of online platforms, which have made it one of the most influential information sources on the Web (Abubakar and Ilkan, 2016), for instance, in the tourism industry (Sotiriadis and Van Zyl, 2013). As a result of technological advances, these new means of communication have led to changes in consumer behavior (Cantallops and Salvi, 2014; Gómez-Suárez et al., 2017), because of the influence they enable consumers to exert on each other (Jalilvand and Samiei, 2012) by allowing them to obtain or share information about companies, products, or brands (Gómez-Suárez et al., 2017).

One of the most comprehensive conceptions of eWOM was proposed by Litvin et al. (2008), who described it as all informal communication via the Internet addressed to consumers and related to the use or characteristics of goods or services or the sellers thereof. The advantage of this tool is that it is available to all consumers, who can use online platforms to share their opinions and reviews with other users. Where once consumers trusted WOM from friends and family, today they look to online comments (eWOM) for information about a product or service (Nieto et al., 2014).

As a result of ICT, today consumers from all over the world can leave comments that other users can use to easily obtain information about goods and services. Both active and passive consumers use this information medium (eWOM). Individuals who share their opinions with others online are active consumers; those who simply search for information in the comments or opinions posted by other customers are passive consumers (Wang and Fesenmaier, 2004).

Electronic word of mouth also provides companies with an advantage over traditional WOM insofar as it allows them both to try to understand what factors motivate consumers to post their opinions online and to gauge the impact of those comments on other people (Cantallops and Salvi, 2014). However, consumers' use of technology to share opinions about products or services (eWOM) can be a liability for companies, as it can become a factor they do not control (Yang, 2017). To counteract this, businesses are seeking to gain greater control of customers' online reviews by creating virtual spaces on their own websites, where consumers can leave comments and share their opinions about the business's products and services (Vallejo et al., 2015). By way of example, in the field of tourism, companies are starting to understand that ICT-enabled media influence tourists' purchasing behavior (Sotiriadis and Van Zyl, 2013).

Understandably, companies view both types of recommendations – WOM and eWOM – as a new opportunity to listen to customers' needs and adjust how they promote their products or services to better meet them, thereby increasing their return. A negative or positive attitude toward the product or service will influence customers' future purchase intentions by allowing them to compare the product or service's actual performance with their expectations (Yang, 2017).

In the field of consumer behavior, some previous studies (e.g., Park and Lee, 2009) have shown that consumers pay more attention to negative information than to positive information (Cheung and Thadani, 2012). For example, the customers most satisfied with a product or service tend to become loyal representatives thereof via positive eWOM (Royo-Vela and Casamassima, 2011), which can yield highly competitive advantages for establishments, businesses, or sellers, especially smaller ones, which tend to have fewer resources. Some studies have suggested that traditional WOM is the sales and marketing tactic most often used by small businesses.

Additionally, eWOM offers businesses a way to identify customers' needs and perceptions and even a cost-effective way to communicate with them (Nieto et al., 2014). Today, eWOM has become an important medium for companies' social-media marketing (Hussain et al., 2017).

#### WOM vs. eWOM

While many authors (e.g., Filieri and McLeay, 2014) consider eWOM reviews to be electronic versions of traditional WOM reviews, this paper aims to summarize and explain the main differences between the two concepts (**Table 1**). The first such difference is credibility as an information source (Cheung and Thadani, 2012; Hussain et al., 2017), since it can influence consumers' attitudes toward products or services (Veasna et al., 2013), for example, with regard to the purchase of tourism services, which are considered to be high-risk (Sotiriadis and Van Zyl, 2013). Luo et al. (2013) have suggested that the anonymity of online messages could have a negative effect on their credibility. In contrast, other

#### TABLE 1 | Differences between WOM and eWOM.

fpsyg-08-01256 July 21, 2017 Time: 15:22 # 3


Source: The author.

studies (e.g., Hussain et al., 2017) have argued that consumers use eWOM more to reduce risk when decision-making. Likewise, eWOM tends to be more credible when the consumer using it has previous experience (Sotiriadis and Van Zyl, 2013).

Message privacy is another feature that sets the two media apart, since with traditional WOM information is shared through private, real-time, face-to-face dialogs and conversations. In contrast, information shared through eWOM is not private and can sometimes be seen by anonymous people who do not know each other. Furthermore, reviews can be viewed at various points in time (Cheung and Thadani, 2012). Indeed, because eWOM reviews are written, consumers and companies can check them at any time; this stands in contrast to traditional WOM, where once the message has reached the receiver, it tends to disappear.

Another salient difference between the two media is the speed of diffusion of the message; eWOM statements spread much faster than WOM statements because of where they are published, i.e., on the Internet (Gupta and Harris, 2010). Online platforms for sharing information (social media, websites, blogs, etc.) are what set eWOM apart from traditional WOM (Cheung and Thadani, 2012). First, they make the reviews accessible to more consumers (Cheung and Thadani, 2012; Sotiriadis and Van Zyl, 2013). Second, because they are written, they persist over time (Hennig-Thurau et al., 2004; Cheung and Thadani, 2012).

### CONCLUSION

This paper has reviewed the literature with a view to providing a clearer understanding of WOM and eWOM in the context of consumer information searches.

To this end, the review found that, in keeping with numerous studies, WOM is both the oldest medium for sharing opinions about products or services and the one most likely to influence consumer behavior, due to the high reliability and credibility transmitted by family and friends. In contrast, few studies have examined the interaction between perceived risk and eWOM source credibility (Hussain et al., 2017).

Notwithstanding the above, the review of the theoretical framework also revealed a gap in the literature on WOM credibility in situations involving multiple or many communicators and receivers and how this ultimately affects the end consumer. This would include, for instance, situations in which one person communicates a message to another, who acts as an intermediary, both receiving the original message and passing it along to a third party, i.e., the end consumer. In such cases, the original message can be altered or distorted, chipping away at the credibility of the WOM review as a source of information. This lends much more strength to written comments and reviews, such as eWOM, which can ultimately reduce risk and increase consumer confidence.

Another feature that distinguishes eWOM from traditional WOM is the speed with which it spreads and the ease of access to it. In this regard, when consumers need information about a product or service, they ultimately turn to online media (eWOM) for two reasons. First, they can get the information more quickly, as there is no need to wait for someone else – a friend or family member – to offer an opinion about what they wish to consume. Second, if they have already received WOM reviews, they can use eWOM to corroborate the information received. Therefore, credibility and speed are the two main features not only distinguishing the two media, but also influencing consumer behavior.

Finally, the analysis of the review showed that these two concepts – WOM and eWOM – while seemingly the same, are at the same time very different. The Internet has transformed traditional WOM into eWOM. The communication of opinions is no longer done interpersonally (i.e., person-to-person or faceto-face), but rather is mediated by ICT. However, the many studies conducted (e.g., Katz and Lazarsfeld, 1966; Brown et al., 2007; Daugherty and Hoffman, 2014; Yang, 2017) agree that they are the media most able to influence consumer behavior and the most often used to obtain information before, during, and after consuming a given product or service. For example, in the field of tourism, eWOM is considered the most influential prepurchase source of travel information (Sotiriadis and Van Zyl, 2013).

### AUTHOR CONTRIBUTIONS

This paper tries to offer a clearer understanding of the two concepts through a literature review and an exploration of how, as a result of advances in ICT, traditional WOM has given rise to eWOM. The author has made an important, direct, intellectual contribution to this paper and has approved it for publication.

### FUNDING

This research was funded by the Spanish Ministry of Economy and Competitiveness under Research Project ECO2014-59688-R ("Planning and implementation of optimal management strategies for physical, online and mobile POSs based on ICT and innovation").

## REFERENCES

fpsyg-08-01256 July 21, 2017 Time: 15:22 # 4


**Conflict of Interest Statement:** The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Huete-Alcocer. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Consumer-Brand Relationships under the Marketing 3.0 Paradigm: A Literature Review

#### Mónica Gómez-Suárez<sup>1</sup> , María Pilar Martínez-Ruiz<sup>2</sup> and Noemí Martínez-Caraballo<sup>3</sup> \*

<sup>1</sup> Finance and Marketing Department, Universidad Autónoma de Madrid, Madrid, Spain, <sup>2</sup> Department of Business Administration, University of Castilla-La Mancha, Albacete, Spain, <sup>3</sup> Economy and Management, Centro Universitario de la Defensa de Zaragoza, Zaragoza, Spain

Consumer-brand relationships encompass several dimensions, most of which have attracted growing research attention during the last years. Building these relationships is especially important in the marketing 3.0 era, where it is suggested that customers will choose those brands that satisfy their deepest needs. With these ideas in mind, this article provides a review of two key concepts implied in such relationships: brand love and customer engagement. Although both conceptions focus on different stages of consumer-brand relationships, they actually cover different perspectives on the same process. Moreover, they come from diverse conceptual paradigms: whilst brand love comes from the psychology discipline, engagement derives from diverse areas of the marketing field (e.g., the service-dominant logic perspective). However, their further empirical developments have taken place in marketing. Besides, both terms appear to be applied to different empirical perspectives: brand love is usually linked to the Fast Moving Consumer Goods industry and customer engagement to services.

Edited by:

Ana I. Jiménez-Zarco, Open University of Catalonia, Spain

#### Reviewed by:

Isabel Llodrà-Riera, Fundació Balears d'Innovació i Tecnologia, Spain Antoni Olive-Tomas, Ramon Llull University, Spain

#### \*Correspondence:

Noemí Martínez-Caraballo noemar@unizar.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 20 December 2016 Accepted: 08 February 2017 Published: 22 February 2017

#### Citation:

Gómez-Suárez M, Martínez-Ruiz MP and Martínez-Caraballo N (2017) Consumer-Brand Relationships under the Marketing 3.0 Paradigm: A Literature Review. Front. Psychol. 8:252. doi: 10.3389/fpsyg.2017.00252 Keywords: brand love, consumer behavior, customer engagement, marketing 3.0, values-driven era

## INTRODUCTION

Society is changing at an ever-increasing rate, especially since the beginning of the 21st century due, among other reasons, to the increasing diffusion of information and communication technologies (ICTs). These changes are producing many modifications in the consumers' behavior, as well as in the relationships they establish with companies, derived from the huge possibilities that ICTs offer. For instance, ICTs allow sharing information about firms, theirs products and their brands at a global level, across boundaries (e.g., Parasuraman and Zinkhan, 2002; Yadav and Varadarajan, 2005). In view of this situation, it is not surprising that by the middle of the first decade of the 21st century the concept of Marketing 3.0 –as known as the "values-driven era"– has emerged. It is a type of marketing that tries to face and respond to the current challenges, derived from globalization issues (Kotler et al., 2010), among others.

In this scenario, one of the trendy research lines in marketing, the study of consumerbrand relationships, reinforces the work of Papista and Dimitriadis (2012). In particular, the emergence of Marketing 3.0 is one of the powerful reasons why this particular research line has been receiving increasing attention during the last years. Marketing 3.0 emphasizes the need to take care of customers not as mere consumers, but as complex and multi-dimensional human beings. Under this paradigm, the role of brands as identifiers of products and firms has been overcome. Companies must posit their brands instead to seek to address to social, economic, and environmental issues as a way of engaging with society (Jiménez-Zarco et al., 2014).

Two of the main concepts regarding consumer-brand relationships are: brand love and customer engagement. Hence, a comprehensive literature review of these concepts is provided in this paper. In particular, we begin by offering a brief review of the relevant literature on Marketing 3.0 and its managerial implications. Then, the concepts of brand love and customer engagement are presented. Finally, conclusions and managerial guidelines are provided.

### CONCEPTUAL BACKGROUND: BRAND LOVE AND CUSTOMER ENGAGEMENT AS KEY CONCEPTS IN THE MARKETING 3.0 CONTEXT

In the values-driven era, people demand to be treated not as just simple consumers; instead, they want to be treated as whole human beings with minds, hearts, and spirits (Kotler et al., 2010). Emerged as a response to the desire of people to growingly express creativity, values and spirituality, Marketing 3.0 makes companies behave as active agents, aiming at positioning themselves as companies whose brands have respect and admiration (Jiménez-Zarco et al., 2014).

This has several implications for brand management. On the one hand, in order to positioning a brand, companies must take into account that the way to differentiate the brand sometimes is not related to the mere fact of attaching the brand itself to a product or service -it should rather link the brand to a particular set of potential emotional benefits that it promises to deliver to the consumer. On the other hand, it is expected that those brands that are acknowledged as ethical elicit positive emotional responses among its consumers and invoke a stronger level of brand affect among them (Glomb et al., 2011; Martínez-Cañas et al., 2016).

Given the relevance of the affective and emotional links usually generated between brands and consumers, companies must take them into account in order to build and manage sustainable brands along time. With this regard, it is interesting to mention how two areas of research have sparked particular interest in the marketing literature because of their special links with emotions: brand love and consumer engagement (Gómez-Suárez et al., 2016).

The first empirical studies carried out to examine these intense consumer-brand relationships were those analyzing the first of these concepts. Nevertheless, Sallam (2014) outlined how it was first introduced by Shimp and Madden (1988), the managerial interest for brand love came after the publication of Roberts (2006). For this author, "lovemarks" were brands that were positioned not only in the mind but also at the heart, causing intrigue, excitement, appreciation and desire among their customers (Pawle and Cooper, 2006). Subsequently, Professor Aaron Ahuvia and his co-authors carried out several research works (e.g., Ahuvia, 1993; Carroll and Ahuvia, 2006; Batra et al., 2012; Ahuvia et al., 2014) in order to conceptualize brand love and to draw several empirical applications.

Brand love represents an intimate experience of the customers –in positive emotional terms– toward the brand. Authors who have previously conducted research on this topic provide different definitions, such as "a product, service, or entity that inspires loyalty beyond reason" (Pawle and Cooper, 2006, p. 39), "the degree of passionate emotional attachment a satisfied consumer has for a particular trade name" (Carroll and Ahuvia, 2006, p. 81) or "a higher-order construct including multiple cognitions, emotions, and behaviors, which consumers organize into a mental prototype" (Batra et al., 2012, p. 2). Regarding brand love, the following papers are worthy of a special mention: Fetscherin et al. (2014), Sarkar and Sreejesh (2014), Huber et al. (2015), Vernuccio et al. (2015), and Kaufmann et al. (2016).

In the management literature, (employee) engagement was first conceptualized by Kahn (1990). High customer engagement, like employee engagement, means that customers present themselves physically, cognitively, and emotionally –i.e., during the service encounter–, particularly since customers have been considered as "partial employees" or "co-producers" in some service situations (Baron et al., 1996; Bendapudi and Leone, 2003) 1 . Nowadays, several organizations consider important consumer engagement and know that it needs to be at the center of the customer service strategy.

In general terms, customer engagement is focused on the interactions between the firm and the customers and is a key research priority of the Marketing Science Institute (MSI). Focusing on customer engagement, the following papers are worthy of a special mention: Brodie et al. (2011), Hollebeek and Chen (2014), and Hollebeek et al. (2014).

According to Hollebeek (2011), there is a lack of consensus pertaining to the definition of engagement-based concepts. Among all of them, we have chosen the two most cited in the literature: customer engagement is "a multidimensional concept comprising cognitive, emotional, and/or behavioral dimensions, [which] plays a central role in the process of relational exchange" (Brodie et al., 2011, p. 3), and/or "the level of the customer's (or potential customer's) interactions and connections with the brand or firm's offerings or activities, often involving others in the social networks created around the brand/offering/activity" (Vivek et al., 2014, p. 406).

Albert et al. (2008) and Batra et al. (2012) develop measurement scales of brand love, which would enable identifying both brands and product categories that might benefit from a consumer-brand relationship. In order to measure customer brand engagement (CBE), Hollebeek et al. (2014) and Vivek et al. (2014) develop and validate CBE scales in different settings.

Although the aforementioned definitions contain common elements, and three papers –Bergkvist and Bech-Larsen, 2010; Sarkar and Sreejesh, 2014; Wallace et al., 2014– have been published that partially investigate the existence of a certain connection between both concepts; when exploring which

<sup>1</sup>Co-creation is "the joint creation of value by the company and the customer; allowing the customer to co-construct the service experience to suit their context" (Prahalad and Ramaswamy, 2004, p. 8).

underlying theories have supported previous research, draws the conclusion that they start from different paradigms.

Basically, it could be said that there has been a fragmented interpretation depending on the research tradition in which they have been supported. Thereby, customer engagement have been developed from several marketing theories, such as the expanded domain of relationship marketing (Morgan and Hunt, 1994), or the service-dominant logic perspective (Vargo and Lusch, 2004, Vargo and Lusch, 2008); whilst brand love is based on theories from the psychology domain, such as the triangular theory of interpersonal love (Sternberg, 1986). Another important issue arises: different names are sometimes used when referring to the same concepts. For instance, the concept of self-congruity –that derives from branding theories– has nearly the same meaning as identity, derived from the identification theory.

Finally, both concepts are frequently applied to different objects of study. In general, brand love has been analyzed in research applied to consumer brands in the Fast Moving Consumer Goods industry, whilst customer engagement has been used mainly in the service sector. However, there are studies that have tried to overcome this limitation. Especially for fulfilling this research gap, in Long-Tolbert and Gammoh (2012), a model have been developed in order to apply brand love to the case of intangible goods.

### CONCLUSION

Consumer-brand relationships have received increasing attention during last years, being Kaufmann et al. (2016) one of the last contributions to this topic. This paper contributes to the understanding of the complex brand relationships consumers have, by exploring the dynamics between brand love and cocreation. In the literature on consumer-brand relationships, there are two concepts that determine a very intense link between them; i.e., brand love and consumer engagement. In this paper, a conceptual delimitation of these two key terms has been done.

To this end, a brief reflection has been started on the context of Marketing 3.0 and how it has boosted the research line dealing with the study of relationships between consumers and brands. Under this paradigm, consumers try to acquire those brands that allow them to especially meet their deeper needs for social, economic, and environmental justice. Namely, consumers no

#### REFERENCES


longer consider only the role of brands as mere identifiers of products, services or companies, but try to go further, basing their brand choices on potential associations and emotional benefits that a specific brand is susceptible to provide them. Then, a conceptual demarcation of terms brand love and customer engagement has been established. In particular, it has become clear how such terms derived from differentiated disciplines and have often been applied to different empirical situations.

Nowadays, given the profound changes taking place in markets, it is necessary to pay attention to how the consumerbrand relationships continue to evolve. In short, among future trends, there should be considered those that are likely to have a greater impact on these relationships, such as the opportunities offered by an efficient management of Big Data and the advent of Marketing 4.0. Big data helps companies to build strong relationships. Marketing 4.0 is the marketing of big data (Jiménez-Zarco et al., 2017). Marketing 4.0, from human-centric to content marketing, helps companies to adapt to the changing nature of customer paths in the digital economy (Kotler et al., 2017). Marketing 4.0 requires: firstly, a depth knowledge about the evolution of marketing, especially about Marketing 3.0, and secondly, an analysis of how technology –not only the Internet and social media– can be used to design marketing strategies that enhance the brand-consumer relationships.

### AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct, and intellectual contribution to the work, and approved it for publication.

#### FUNDING

This research was funded by the Spanish Ministry of Economy and Competitiveness, Research (Project reference: ECO2014- 59688-R, Programa Estatal de Investigación, Desarrollo e Innovación Orientada a los Retos de la Sociedad, Plan Estatal de Investigación Científica y Técnica y de Innovación 2013–2016). Moreover, financial support from research group CREVALOR, funded by the Diputación General de Aragón and the European Social Fund, is gratefully acknowledged.



**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Gómez-Suárez, Martínez-Ruiz and Martínez-Caraballo. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Consumer Participation in Co-creation: An Enlightening Model of Causes and Effects Based on Ethical Values and Transcendent Motives

Ricardo Martínez-Cañas \*, Pablo Ruiz-Palomino, Jorge Linuesa-Langreo and Juan J. Blázquez-Resino

Business Administration Department, University of Castilla-La Mancha, Cuenca, Spain

In the current highly interconnected modern world, the role of consumers has changed substantially due to their active collaboration with companies in product and process innovation. Specifically, consumer participation has become key to the development of successful products and services, as companies have come to rely more and more on consumers' opinion as a source of innovative ideas and brand value. However, whereas existing research has focused on identifying the different elements involved in consumers' co-creation, there is still the need to comprehend better this complex mechanism by integrating distinct dimensional insights. With an integrative review of research into three important perspectives, one nurturing from the Service-Dominant logic, another one based on the information and communication technologies (ICTs) platforms, and (the ethical values-driven) Marketing 3.0 paradigm, this article proposes a conceptual framework in which consumers' ethical values and transcendent motivations play an important role in encouraging their engagement in co-creation activities. In this connection, and with consumers increasingly embracing the need to fulfill a social and ethical function in society, the co-creation process is here comprehended as a means to emphasize the social and moral aspects of co-creation. This article also identifies the important, supportive role of the Marketing 3.0 paradigm and Web 3.0 tools to initiate the co-creation process, as well as the important valuable benefits attained by both companies and consumers after consumers engage in this process. Importantly, these benefits are highlighted to increase when ethical products are the object of these co-creation activities. All these insights have notable implications for both research and managerial practice.

Keywords: value co-creation, ethical values, transcendent motives, information and communication technologies, ethical products, Marketing 3.0, conceptual paper

## INTRODUCTION

In the recent times value co-creation has emerged as a major strength of companies to remain and gain competitiveness (Zwass, 2010). Defined as a holistic management strategy focused on bringing distinct agents together to produce mutually valued outcomes (Prahalad and Ramaswamy, 2004), this is increasingly utilized by companies as a source of corporate reputation, brand value,

#### Edited by:

Alicia Izquierdo-Yusta, University of Burgos, Spain

#### Reviewed by:

Eva Reinares Lara, Universidad Rey Juan Carlos, Spain Natalia Medrano, University of La Rioja, Spain

> \*Correspondence: Ricardo Martínez-Cañas ricardo.martinez@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 18 March 2016 Accepted: 11 May 2016 Published: 26 May 2016

#### Citation:

Martínez-Cañas R, Ruiz-Palomino P, Linuesa-Langreo J and Blázquez-Resino JJ (2016) Consumer Participation in Co-creation: An Enlightening Model of Causes and Effects Based on Ethical Values and Transcendent Motives. Front. Psychol. 7:793. doi: 10.3389/fpsyg.2016.00793 and competitive advantage (Cova and Dalli, 2009). Beyond the conception of market as company centric-based, where economic exchange is about making and distributing things to be sold (Good-Dominant Logic, see Vargo and Lusch, 2004, 2008), companies are increasingly conceiving markets as the intersection of companies, network partners, and consumers to co-create value (Service-Dominant Logic, see Vargo and Lusch, 2008; Vargo et al., 2008; Williams and Aitken, 2011). This new understanding has guided in practice to conceive consumers, along with their personalized experiences, not as passive but rather active players to create value in designing and developing products/services (Prahalad, 2004; Prahalad and Ramaswamy, 2004).

While co-creation research has importantly advanced our understanding around the concept in the last decade (Prahalad and Ramaswamy, 2004; Holbrook, 2006; Grönroos, 2008; Payne et al., 2008; Vargo and Lusch, 2008; Zwass, 2010; Brodie et al., 2013), relatively little knowledge exists about how consumers engage in co-creation. Whereas, there have been some major attempts to understand this process (see Payne et al., 2008; Brodie et al., 2013), the understanding of the concept is still far complete. Our review of literature notes that interesting pieces of this complex puzzle are still missing (see Arvidsson, 2011), and an integration of the different existing perspectives around the concept, which has not yet been well addressed in literature (see Edvardsson et al., 2011), appears to be essential.

One important hitherto area of research has involved studies around the co-creation processes nurturing from the Service-Dominant logic (S-D logic), which posits consumers as resource integrators, value creating entities as traditional companies are (Vargo and Lusch, 2008; Edvardsson et al., 2011). With this theory in mind, previous research has studied consumers' involvement in co-creation to understand elements involved for the process to occur (Cova and Dalli, 2009; Brodie et al., 2013; Roberts et al., 2014). One of these important elements refers to motivations driving consumers to participate in co-creation activities. Specifically, and from a psychological perspective, studies have revolved around either intrinsic or extrinsic motives to explain why consumers participate in value co-creation processes (Roberts et al., 2014). In addition, under the umbrella of this (S-D logic) perspective, another important stream of research has focused on examining and identifying the positive outcomes deriving from these processes (i.e., consumer satisfaction, consumer learning, consumer brand loyalty; Payne et al., 2008; Bowden, 2009; Jaakkola and Alexander, 2014; Luo et al., 2015).

One another perspective encompasses an array of studies (see Kalaignanam and Varadarajan, 2006; Zwass, 2010; Garrigos-Simon et al., 2012) stressing the significant influence of information and communication technologies (ICTs) platforms to initiate and develop this process. These studies have much emphasized how the rapid diffusion of the ICTs have facilitated both social interaction and virtual communities, which notably enhances productive co-creation processes. Indeed, important elements that have facilitated consumers' access to information and social relationships include the unstoppable spread of the Internet, the ICTs, the Web 2.0 (social networks), consumer relationship management, and even the more recent Web 3.0 platforms, which offer improved interaction and learning through data mining, semantic webs, and artificial intelligence. These technologies and tools support a better understanding of consumers' existing and potential needs and problems. In particular, Web 3.0—with the support of social media tools and intelligent technologies—enables bidirectional conversations between consumers and companies, and the Web 3.0 technology platforms are designed to engage consumers in a collaborative interaction that provides mutually beneficial value (Choudhury and Harrigan, 2014). Interactions thus, are part of the modern, global era, which are greatly favored by ICTs, thus, in turn, contributing to foster consumers' active and valuable collaboration with companies in product and process innovation.

Although both perspectives are critically formative of the existing body of research, in the last times some studies (see Payne et al., 2008; Edvardsson et al., 2011) have discussed the need to introduce new theories to better understand the consumer's cocreation process. In this connection, recent literature has begun to nurture from a new, very incipient perspective (i.e., Kotler et al., 2010; Arvidsson, 2011; Williams and Aitken, 2011), the ethical values-driven Marketing 3.0 paradigm (Kotler et al., 2010), to investigate this process. This new understanding turns around the idea that consumers are increasingly seeking solutions to their own concerns and are interested in building a better world, guided by their ethical values in purchasing decision processes (e.g., Hollenbeck and Zinkhan, 2010). It understands, as observed in several studies (i.e., Shaw et al., 2005), that consumers, prior to show affinity to brands, attend to the benevolence, security, equality, and environmental responsibility that brands and products exhibit. In other words, consumers' choices of products and services are increasingly based on the extent to which they permit to fulfill their higher-order needs for social, economic, and environmental justice (Kotler et al., 2010). This applied to our research subject would entail other type of motives, beyond those of an intrinsic or extrinsic nature, for consumers to involve in co-creation activities, namely, transcendent motives (e.g., Pérez-López, 1998; Guillén et al., 2014). Including thus transcendent motives in our understanding of the issue might fill a research gap that exists in the current knowledge about how and why consumers engage in value co-creation processes.

Given the attractive this new perspective might offer to gain better understanding of consumers' co-creation processes, this study tries to utilize and integrate it together with the other aforementioned areas of research. Specifically, and following examples of Payne et al. (2008) and Edvardsson et al. (2011), this article seeks to do an integrative literature review around these three important perspectives, the S-D logic-based area of research, the ICTs platforms-based understanding, and the ethical valuesdriven Marketing 3.0 paradigm. Thus, this article tries to shed light on several fundamental questions:

(a) What are the antecedents (i.e., extrinsic, intrinsic, or ethical motivators) of consumers' engagement in co-creation processes?, or in other words, why do consumers engage in value co-creation activities with companies?


In seeking answers to these questions, our conceptual approach and subsequent model tries to develop a better understanding of both antecedents and consequences of consumers' engagement in co-creation activities. Importantly, our study introduces missing elements in this area, very related to ethics and transcendent motives, which help form a substantial source of information around the principal elements involved in consumers' cocreation processes. It should be of assistance to both scholars and practitioners who seek to understand or optimally manage consumers' engagement in co-creation activities.

After this introduction, our second section describes the method (i.e., integrative literature review) we used to gain a new enlightening perspective on the consumers' co-creation process. The third section puts forward the conceptual analysis we have undertaken regarding the different elements involved in understanding consumers' engagement in co-creation activities, which produces a series of theoretical propositions. Next, a fourth section synthesizes and links all these elements in a form of conceptual model that integrates both causes and effects of the co-creation process. In the final sections, the article outlines a series of conclusions and offers compelling implications for both research and practice. Limitations of our study, and avenues for future research are also finally outlined.

### METHODS

In this paper an integrative literature review has been conducted as a distinctive form of research to generate new knowledge to literature (Torraco, 2005; Yorks, 2008). Specifically, our integrative literature review aims to elucidate good understanding around the important elements involved in consumers' intentions to co-create with companies. This method has emerged in recent years as a relevant one to criticize and synthesize various streams of literature (Shuck, 2011), and importantly, to set up new theoretical frameworks and research agendas in the social sciences field (i.e., Carasco-Saul et al., 2015; Mercurio, 2015).

Although an integrative literature review can be also used to investigate mature topics and emergent topics, the latter ones, which are relatively new and lack a comprehensive state of art, can benefit importantly from holistic conceptualizations and literature syntheses. In this paper we address an emerging topic in the marketing and consumer behavior literature, so accordingly, this method could help substantially to bring into light new perspectives (i.e., ethical values, Marketing 3.0, transcendent motives) in our understanding on both antecedents and positive effects of consumers' engagement in co-creation activities. Thus, thanks to using this research method, a value-added theoretical contribution widening the classical economical perspective of the value co-creation phenomenon, has been feasible.

Following the criteria used in this method to conduct this type of research proficiently (Torraco, 2005; Yorks, 2008), the most important bibliographical databases were primarily selected. Research databases such as Web of Science, Scopus, PsycoInfo, ABI/Inform, JSTOR, the Academy of Management database, EBSCO Academic Search Premier, and Google Scholar were utilized to access the literature. Next, these databases were used to conduct an electronic search for published articles around our subject matter of this study, through including primary keywords. With a coverage of databases during a 15 years period (2000–2015), search words included "value co-creation" AND (business OR management OR psychology OR organizations OR motivation OR ethics OR transcendent OR marketing OR Web 3.0). In addition, criteria used for retaining or discarding articles was threefold: relevance of the title, diffusion relevance (times cited), and inclusion in the abstract of terms such as "motivation" "transcendent" "value co-creation" "marketing" "Web 3.0" and "ethics." Also, a deep inclusion-exclusion criteria was used after a complete reading of each pre-selected article. Finally, due to their important scholarly status, other relevant references were included.

All these articles were selected to analyze the existing research under the perspective of elucidating links between the different elements involved in the consumers' value co-creation phenomenon. Thus, the selected papers for our integrative review show their high relevance to better understand our topic of interest. Specifically, they allowed us to widen the existing current perspective around the theme by including the important role of ethical values, Marketing 3.0 and transcendent motivations in driving consumers' engagement in value co-creation activities. Thanks to this theoretical integration our article represents valuable research in conceptual terms, providing direction for researchers and practitioners who are interested in studying or developing, leveraging, and managing, value co-creation activities.

## ANALYSIS AND PROPOSITIONS

#### Co-Creation and Consumer 3.0 Value Co-Creation

The value co-creation concept often refers to company mechanisms, such as co-design (Lusch et al., 2007), coproduction (Auh et al., 2007; Sanders and Stappers, 2008), consumer participation (Bendapudi and Leone, 2003; Bagozzi and Dholakia, 2006; Fang et al., 2008; Chan et al., 2010; Olsen and Mai, 2013), user innovation (Barki and Harwick, 1994; Lagrosen, 2005), customer engagement (Bowden, 2009; Jaakkola and Alexander, 2014), pro-consumption (Richards, 2011), or co-innovation (Lee et al., 2012). In essence, value co-creation ultimately is a holistic management initiative or economic strategy that brings different parties together to produce a mutually valued outcome (Prahalad and Ramaswamy, 2004).

Research into the value co-creation concept has importantly configured and evolved thanks to the emergence of the S-D logic, which brings to light the important ingredient consumers (and other agents) represent in product and process innovation (see Vargo and Lusch, 2008; Vargo et al., 2008; Williams and Aitken, 2011). However, based on this S-D logic new understanding of creating value though, various other theoretical approaches have been emerging around the concept (Saarijärvi et al., 2013), including the service science, service logic, many-to-many marketing, social constructionist, new product development, or post-modernism perspectives. **Table 1** summarizes these key approaches, ideas, concepts, and authors, and shows they differ, to some extent, in their characteristics and locus of attention (e.g., companies, customers, communities, networks). Consequently, value co-creation as a concept lacks a clearly united basis for further development. Yet this divergence also provides an interesting starting point for addressing important questions about who benefits from the created value, what kind of resources are used, and what mechanism (or technology) defines how company resources get integrated into customer processes (Saarijärvi et al., 2013).

Analyzing the main differences among these theoretical approaches in detail reveals that value co-creation is based on interactive processes, promoted by agents with valuable resources that they could offer up for integration (Prahalad and Ramaswamy, 2004). In addition, value co-creation emphasizes joint efforts by companies, consumers, and other agents, such that reciprocity and mutual dependence are particularly important in defining the interdependent roles associated with the production of services and value creation (Vargo et al., 2008). In the participation among agents, this co-created value arises in the form of personalized, unique experiences derived from the valuein-use for the customer or value-in-context in general. These benefits, together with ongoing revenues, learning, and enhanced market performance, can drive some desired effects for both companies (e.g., trust, commitment, loyalty, risk reduction, cost effectiveness) and consumers (e.g., empowerment, commitment, satisfaction, learning, personalized experiences).

According to a classical value creation approach, companies offer innovative products (Kirca et al., 2005) by leveraging their distinctive, differentiated capabilities to create great value for consumers and achieve competitive advantages. In the value co-creation paradigm, companies instead co-create such benefits together with consumers (or other agents), with a more humanistic view, which ultimately might enhance consumers' loyalty, based on their own perceptions (Slater and Narver, 1995; Füller, 2006). Furthermore, consumers must be willing and able to interact with companies and contribute to the process, which constitutes a key challenge (Lengnick-Hall et al., 2000; Sawhney and Prandelli, 2000; Auh et al., 2007; Füller and Matzler, 2008). Understanding consumers is not enough to ensure new product success; consumers also must be active or proactive (Lagrosen, 2005), as well as intrinsically or extrinsically motivated to share their knowledge, ideas, and preferences with companies (Füller, 2006). For example, consumers' leit motive could relate to activities that lead to unique experiences, which then would involve both customer participation and a connection to the experience (Shaw et al., 2011). Ensuring the success of a new product or service thus requires (among other factors) a more humanistic, detailed understanding of consumers' ethical values and transcendent motives, which determine their behavior. But acknowledgment of the concrete exchange situation (product-service characteristics, technological platform) also is critical. Therefore, using a marketing strategy that is oriented toward social aspects and defining the appropriate role of ICTs and Web 3.0 platforms represent important elements.


#### Marketing 3.0, Web 3.0, and ICTs Platforms

Marketing 3.0 represents the most recent marketing paradigm, with the key assumption that companies treat consumers as human beings with intelligence, heart, soul, and spirit (Kotler et al., 2010). As such, it is a prominent philosophy, gaining relevance among consumers who increasingly recognize the effects of unpredictable social, economic, and environmental changes on them (Kotler et al., 2010). Previous paradigms included Marketing 1.0, which emerged during the industrial, "product-centric" era and focused on mass sales of products based on functional value propositions; the marketing department's activities centered solely on the product or service for sale. An enhanced version, Marketing 2.0, arose during the information age, such that companies adopted an emotional value proposition. That is, Marketing 2.0 is based not on product transactions but on the relationships that allow companies to engage consumers with messages and individualized services and products (Corbae et al., 2003). Consumers differ in their preferences, so companies must segment the market and develop unique products for different consumers (Kotler et al., 2010).

The new paradigm of Marketing 3.0 implies that we are at the dawn of a "values-driven" era, characterized by populations who want to satisfy functional, emotional, and spiritual needs through their consumption. Marketing 3.0 seeks to satisfy the whole person (mind, body, soul). This evolution to more humancentric value propositions is shaping the future of marketing in three main ways (Kotler et al., 2010). First, mass participation and co-creation through collaborative marketing reflect how modern social media and the Internet have tapped into natural human desires for connectivity and interactivity. Companies thus seek collaborative marketing strategies, such as product or service co-creation, with consumers, employees, channel partners, and other companies that have similar goals and values. Second, in the globalization paradox, technological advances have created truly "global citizens" who still want to be considered individuals. Accordingly, marketing needs to address both local and global communities simultaneously. Third, the rise of a creative society and human–spirit marketing encourages creative people, who tend to innovate, collaborate, and express themselves more than others, to pursue their self-actualization but also demand originality and trendiness in the products and services they consume.

Therefore, the conceptual approach of Marketing 3.0 entails a redefined triangle: Rather than brand, positioning, and differentiation, it builds on this formula to suggest a 3-I model, encompassing identity, image, and integrity (Kotler et al., 2010). The first I, identity, reflects the relationship between positioning and brand and seeks to address the rational portion of the value proposition. Image instead lies at the juncture of differentiation and brand and strives to capture the emotions of the target audience. Finally, integrity represents the intersection of positioning and differentiation, and it aims to fulfill the brand promise authentically while fostering trust, commitment, and loyalty. This 3-I model demonstrates how more intangible and social factors can determine the real and perceived value created by a company.

Accordingly, we anticipate that, in line with the Marketing 3.0 paradigm, value co-creation is an integrated process of proactive, creative, and social cooperation among companies and consumers. Through various value co-creation activities, organizations attract consumers, and engage them in discussions of their emotions, feelings, and expectations, thereby generating a constructive, deep exchange of ideas, resources, and services (Piller et al., 2010). Furthermore, these exchanges are more plausible as a result of the new advances in ICTs, including Web 3.0 platforms. These platforms represent a third generation of Internet-based services that collectively comprise what might be called "the intelligent Web," which includes semantic webs, micro-formats, natural language searches, data mining, machine learning, recommendation agents, and artificial intelligence technologies (Markoff, 2006). These services explicitly emphasize machine-facilitated understanding of information as a means to provide a more productive, intuitive user experience. Therefore, Web 3.0 enables customers to converge with companies through several emerging technology trends, such as ubiquitous connectivity, network computing, open technologies, open identity, and a more intelligent web.

From another perspective, Web 3.0 and other ICTs platforms constitute the technical elements needed to implement Marketing 3.0 strategies, because they enable consumers to create value by facilitating their collaborations with companies (Kalaignanam and Varadarajan, 2006) while also increasing the adaptation and personalization of products, brands, and services by and for different users, according to their own needs (Garrigos-Simon et al., 2012). Such technological advances offer new tools that consumers can use to interact, as well as incentives for creating new products and services. The ubiquity of the Internet, Web 3.0, and ICTs also allows users to interact widely and easily, with both companies and other users. The Internet has increased consumers' power, through two main processes: reformulating the identity of each user (through interactions with others, learning processes, and the creation of social links) and increasing users' efficiency and skills (Amichai-Hamburger et al., 2008). These tools also have an important role in helping companies gain advantages for the design and delivery of customized products that maximize consumers' satisfaction (Du et al., 2006). However, more skilled and powerful consumers need ICTs to help them proactively generate and evaluate new ideas, improve product details, select and personalize preferred prototypes, experience new product features (e.g., through simulations), obtain and share new product information, and participate in the development of new products (Füller, 2006). Thus, we propose:

Proposition 1: Web 3.0 platforms and the generalization of new and advanced ICTs boost consumers' engagement in value co-creation activities.

### Consumers' Co-Creation Motives: Ethics and Transcendence

#### Consumers' Motivations to Co-Create

Motivation is an antecedent of human behavior, explaining why people behave in certain ways, what provokes these behaviors, and what directs subsequent voluntary actions (Deci and Ryan, 1985; Nambisan, 2002). Prior literature explicates what motivates people to act, using various theories that attempt to detail the entire human motivation process (Ambrose and Kulik, 1999). Relying on seminal work on human motivations by Guillén et al. (2014), we offer a simplification and integrative review, with the goal of providing theoretical support for the motivation process involved in co-creation. Thus, we present two basic motivational approaches: Maslow's (1943) and Herzberg's (1968).

Maslow's Theory of Human Motivation (1943) classifies motivations according to whether they seek to meet basic, lowerorder, physiological needs (food, water, safety, and security) or higher-order needs linked to social activities, such as esteembuilding, self-actualization, or continuous self-improvement. These needs act as motivators until they are satisfied, though some exceptions are possible (Maslow, 1943). This theory is based on two essential pillars: Human needs follow a hierarchical pattern, and there is a dynamic between them. Thus, the motivation to satisfy a higher-order need should exist only if lower-order needs already have been satisfied.

Expansions of Maslow's framework generally propose similar classification patterns. For example, Herzberg et al. (1959) rely on Maslow's description of the hierarchy of needs and divide motivations into hygiene factors (i.e., company policy, relationship with peers, or security) and motivator factors (i.e., achievement, recognition, responsibility, advancement). For these authors, only the latter are true motivators, because the hygiene factors actually cause de-motivation if people lack them, whereas their presence does not exert an effect in motivational terms. McClelland (1962) also identifies three types of needs (achievement, power, and affiliation) that prompt three associated motivations. Finally, Alderfer's study (Alderfer, 1969) extends Maslow's theory by categorizing human needs into three types: existence, which comprises Maslow's basic needs; relatedness, encompassing Maslow's social and external self-esteem needs; and growth, which connects closely with Maslow's internal self-esteem and self-actualization needs. All in all, these theories provide a greater general understanding of human motivations and needs, for both managerial research and practice.

Another important description of human motivations comes from Herzberg (1968), who distinguishes extrinsic from intrinsic motivational factors. When people are intrinsically motivated, they experience interest and enjoyment, feel competent and selfdetermining, and hold an internal locus of control, such that they perceive themselves as the masters of their destinies and outcomes, through their behavior. Conversely, when people are extrinsically motivated, they need external factors such as money or verbal support to motivate them to act. Thus, intrinsic factors are inherent to the person/actor, whereas extrinsic factors are facilitated outside of the person/actor (Heath, 1999).

Again, others have developed and enriched Herzberg's (1968) intrinsic–extrinsic theory with subsequent research. For example, self-determination theory (Deci and Ryan, 1985; Ryan and Deci, 2000) relies heavily on the concepts of intrinsic and extrinsic motivation, such thatcompetence (to succeed in difficult tasks and be able to achieve expected results) and autonomy (to have the ability to choose) needs are described as intimately related with intrinsic motivations, whereas relatedness (to establish a sense of mutual respect and trust with others) is classified as an extrinsic type of motivation.

As summarized in **Table 2**, consumers then might be encouraged to participate in co-creation activities to attain financial rewards or acquire useful skills for career advancement purposes, as well as personal relationships and social capital resources that help construct their own identity (e.g., Zwass, 2010; Roberts et al., 2014). Consumers might also participate if they believe doing so facilitates their access to social standing and reputation (Zwass, 2010; Chen et al., 2012). From this activity, they can learn what they appreciate most (Wasko and Faraj, 2000; Zwass, 2010), and in many cases, participation appears valuable in itself, enabling them to meet their self-esteem, selfefficacy, and self-expression needs (Bandura, 1995; Kollock, 1999). Furthermore, their interests might be motivated altruistically (Kollock, 1999; Roberts et al., 2014) or because of feeling joy and enjoyment (i.e., hedonic motivations) by doing what they love (Nambisan and Baron, 2009; Roberts et al., 2014). In general, our literature review thus elucidates various different needs that lead consumers to co-create, which can clearly be grouped into the higher–lower or intrinsic–extrinsic motivational taxonomies (see **Table 3**).

In **Table 3** we connect Maslow's (1943) and Herzberg's (1968) classical taxonomies of human motives, in line with Guillén et al. (2014), and our deep review of literature on consumers' motivations to co-create. That is, in the columns of **Table 3** we document extrinsic and intrinsic motivations, according to Herzberg's (1968) distinction, and in the rows, we present higherand lower-order needs, according to Maslow (1943). On the one hand, intrinsic motivations relate to the nature of the activity itself and are rooted in the personal satisfaction that can be achieved by performing the activity (Kozinets, 2002), but extrinsic motivations are utilitarian in nature and associated with attaining external, functional, and practical incentives, distinct from the activity per se (Daugherty et al., 2008). On the other hand, lower-order needs are associated with preserving physiological, subsistence (i.e., food, water), safety, and security (i.e., safety, security) needs, whereas higher-order needs have to do with social activities (i.e., love, esteem) and meeting selfactualization aspirations. When consumers who participate in co-creation activities are motivated extrinsically to meet lowerorder needs, practical purposes are their real motives (e.g., financial rewards, career advancement; see **Table 3**). When they are motivated intrinsically and seek to meet lower-order needs, they really participate in co-creation activities for practical purposes related to learning and enjoying the personal hedonism they derive from co-creating new and unique goods. With regard to higher-order needs, consumers often focus on relatedness and likely participate in co-creation activities for extrinsic motives (e.g., access to social capital, personal relationships, identify with co-creating communities and projects, gain social standing, and renown). Finally, to meet higher-order needs, self-esteem, selfefficacy, and self-expression can prompt intrinsically motivated consumers to participate in co-creation activities. Thus, given that both intrinsic and extrinsic motivational aspects play a role

#### TABLE 2 | Consumers' motives for participating in co-creation processes.




Source: Based on Guillén et al. (2014).

in explaining consumers' willingness to engage in co-creation activities, we propose:

Proposition 2A: Consumers engage in co-creation activities to receive external goods, beyond performing the activity itself, reflecting their extrinsic motivation.

Proposition 2B: Consumers engage in co-creation activities to receive internal goods related to performing the activity itself, reflecting their intrinsic motivation.

Although this integrative revision of consumers' motivations to co-create is new to extant literature and offers a clearer general understanding of this issue, some necessary elements are still missing. According to Guillén et al. (2014), both types of motivation classifications (i.e., higher–lower and extrinsic– intrinsic) need to expand to include perspectives that reflect the moral content of motivation, as it relates intrinsically to human life. Thus, even the expanded, integrated view in **Table 3** is insufficient for explaining consumers' participation in co-creation activities, especially in the new era in which community members freely share innovative ideas and content to facilitate and leverage value co-creation. There must be other motives, beyond those described in **Table 3**, which can help us understand why the process of participation in co-creation activities has become so prominent. We posit that these motives revolve around ethical and transcendent issues. In particular, modern consumers devote increasingly more emphasis to ethical values in their purchase decisions (e.g., human welfare, social justice, environmental factors; Shaw et al., 2005). Creative people also might participate for spiritual motives rather than to attain material benefits (Kotler et al., 2010), so consumer participation in value co-creation activities likely aims to improve the usefulness, value, and service of the new product to society. Thus, consumers' willingness to participate is based not only on intrinsic or extrinsic motivations but also on transcendent motives, including the benefit their collaborations have for others in wider society.

#### Ethics and Transcendence in Consumers' Motivations to Co-Create

The service-dominant (S-D) logic (Abela and Murphy, 2008) places strong emphasis on service and the co-creation of value as essential elements for marketing area; it also has had powerful influences on marketing practitioners and researchers (Williams and Aitken, 2011). In this logic, each individual consumer is a resource-integrating, value-creating enterprise (Vargo and Lusch, 2008) that companies must motivate by embedding their business actions in line with the value-laden societal context (see **Figure 1**). To encourage consumers' participation in cocreation activities, businesses need to behave in accordance with the values that motivate those consumers (Williams and Aitken, 2011). In the modern era, ethics is one such value. That is, in the era of Marketing 3.0., consumers look to products and services not just to meet their needs but also to achieve their spiritual and moral interests and needs (Kotler et al., 2010). Increasingly, consumers are more and more concerned about the effects of their purchase choices, both for themselves and for the world around them (Harrison, 2005). Accordingly, they increasingly look for solutions for their own concerns about how to make the

global world a better place, such that they are guided by ethical values in their purchase decisions (Shaw et al., 2005; Hollenbeck and Zinkhan, 2010). For example, the extent to which a product provides freedom of choice, independence, and curiosity are key assessments, and the brand needs to inspire a sense of benevolence, security, equality, environmental friendship, and rules conformity (Shaw et al., 2005).

This ethical axis governing consumers' purchase decisions also offers a proxy for consumers' growing concerns about society's welfare. In modern environments, consumers' decisions to participate in co-creation activities evolve mainly according to their altruistic desire to contribute (Kollock, 1999; Zeityln, 2003; Zwass, 2010; Roberts et al., 2014; see **Table 1**). Thus, transcendent motives based on others' rather than self-interest (Pérez-López, 1998; Guillén et al., 2014) can prompt participation in co-creation activities. As originally proposed by Pérez-López (1998), transcendent motives place concern for others' needs and a sense of service in central positions, prompting people to shift from self-interested to others' perspectives. Consumers guided by these motives, through their actions, likely seek to satisfy others' needs rather than their own, similar to the way that people guided by ethical values such as solidarity, service, or altruism might be inclined to do. Furthermore, transcendently motivated people worry about others' authentic (moral) human goods (Melé, 2009: 207), such as truth, beauty, work, friendship, life, and human dignity (Melé, 2009); have a sense of stewardship; and aim to transcend the individual domain and take the impact of their actions on others, both known and unknown, into account.

According to prior literature, this transcendent dimension, with its close link to ethics, offers a compelling, new, and enriched framework to understand human motivations (Argandoña, 2008; Pastoriza et al., 2008; Melé, 2009; Ferrero and Calderón, 2013; Guillén et al., 2014). As such, and with the recognition that the Marketing 3.0 depends on Web 3.0 platforms and is based on values, it is useful and appropriate to address the motivations of consumers to engage in co-creation activities in this setting. Again following Guillén et al. (2014), we propose expanding the traditional taxonomy of motivations to include this new, third type of motive, beyond intrinsic and extrinsic ones. As we show in **Table 4**, Herzberg's (1968) motivational framework now is enriched with the incorporation of motivations specifically oriented toward being concerned about others' welfare, in the form of transcendent motives.

Applied to a consumer context, and specifically to engagement in co-creation activities, a transcendent motivation might mean practices such as collaboration, cooperation, help, and service, in an effort to meet lower-order needs. Consumers might engage in these practices during the realization of co-creation activities for practical reasons related to their needs to grant knowledge, experience, skills, or competencies that they have acquired and believe might be useful to others. Consumers also might cooperate in the hope they can receive help when they are in need. However, when consumers engage in co-creation activities to meet their higher-order needs, the collaboration is usually regarded as an end in itself. The consumers think authentically about transcending their personal sphere to acknowledge the welfare of wider society and contribute to the common good. In contributing to the common good, consumers seek to bring about a society that features, for example, unpolluted air, social cohesion, educational goods, environmentally friendly products and services, and healthy and practical offerings (Melé, 2009). Thus, consumers can fulfill their spiritual needs and develop their human side when participating in co-creation activities, which constitutes an increasingly critical demand among creative consumers today (Kotler et al., 2010).

**Table 4** thus synthesizes the several motives that consumers exhibit when they decide to collaborate and engage in cocreation activities. In this new era, consumers increasingly emphasize ethical values and seek to make purchase decisions in a conscious manner, such that they think carefully about the environmental, ethical, and social costs (Beagan et al., 2010). Our extended taxonomy, following Guillén et al. (2014), incorporates a transcendent motivation dimension and reveals various motives arising on the scene that influence consumers' intentions to participate in co-creation activities, with special attention to ethical, transcendent ones. Although the influences of these motivations vary in strength, they are complementary, supplementary, and potentially simultaneous in both time and action. In the modern era of ethical consumption (Harrison, 2005; Shaw et al., 2005; Beagan et al., 2010) and Marketing 3.0 (Kotler et al., 2010), the transcendent motives occupy perhaps the most critical role in encouraging consumers to participate in co-creation activities. Consistently, we propose:

Proposition 3: The ethical values–driven Marketing 3.0 era boosts consumers' transcendent motivations in relationships and decision making about brands and companies (e.g., purchases, collaboration).

Proposition 4: Consumers engage in co-creation activities to meet their need to give to others, care for others' welfare, reflecting their transcendent motivation.

#### Consumer Co-Creation Processes Direct Effects on Consumers and Companies

In seeking new ways to create consumer value, current marketing developments, such as the S-D logic, may prove especially useful. The S-D logic is based on the premise that companies do not deliver value but rather work out value proposals. Consumers themselves individually create value by using or consuming products and services. This new approach also emphasizes that the customer's participation in the product and service experience is indispensable for creating more value, such that both consumers and company employees are active participants


Source: Based on Guillén et al. (2014).

in the creative process. Two further elements are implicit to this process and should be fostered by companies: consumer empowerment and consumer engagement. Both elements have been addressed repeatedly in co-creation literature as essential to allow for the process to flow and generate positive outcomes for both consumers and companies.

Consumer empowerment implies that the company delegates power to consumers to co-create new products and services (Zimmerman and Warschausky, 1998; Cova and Pace, 2006; Füller et al., 2009). This delegation is increasingly possible in the new era; due to new technologies, consumers have been enabled to interact with the world on various levels (e.g., personal, dyad, group, community), as well as observe and experience distant things as if they were real (Kozinets et al., 2004). The Internet and new Web 3.0 platforms provide consumers with an accessible medium to express their opinions and observations about purchase decisions and product/service characteristics; to share knowledge with companies; and to engage in new product or services designs, in meaningful, challenging new product development tasks or other product-related activities (Füller et al., 2009).

Accordingly, consumer empowerment strengthens consumers' perceptions of their own self-determination and self-efficacy (Füller et al., 2009). Thanks to advances in new technologies (Web 3.0; Cova and Pace, 2006), consumers increasingly combine their resources and skills with others' resources to create virtual spaces in markets, where they exert powerful influence, establish their credibility, and can develop their own identity (Cova and Dalli, 2009; Füller et al., 2009). The Internet and new ICTs also allow consumers to interact with others, role play, test their social skills (which strengthens their sense of self-identity), and enjoy mastery experiences (which increases self-efficacy perceptions) (Hamburger et al., 2008; Füller et al., 2009). In turn, other positive outcomes for consumers and companies are likely. First, in terms of the effects for companies, consumer empowerment may cause consumers to perceive that the brand that has assigned them more power produces higher quality products and services, leaving them more motivated and committed to co-creation activities (Zeithaml and Bitner, 2003), as well as the related brands and companies. Second, regarding the effects on consumers, the perceived quality of consumers' own contributions to the co-creation process should enhance their satisfaction, with both their own services and the tasks (Kelley et al., 1990; Vega-Vazquez et al., 2013). When these positive emotional experiences occur repeatedly, consumers' loyalty to the focal brands and companies gets reinforced, such that a virtuous cycle initiates (Lam et al., 2004).

Consumer engagement is also essential to co-creation processes, and those processes may vary depending on its level (DeFillippi and Roser, 2014). Patterson et al. (2006) define consumer engagement as the level of physical, cognitive, and emotional presence the consumer devotes to a relationship with an organization, involving vigor, dedication, absorption, and interaction. Such traits are fostered by Web 3.0 platforms that enable consumers to share, socialize, learn, advocate, and co-develop (Brodie et al., 2013). As a result, high-quality relationships (consumer-to-consumer and consumer-tobusiness) arise from continuous dialogue (Prahalad and Ramaswamy, 2004; Jaworski and Kohli, 2006; Auh et al., 2007), with important and valuable outcomes for both consumers and companies.

Specifically, through the social relationships established in the co-creation process, consumers engage easily in dialogue with others in each stage of the product design or delivery process (Payne et al., 2008), which induces consumer learning. This dialogue during the co-creation process encourages shared emotions, behaviors, and knowledge (Payne et al., 2008), resulting in interactive processes of learning (Ballantyne, 2004). Through virtual experiential interactions and encounters, consumers perceive that their engagement in the co-creation activities ensures the utilization of their own personal resources (Payne et al., 2008), while also helping them improve and reach higher levels of learning and knowledge, because with others, they have the opportunity to create value through customized and co-produced offerings. Co-creation processes enable them to communicate directly with one another and share their experiences, which also can lead to personalized interactions, depending on how each consumer prefers to interact with the company (Prahalad and Ramaswamy, 2004). As a result, consumers' co-creation activities likely boost consumer satisfaction with the co-creation process and their maintained social relationships (Bowden, 2009), which then becomes the seed of social capital within the social community (e.g., Nuttavuthisit, 2010; Linuesa-Langreo et al., 2015) and ultimately should boost consumers' creative thinking (e.g., Füller et al., 2009; Ramaswamy and Gouillart, 2010).

With regard to the benefits for companies, several studies show that consumer co-creation processes increase consumer trust in the community setting in which their social relationships have developed (e.g., Casalo et al., 2007; Füller et al., 2009; Hollebeek, 2011; Brodie et al., 2013), as well as consumer loyalty (e.g., Andersen, 2005; Casalo et al., 2007; Schouten et al., 2007) and commitment to the social community (e.g., Chan and Li, 2010). The involvement of these external agents in the cocreation process, who work as partial employees (Dong et al., 2008), leads to new product ideas that can satisfy new and emerging needs, as well as lowered product development, design, and marketing costs (Ramaswamy and Gouillart, 2010). Because consumers prefer not to have products and processes imposed on them, co-creation processes help differentiate the company from competitors, because they enable consumers to co-construct to suit their own contexts and needs (Prahalad and Ramaswamy, 2004). The important benefits from encouraging consumers to engage in co-creation processes thus entail, for example, cost effectiveness, risk reduction (Prahalad and Ramaswamy, 2004), and differentiation (Ramaswamy and Gouillart, 2010).

Of all the benefits for companies though, consumer brand loyalty is perhaps the most well documented one, and are defined as a "deeply held commitment to re-buy or repatronize a product/service consistently in the future" (Oliver, 1999: 34), Brands usually evoke emotional and symbolic issues (Aaker, 1999), including two basic elements—authenticity and sincerity—that largely define the image that consumers develop about a specific company (Keller, 1993; Kaplan and Haenlein, 2010) and the affect and loyalty they devote to a brand (Aaker, 1999). Brand perceptions result from consumers' relationships with brands, and these relationships mimic interpersonal relationships (Fournier, 1998), so satisfactory cocreation processes logically should cause perceptions of brand authenticity and sincerity. Co-creation actively seeks to facilitate interactions between consumers and companies, so creating and maintaining an authentic, open dialogue and incorporating consumer needs in new product or service development processes should enhance brand authenticity perceptions. If the co-creation process also gets communicated to the general target group with a sincere storyline, brand sincerity perceptions should grow stronger (Dijk et al., 2014). Co-creation creates such close consumer–brand interrelationships that consumers' brand loyalty is increasingly probable in these settings (Luo et al., 2015).

In summary, important benefits derive from new business insights into co-creation, for both consumers and companies. **Table 5** summarizes these benefits, revealing the diverse, complex value that consumer co-creation processes provide. On these theoretical grounds and arguments, we propose:

Proposition 5: When consumers engage in co-creation activities, it boosts value for both consumers and companies, in multiple forms.

#### Co-creation of Ethical Products: Effects on Consumers and Companies

Consumers are increasingly willing to integrate ethics into their product purchase decisions. By "ethical" products, we mean products that reflect one or several social, moral, or environmental principles that could influence purchase decisions. A product cannot be ethical per se, but it can be augmented by ethical considerations or attributes that are perceived positively (Crane, 2001). Thus, because ethical consumers actively "adhere" to social and environmental principles (Strong, 1996), the presence of ethical characteristics in the product might enhance consumer engagement in co-creation processes today, in the new ethical valued-driven Marketing 3.0 era. The more an object relates to these (ethical) values, the more involving it may be for consumers (e.g., O'Cass, 2001).

As noted previously, consumers' co-creation offers important benefits for both consumers and companies; we also posit that these benefits increase when the product or service they are co-creating features ethical characteristics. For example, at the company level, emerging research on brand building processes suggests that consumers' trust in brands depends on whether they perceive that brands are ethical or, more important, offer products and services that are just, honest, and trustworthy (Singh et al., 2012). Considering that altruistic and courteous behaviors prompt enhanced positive affect (Sung and Kim, 2010), a brand that is perceived as ethical will elicit positive emotional responses among its consumers and invoke a stronger level of brand affect among them (Glomb et al., 2011). At the consumer level, when co-creation processes involve ethical products and services, consumers gain more value from their participation. In parallel with findings that reveal that committing moral deeds creates a sense of purpose, meaning in life, and relative gains in happiness (Hoffman et al., 2014), consumers who co-create ethical products and services likely experience good feelings and increased consumer satisfaction, as well as better, more personalized co-creation experiences. Also, research reveals that long-term, supportive collaborations are likely to be enhanced when the decisions that each party to the relationship makes evoke perceptions of fairness or ethicality (Ruiz-Palomino et al., 2013).

Brand loyalty tends to apply to ethical companies because it is based on a consumer commitment for future transactions. Thus, when consumers perceive fairness in a company's service or product transaction, as well as in the process for handling customer claims, repurchase intentions grow (Hellier et al., 2003), which in turn may increase loyalty to the company. Customer– brand relationships cannot be sustained in the face of ethical misconduct by the company (Roman, 2003; Huber et al., 2010). In view of these arguments, the consumer co-creation process for developing ethical products should lead to increased positive benefits for both consumers and companies. Thus we propose:

Proposition 6: The value of consumers' engagement in cocreation activities for both companies and consumers is higher when co-creation involves ethical products.

### INTEGRATIVE MODEL OF THE CO-CREATION PROCESS, ITS CAUSES AND ITS EFFECTS

Reflecting our integrative literature review and the resulting theoretical propositions, we developed the integrative model of the causes and effects of consumers' engagement in cocreation activities in **Figure 2**. This model incorporates both theoretical and empirical contributions from prior literature

#### TABLE 5 | Positive effects of co-creation.


and seeks to affirm a better understanding of the process of consumer co-creation, both is antecedents and consequences. Furthermore, it includes ethics as an essential element. As described in the previous sections, several areas of research provide the foundation for this proposed model, including consumer co-creation (e.g., Prahalad and Ramaswamy, 2004), Web 3.0 platforms (e.g., Kalaignanam and Varadarajan, 2006), motivation (e.g., Maslow, 1943; Herzberg, 1968; Deci and Ryan, 1985; Pérez-López, 1998), ethical product characteristics (e.g., Crane, 2001), and the positive effects of co-creation on both consumers and companies (e.g., Prahalad and Ramaswamy, 2004; Füller et al., 2009; Ramaswamy and Gouillart, 2010; Brodie et al., 2013; Luo et al., 2015).

The first part of the model depicts the essential elements that initiate and influence the co-creation process. In particular, the development of the Internet and ICTs allows companies to provide Web 3.0 platforms and helps consumers interact widely and easily with both other agents and the company to cocreate new products and services (P<sup>1</sup> in **Figure 2**). These new interactive spaces also support strong relationships and a sense of social community, through the easier, increased interactions with others, learning processes, and social links. This model also includes motivational factors that might lead consumers to engage in co-creation activities. Extrinsic motivations (financial rewards, personal relationships, identity construction, social standing, and renown) and intrinsic motivations (hedonic factors, learning, self-esteem, self-efficacy, self-expression) are well-known from prior literature (P2A and P2B in **Figure 2**). However, in modern society, with its focus on ethical and social values, consumers seek out companies that offer products and services to address social, economic, and environmental problems (Kotler et al., 2010). Creative people also prefer to follow spiritual, social, or ethical motives, rather than simply attaining material or personal goals (Kotler et al., 2010). In this sense, consumer participation in value co-creation activities is more likely to reflect transcendent motivations to improve the usefulness of products or services and offer value to society (P<sup>4</sup> in **Figure 2**). In a values-focused era, ethical values provide guides for consumers, and ethical values, such as sharing knowledge, experience, skills, or competencies or contributing to the common good, can underlie the transcendent motivations of consumers to engage in co-creation activities (P<sup>3</sup> in **Figure 2**).

The second part of the model highlights the positive effects of co-creation processes on both consumers and companies (P<sup>5</sup> in **Figure 2**). Value co-creation thus is based on interactive, social processes promoted by consumers and companies, in which valuable resources are integrated and value is distributed among agents (Prahalad and Ramaswamy, 2004). Thus, consumers work as partial employees who adhere voluntarily to the inspiring community project, promoted by companies. Two core outcomes for consumers derive clearly from these processes: consumer empowerment, spanning self-determination and selfefficacy perceptions, and consumer engagement, which produces access to social relationships, creates effects that are positive in and of themselves (i.e., consumer satisfaction with the service, consumer learning, personalized experiences of co-creation), and sparks synergic effects and benefits for the social communities and companies involved (trust, commitment, loyalty). The model also reveals that companies benefit from cost effectiveness (consumers as partial employees), risk reduction (Ramaswamy and Gouillart, 2010), market differentiation (Prahalad and Ramaswamy, 2004), and brand loyalty (Kim and Slotegraaf, 2015; Luo et al., 2015). Finally, the entrance of ethical products on the scene of co-creation processes causes the previously described positive effects, for both consumers and companies, to increase substantially (O'Cass, 2001; P<sup>6</sup> in **Figure 2**).

The model highlights the important value generated by consumers and companies. Companies should leverage the benefits of fostering consumer empowerment and engagement in co-creation processes. Because we increasingly are moving toward a world in which value results from an implicit negotiation between the individual consumer and the company

(Prahalad and Ramaswamy, 2004), the co-creation of value with consumers is a new business model that companies should embrace, so that they can compete in an efficient, differentiated manner (Prahalad and Ramaswamy, 2004; Ramaswamy and Gouillart, 2010). Although continuing improvements to Web 3.0 platforms will help feed the consumer co-creation process (dashed lines in **Figure 2**), they are useless without consumers' intentions to engage in future co-creation activities. Such intentions can be fostered if consumers perceive that they will gain multiple forms of benefits from engaging in co-creation activities, as well as insofar as they participate more in such processes, which should give them confidence in their ability to complete tasks and participate in value co-creation, lead them to take ownership of the activities, reduce their risk perceptions, and allow them to enjoy the whole experience more (Dong et al., 2008). A feedback process that indicates how, ceteris paribus, increased consumer participation in co-creation causes the process to continue is also represented in the model (dashed lines in **Figure 2**).

## CONCLUSIONS

Our research represents a relevant contribution to the fields of business and marketing, specifically to the area of consumers' co-creation value. Our first contribution is the valuable synthesis of representative literature and expanded, diversified knowledge that we have gained around the co-creation process involving consumers. Through an integrative literature review methodology (Torraco, 2005; Yorks, 2008), we have developed a conceptual model to describe both the antecedents and the positive outcomes of consumers' engagement in co-creation processes. Based on ethical theory, we have provided a fresh new understanding by offering sound support to the role that ethical values and transcendent motives play in boosting consumers' cocreation processes, which fills a missing gap in afore-mentioned literature review.

The second important theoretical contribution refers to the advances that our article represents for the better understanding of the reasons why consumers engage in value co-creation activities with companies. Specifically, our study incorporates transcendent motives as a complement and extension to the classical intrinsic–extrinsic motivation theory (Herzberg, 1968) to understand this process. Although abundant literature emphasizes intrinsic (i.e., hedonic motivations) and extrinsic needs (i.e., financial rewards) as responsible for consumers' engagement in co-creation activities, our literature review elucidates other important, transcendent needs in this area. Consumers' willingness to participate is based not only on intrinsic or extrinsic motives but also on the need to benefit, through their collaborations, third parties in the wider society, which corresponds to the new ethical values-driven Marketing 3.0 era.

Our third contribution refers to the adequate specification that our model makes about the influential roles of new technologies—web 3.0 platforms—and the Marketing 3.0 paradigm to better understand the process leading consumers to engage in value co-creation processes. In these new contemporary times, consumers increasingly are emphasizing ethical values and seeking to make purchase decisions in a conscious manner, by carefully thinking about the ethical, environmental, and social costs derived. Accordingly, the way consumers are conceived and seen must be changed in order to adapt to the new times successfully. Consequently, the ethical values-driven Marketing 3.0 paradigm, which has recently emerged, occupies an essential role to understand the process of consumers' engagement in value co-creation processes. Under the umbrella of this paradigm, companies are seen as entities that must treat consumers as human beings, with intelligence, heart, soul, and spirit, as well as aspire to live such ethical values as cooperation, friendship, and human welfare. Therefore, as a result of consumers perceiving this congruence in ethical values with companies, the co-creation process is helped to start and develop over time, which is clearly identified in our conceptual model. Of course, to implement and follow the Marketing 3.0 paradigm and thus promote collaboration with consumers adequately, companies need the new advances in ICTs. Web 3.0 platforms, and its continuous improvements, thus, offer new tools that consumers can use to interact with companies and other agents, as well as incentives for creating new product and services.

Finally, we importantly contribute to literature by identifying the positive outcomes and the process by which these are obtained when consumers' engagement in co-creation activities occurs. In this sense, one important part of our conceptual model refers to the specific positive outcomes that these activities entail for both companies, and importantly, consumers, which aims to fill a void in literature. Our integrative review of literature revealed that these activities might result in important, very valuable outcomes for companies (i.e., consumer loyalty) and consumers (e.g., consumer satisfaction), but depend on whether the co-created products or services feature ethical characteristics. The presence of ethical characteristics in the product or service co-created is sensitive so as to strengthen the value created for both consumers and companies. Because one important motivation for co-creation is the value which is created for others, when co-creation processes involve ethical products and services, consumers tend to gain more value from their participation. In parallel with findings that reveal that committing moral deeds creates a sense of purpose, meaning in life, and relative gains in happiness, consumers who co-create ethical items are more likely to experience good feelings and satisfaction as well as better personalized co-creation experiences. Also, these consumers are expected to increase their trust, commitment and loyalty to the brand and the company with which they are collaborating. All these positive outcomes would also be fostered insofar consumers perceive they gain multiple forms of benefits from engaging in cocreation activities, as well as they will participate more and more in such processes.

All in all, our integrative review of literature offers new understandings around the engagement process of consumers in co-creation activities, and has given rise to some interesting conclusions that should be well considered in business and marketing management. Companies must collaborate with consumers, who play a key role in generating value and competitive advantages by providing information, fresh ideas, and co-creating new, improved products and services. They are sources of creativity as well as sources of social, ethical values imprints in the product design, and development processes, which is imperative to be successful in the new contemporary times. The consumer role has evolved so much that today consumers are now described as active agents, protagonists, or value co-creators. These roles also converge to describe not just now actively and constructively consumers are today but also the importance of their market experiences, joint activities, and relationships with companies. That's why managers should acquire an integral understanding of the antecedents making consumers engage in value co-creation activities. With this in mind, our integrative review allows us to highlight the important role played by that instrumental devices (i.e., Web 3.0 platforms) and personal mechanisms (i.e., intrinsic, extrinsic motives) in fostering these activities. However, new to literature, special emphasis has been laid on ethical values and, specifically, on personal transcendent motives as antecedents. Modern society is increasingly focusing on ethical values, along with consumers seeking companies that offer creative products, services that truly solve the current social, economic, and environmental problems. Humanity, morality, and spirituality are common elements behind these creative solutions (Zohar, 1990). In this vein, to encourage consumers' engagement in co-creation activities and creativity, companies' alignment with ethical and transcendent values should not be obviated, nor the design and development of ethical, responsible items.

### IMPLICATIONS FOR RESEARCH AND PRACTICE

For academic researchers, our theoretical model and integrative literature review should serve as stimulants of further studies on the topic, in that they provide a reference or starting point for additional research. Our findings demonstrate the need for understanding the causes and effects of consumers' participation in co-creation processes, as well as the positive effects for both consumers and companies. The findings further suggest that when consumers feel empowered, with passion and a sense of ownership, they are willing to contribute extensively for the benefit of the company. Thus, companies need to learn about their consumers' desires and needs, beyond normal exchange processes, especially considering the importance of ethical and social values for consumers today.

For managerial practice, various implications emerge from this study. The first one is related with the potential value which consumers who are willing to participate in value co-creation activities. Managers should assess the potential value of their communities of proactive customers for a greater innovation, brand loyalty, differentiation, and augmentation of competitive advantage to their companies. Consequently, organizations should take a long-term view of their customer relationships, rather than a short-term financial perspective. Within a shortterm approach, companies are unlikely to boost consumers' co-creation participation; consumers need to perceive their relationship with the company as equitable, which takes time. A long-term perspective is also more suitable, considering consumers' increasing interests in ethical and social values. It might ensure that consumers' transcendent motives align with companies' interests to contribute, ultimately, to the general welfare of society.

One second implication is related to the necessary commitment of the top management team with favoring these co-creation processes, and thus allow for these processes to develop optimally. Managers who seek new ways to involve consumers in co-creation activities should institute cultural changes in their organizations. Co-creation initiatives require flexible organizational structures, oriented toward engaging in frequent contacts with consumers and meeting their needs. This view is important to build emotional bonds between consumers and companies as well as to encourage consumers' engagement in co-creation activities (Grayson, 1998). Accordingly, managers should take this commitment in mind when planning, organizing, leading, and monitoring the activities of their employees.

One third implication is the creation and maintenance of relevant communication channels with consumers through platforms such as Web 3.0. These efforts can translate into enhanced encounters, supporting cognition, emotion, and action-based learning for both consumers and companies, and thus result in a proactive community that fosters consumers' loyalty to the community and to the company. Although consumers' engagement in value co-creation activities strengthen these ties, and might enhance the level of community commitment on itself, the more companies invest and develop efforts in connecting consumers, the better to retain on loyal both to the community and the company. The design of good interaction channels with consumers is thus an important element to implement the value co-creation strategy, and thus make the creative process initiated with consumers work properly.

Finally, companies should be cautious about the negative potential consequences of empty value co-creation strategies. Due to the low cost and popularity of social media, many companies almost blindly take for granted advantages to initiate various value co-creation activities. However, these activities are not just about interacting with customers and managers; they rather require a strategy based on careful planning and implementation through Web 3.0 tools. Thus, managers should create policies based on ethical principles and the Marketing 3.0 paradigm to obtain a long term and effective business development. Business brands should behave ethically regardless of the potential impact on the bottom line. However, in a highly interconnected world that has made brands more transparent, truly ethical behavior of companies and customers based on transcendent motives will be necessary to succeed in any marketplace.

### LIMITATIONS AND FURTHER RESEARCH

Despite these contributions, we acknowledge some limitations and accordingly propose new avenues for research. First, empirical data are needed to validate the conceptual model and theoretical propositions. In particular, **Tables 2**–**5** synthesize findings and suggestions from prior literature; those summaries should be tested empirically in the particular setting of consumers who are participating in co-creation activities. Other moderating effects also can be examined, such as social justice perceptions or relevant contextual variables. For example, studies might analyze the effectiveness of cocreation processes when consumers perceive a balance in their relationship with the company, or else document any negative effects that arise when a consumer perceives an unbalanced relationship. Second, we focused on business-to-consumer interactions, but other beneficial relationships may also tend to arise, whether business-to-business links or relationships with government or third-party agencies. Additional research focused on understanding the co-creation processes that arise from these relationships thus could helpfully nourish from integrating the perspectives addressed here. Third, in line with our research goals, we concentrated on the positive effects of consumers' engagement in co-creation processes. However, researchers could address the specific variables that might produce negative aspects, such as negative consumer sentiments, bad brand experiences, or trust reduction. Finally, a longitudinal study would offer a more dynamic view of value co-creation and help determine how these effects evolve over time.

## AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication.

### FUNDING

This work was funded by The Ministry of Economy and Competitivity (Spain), Research Project reference ECO2014-59688-R, Programa Estatal de Investigación, Desarrollo e Innovación Orientada a los Retos de la Sociedad, Plan Estatal de Investigación Científica y Técnica y de Innovación 2013–2016.

### REFERENCES


retail spectacle. J. Consum. Res. 31, 658–672. doi: 10.1086/ 425101


Maslow, A. (1943). Motivation and Personality. New York, NY: Harper and Row.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Martínez-Cañas, Ruiz-Palomino, Linuesa-Langreo and Blázquez-Resino. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# A Perspective on Consumers 3.0: They Are Not Better Decision-Makers than Previous Generations

#### Petr Houdek1,2,3 \*

<sup>1</sup> Faculty of Business Administration, University of Economics, Prague, Czech Republic, <sup>2</sup> Faculty of Social and Economic Studies, Jan Evangelista Purkyne University, Ústí nad Labem, Czech Republic, ˇ <sup>3</sup> Center for Behavioral Experiments, Prague, Czech Republic

This perspective article builds upon the theory of local thinking in interpretation and prediction of consumer behavior in a contemporary world of information overload. It is shown that even informed and socially and environmentally responsible consumers (consumers 3.0) exhibit selective recall, limited attention, and bounded search in the perception and interpretation of price and quality of purchases. Their decisions fall into local cognitive frames, which specifically focus attention only on a narrow structure and content of the choice. The cognitive frames can be established by recent or regular purchases, but also extreme or primary purchase experiences. The article includes a short conceptual review of car, food, clothing, insurance, drugs, paintings, and other product purchases showing that the local cognitive frames often lead to bad bargains across various sectors. The article presents several suggestions for future research.

#### Edited by:

Maria Pilar Martinez-Ruiz, University of Castilla-La Mancha, Spain

#### Reviewed by:

Carmen Selva-Sevilla, Profesora de Universidad, Spain Omid Kamran-Disfani, University of Missouri, USA

#### \*Correspondence:

Petr Houdek petr.houdek@gmail.com

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 07 March 2016 Accepted: 20 May 2016 Published: 03 June 2016

#### Citation:

Houdek P (2016) A Perspective on Consumers 3.0: They Are Not Better Decision-Makers than Previous Generations. Front. Psychol. 7:848. doi: 10.3389/fpsyg.2016.00848 Keywords: anchoring, decision-making biases, cognitive frames, consumer 3.0, limited attention, local thinking

### INTRODUCTION

The aim of this perspective article is to show that even in age of an informed and socially and environmentally responsible consumer (consumer 3.0), his/her style of decision-making has not changed much. People still tend to overrate the importance of the information that is vivid, emotional, recent, or frequent and therefore easily appears in their minds, as Tversky and Kahneman (1974, p. 1127) famously wrote: "It is a common experience that the subjective probability of traffic accidents rises temporarily when one sees a car overturned by the side of the road." The immediate experience creates a specific cognitive frame under which people interpret further information. In an insurance choice the consumers thus prefer insurance for a high probability, low consequence risks, and downplay insurance for a low probability, high consequence risks (Browne et al., 2015), or that insurance take-up spikes the year after a catastrophe and then steadily declines as salience of the event fades off (Gallagher, 2014). Consumers are framed by immediate weather to buy goods with advantageous attributes in that particular weather, underestimating different future circumstances or changes in their preferences, followed by high returns rates (Conlin et al., 2007; Busse et al., 2015). Examples of other cognitive frameworks are abundant.

This article elaborates the concept of cognitive frames in consumer research to offer a simple unifying framework of the forces influencing consumer decisions in contemporary complex consumer environment (see also Olshavsky and Granbois, 2002; Uncles et al., 2002; Foxall, 2010). I utilize a behavioral theory about consumers' local thinking that has been conceptualized in a

series of articles by Gennaioli and Shleifer (2010) and Bordalo et al. (2012; 2013; 2015) and is itself inspired by a model of selective recall (Kahneman and Miller, 1986). This approach offers a fruitful conceptualization of consumer decision-making across the retail sector.

The article follows by showing that consumers do not consider all relevant information when shopping. Their behavior is better described by selective recall, limited attention, and bounded search. In the third section a theory of consumer local thinking is introduced. When demonstrating its relevance, I particularly refer to contemporary studies about consumer behavior in market settings to avoid criticism of laboratory and/or survey research. Although laboratory-based studies have proved useful in measuring many aspects of shopping cognition and behavior, they may suffer from problems of low external validity and non-replicability, their participants do not necessarily represent broader populations, and it is generally difficult to capture highly dynamic and incentivized aspects of markets in a lab (Hertwig and Ortmann, 2001; Henrich et al., 2010; Open Science Collaboration, 2015).

### LIMITED SEARCH AND ATTENTION

The cognitive frames approach of consumer decision-making is interlinked with the findings that consumers' willingness (or capability) to seek out new information, their attention and memory are not limitless (DellaVigna, 2009; Iyengar and Kamenica, 2010; Houdek and Koblovský, 2015) and consumers may not realize these aspects of their bounded rationality (but there is a great heterogeneity in consumers' shopping sophistication). Moreover, the natural reaction of firms to bounded rationality of some consumers is to make their price policies more complex and less transparent to further hamper consumers' ability to compare products (Hortaçsu and Syverson, 2004; Carlin, 2009). In many market conditions there is no reason to expect that competition among firms will push to greater price transparency, e.g., if median consumers focus only on a base price and ignore shrouded add-ons, then firms with transparent pricing lose to non-transparent firms (Gabaix and Laibson, 2006).

Consumers focus only on a small fraction of products. An average consumer is deciding between two to five cars while 160 brands of them exist; when there are flooding coffee options consumers are making a decision only between three and four alternatives – essentially regardless of the range of possible substitutes, a consumer takes into account always only a few units of alternatives (Hauser and Wernerfelt, 1990). The advent of internet search engines and comparative sites dramatically increased opportunities to choose the best deals by allowing consumers to engage in low-cost price comparisons (Brown and Goolsbee, 2002; Zettelmeyer et al., 2006), nevertheless the prices on the internet are far from converging to the law of one price.

The amount of time spent on shopping (including browsing and buying) has not changed at all in more than 30 years (Ott, 2011). Consumers are still affected by the most salient sales (Grubb, 2015). Additionally, "advances in search technology are accompanied by investments by firms in obfuscation" (Ellison and Ellison, 2009, p. 427). Ellison and Ellison show that charging a low price for a low-quality product on a price comparison website actually increases retailer's sales of medium- and highquality products on the retailer's own website. Consumers are influenced by the salient (top) positions of low-quality product on comparison website, they move to retailer own website, where they are persuaded to buy a higher-quality higher-price product instead (see also Chioveanu and Zhou, 2013 and Spiegler, 2014).

Moreover, complex pricing policies and (perceived) quality differences of many product categories do not always allow direct comparison (Greenleaf et al., 2016). Consumers must rely on limited search and are thus heavily influenced by various levels of information disclosure (Brown et al., 2010) and/or advertising, which exacerbate their confusion. Consumers' take-up of credit card offers responds much less to the interest rate on credit cards than to a photo of an attractive woman in a mailing marketing campaign (Bertrand et al., 2010). Consumers are willing to pay high premium for branded product, which is homogeneous to non-branded alternatives. In their case study of headache remedies Bronnenberg et al. (2015) found that more informed consumers (such as physicians and pharmacists) choose national brands over store brands only 9% of the time, compared to 26% of the time for the average consumer.

Consumers pay attention largely to digits that are on the left in a string of numbers. The left-digit bias creates a salient break-point, which is why items priced at 0.99 or 3.99 dollars are considered much better deals than items priced at 1.00 or 4.00 dollars. The bias is not governing only trivial purchases. At American used-car market Lacetera et al. (2012) found that vehicles with odometer values between 79,900 and 79,999 miles were sold on average for approximately 200 dollars more than cars with odometer values between 80,000 and 80,100 miles, but for only 10 dollars less than cars with odometer values between 79,800 and 79,899. Such focal points could be found in a number of consumer choices. Scott and Yelowitz (2010) detected that prospective grooms are willing to pay premiums upward of 18% for an engagement ring's diamond that is onehalf carat rather than slightly less than a half carat and between 5 and 10% for a one-carat rather than a slightly less than one-carat stone.

When price promotions occur, many consumers fail to switch to purchasing a package that has the lowest unit price, which suggests a lack of price awareness. In their study on quantity surcharges Clerides and Courty (2015) showed a decrease of only 27% in the sales of disadvantageous large laundry detergents relative to the sales the week preceding the surcharge. In the utilities sector, households rarely search for alternative retailers, and when they do search, the current retailer has a significant brand advantage (Hortaçsu et al., 2015). Nevertheless, real-time feedback on quantity of electricity consumed via an in-home display could have a substantial impact on price elasticity of households (Jessoe and Rapson, 2014). The advent of technology using big data about consumption could lead to higher price and quality transparency. It could be expected that consumers' sensitivity to price and quality features of products will grow.

### CONSUMERS' LOCAL COGNITIVE FRAMES

Cognitive frames create simplified mental models of choices. They constrain how consumers respond to stimuli by selectively organizing information from the consumers' environment and by focusing attention only on a narrow structure and content of the perceived attributes of a choice. They thus create specific expectations about a product's price or consumption quality (Gennaioli and Shleifer, 2010; Bordalo et al., 2013, 2015). For example, a classic study by Hoyer (1984) found that while choosing a laundry detergent, consumers examine on average only 1.4 packages and only 11% of them look at more than two. Hoyer concluded that 91% of the consumers were governed by a simple, one-statement reason for their choice (e.g., "I like it" tactics).

The expectations about prominent attributes of a product, or consideration sets (Roberts and Lattin, 1997; Houdek, 2016), are at the top of a consumer's mind and they lead to a particular interpretation of a choice at hand and affect the alternatives consumers consider (Tversky et al., 1988). Other aspects of the choice are reflected or recalled only partially or not at all. This neglect makes consumers leave out better products from their consideration sets.

The expectations (cognitive frames) can be created by (i) deeply remembered psychological states – strong personal or primary experiences could create a general tendency for evaluating all following products and create a long lasting consumption pattern (see Strong and Primary Memories, Attribution Error). They are further formed by (ii) recalls from immediate past or from frequent consumption (see Immediate and Frequent Consumption, Anchoring Bias); the more often or saliently consumers were experiencing some attribute of choice, the stronger cognitive frame of this choice they then apply.

The expectations could originate from (iii) transient mood and affects (see Situational Influences, Emotion, and Mood, Affect Bias) or just in other words (iv) they can be created by momentarily salient external circumstances; typically, from information supplied by marketing devices or projections of current tastes influenced by immediate context to the future (see Situational Influences and Projection Error). Moreover, an attribute of choice options that is different from the expectations is most salient and the consumer overrates it in their decisionmaking.

### Strong and Primary Memories, Attribution Error

Expectations could be derived from primary and/or extreme experiences that consumers more easily recall from memory. The most apparent display of this influence is the loyalty of a consumer to a brand, which originates along with feelings of a satisfactory purchase (Chaudhuri and Holbrook, 2001). The customer is then attracted to the brand because of the trust, nostalgia, familiarity, or resistance to change. The favorite brand is prominent in the consumer's mind in a purchase, and other brands may not be considered relevant (Smith and Brynjolfsson, 2001). The consumer must be persuaded to even consider their properties (Eliaz and Spiegler, 2011). Past consumption thus strongly impacts the current one and preference of a certain brand can persist in the long term (Bronnenberg et al., 2012).

In the same way some kinds of consumption can influence a consumer's satisfaction even long after the purchase had been made. Memories of it influence the consumer, either positively or negatively, in the following decisions, as they bring an emotional context of the previous purchase into the current decisionmaking, leading to attribution error. For instance, honeymoon, holiday, or recreation can steer the long-term choices where to go to a new holiday and how to spend it, because the pleasant nostalgia or conversely strong regret will be very prominent in the consumer's mind. As Simonsohn (2010, p. 272) noted: "[F]oods tasted for the first time on an empty stomach are remembered as more enjoyable and might hence be disproportionately likely to be purchased again."

Strong primary experiences can produce a similarly strong impact. Children learn consumer behavioral patterns and preferences from their parents, either by direct imitation or shared preferences (Viswanathan et al., 2000). Primary experience can impact car purchases: there is a strong correlation between the brand of cars purchased by parents and their children, even after controlling for shared demographic or geographic characteristics (Anderson et al., 2013). This correlation is strongest for cars purchased at the time when the children had still lived together with their parents. Anderson et al.'s (2013) findings mean that prices of cars for young families can be lower, as the car company creates loyalty in a future generation of customers. Most other studies about intergenerational preference transfer concerns food (Birch, 1999), but research into this phenomenon in other kinds of goods is virtually non-existent.

### Immediate and Frequent Consumption, Anchoring Bias

People are anchored by their immediate or repeated past consumption and/or experience (Simonson and Tversky, 1992). A study (Simonsohn and Loewenstein, 2006) showed that families moving from areas with high rent into cheaper areas spend more on their rent and rent larger homes on average. In contrast, families moving from cheaper into more expensive areas tend to spend less and live in smaller homes (both findings are robust after controlling for the effects of wealth, taste for expensive housing, or mis-estimation of local housing costs etc.). Even after moving, the families were still influenced by their past in their decision about how much they would spend on rent. A similar effect has been found in commuting time; the longer one had commuted in their previous home, the longer they commute in their new location as well (Simonsohn, 2006).

The same anchoring effect, where assessment of a choice is influenced by recent contextually relevant experience, can be found in a number of other situations. Paintings previously sold on "hot" markets for high prices usually continue achieving higher prices in auctions than comparable paintings previously sold in "cold" markets (Beggs and Graddy, 2009). Sellers of houses bought on a growing market, currently facing nominal loss due to

unfavorable market situation, keep their houses on the market for a longer time and higher asking price, and eventually sell them for a higher price than sellers who had initially bought comparable houses for lower prices (Genesove and Mayer, 2001). Consumers' willingness for a purchase can also be affected by completely irrelevant anchors present at the moment of a decision about a purchase (Ariely et al., 2003), however, also see Maniadis et al. (2014) for the criticism of these findings. However, there are not many studies about the dynamics of anchor updating or about particular shopping features, which could (or couldn't) sway consumers to succumb to an anchor.

### Situational Influences, Emotion, and Mood, Affect Bias

The cognitive frames could be derived from external circumstances and internal psychological states that specifically enhance certain attributes of the product and attract attention to it. As mentioned in the introduction, people prefer insurance against events that come to their mind more easily, not necessarily against events that might have a significant impact on their wellbeing (Gallagher, 2014; Browne et al., 2015).

The local consumer thinking approach can build upon the research of the influence of affect and mood on consumer decision-making (Cohen et al., 2008). People under different emotional states interpret information differently, exhibit different preferences and salient memories, creating specific cognitive frames influencing their behavior (Schwarz, 2000). Consumers' decision-making is in tow of emotion and visceral factors such as satisfaction, hunger, fear, or exhaustion.

Weather changes are often used as a research tool of the impacts of mood on consumer behavior: during sunny weather, people are happier, while in cloudy conditions, their mood is lower. Restaurants receive worse ratings in rainy weather or heat waves (Bakhshi et al., 2014), sunny weather correlates with higher tips (Rind, 1996) and higher willingness to shop using one's mobile phone (Reinaker et al., 2015). Results concerning the impact of mood on consumer decision-making show the importance of situational influences on the consumers' mood (Cohen et al., 2008). There are currently a number of mobile apps tracking or triggering emotional changes, be they weather or social apps (Kramer et al., 2014). They offer new research opportunities for a more detailed identification of the influence of emotional framing in online or mobile shopping.

### Situational Influences and Projection Error

Another example of cognitive frame is that consumers tend to exaggerate the degree to which their future tastes will resemble their current tastes, i.e., they are projecting their current tastes influenced by immediate context to the future, decision-making known as projection bias (Loewenstein et al., 2003). Using catalog orders of cold-weather items Conlin et al. (2007) find that consumers are over-influenced by current weather. If the order-date temperature declines by 30◦ Fahrenheit, the return probability increases by 3.95%. Weather had moved in a direction that made the item more valuable and/or salient at the time the consumers had been buying, but these biased valuations did not have to hold the same in the future; as a result, the likelihood of returning the item increased. Projection bias affects advance sales for an outdoor movie theater the same way (Buchheim and Kolaska, 2016). Another example is a study by Gorn et al. (1993) demonstrating that in evaluation of a speaker system, customers are influenced by the music currently playing on it.

Projection bias does not emerge only in relatively small value purchases. Days that are warmer and skies that are clearer than seasonal averages influence customers to buy more convertibles. On the contrary, surprising snowstorms increase the sales of cars with four-wheel drive (Busse et al., 2015).

## CONCLUSION

The theory of consumer local thinking is fairly simple. It states that some information is more salient for the consumers under certain circumstances, drawing disproportional attention to a narrow characteristic of the goods and making even sophisticated and responsible consumers underestimate other, often more relevant information. Future research must specifically identify which attention-drawing measures (i.e., economic, social, and/or environmental sustainability) work, how cost-effective they are and for what kinds of products and customers they work best.

The consumers' behavioral biases often lead to bad bargains, further exploited by firms to their profit (Grubb, 2015). Despite a body of literature on nudging people toward better decisionmaking (Thaler and Sunstein, 2009), there are not many real interventions successfully de-biasing consumers in mentioned inept decision-making. Nevertheless, even a small short-term attention shock can improve households' decision-making. For example, participants of a survey including questions focused on checking account overdraft fees subsequently show a lower probability of checking overdrafts in their real financial decisions (Stango and Zinman, 2014) or see (Bhargava and Manoli, 2015). Future research should examine more techniques to debiasing some consumers or address the need of regulatory interventions.

## AUTHOR CONTRIBUTIONS

The author confirms being the sole contributor of this work and approved it for publication.

## FUNDING

This research received support from the grant VŠE IP300040.

## ACKNOWLEDGMENTS

I am very grateful to Ludmila Hadincová and Marek Vranka for their help and inspiring comments and Julie Novakova for editing the English version of the manuscript. I thank reviewers for useful suggestions, which improve the quality of the paper.

### REFERENCES

fpsyg-07-00848 June 1, 2016 Time: 12:11 # 5


**Conflict of Interest Statement:** The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Houdek. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

fpsyg-07-00848 June 1, 2016 Time: 12:11 # 6

# I Will Do It If I Enjoy It! The Moderating Effect of Seeking Sensory Pleasure When Exposed to Participatory CSR Campaigns

#### Salvador Ruiz de Maya\*, Rafaela Lardín-Zambudio and Inés López-López

Marketing Department, University of Murcia, Murcia, Spain

In an attempt to gain differentiation, companies are allocating resources to corporate social responsibility (CSR) initiatives. At the same time, they are giving consumers a more active role in the process of creating value. In this sense, consumer participation represents a new approach to gain competitive advantage. However, the effectiveness of consumer participation in CSR campaigns still remains unknown. With the purpose of shedding light on this issue, this paper shows that participatory CSR campaigns lead to greater consumer perceptions of CSR, which in turn results in more favorable attitudes toward the company. Furthermore, the effect is stronger for sensory pleasure seekers, whose involvement with the experience is greater. The findings contribute to the CSR literature and reveal important implications for marketers.

#### Edited by:

Maria Pilar Martinez-Ruiz, University of Castilla-La Mancha, Spain

#### Reviewed by:

Rita Coelho Do Vale, Catolica LIsbon-School of Business and Economics, Portugal Sebastian Molinillo, University of Malaga, Spain

\*Correspondence:

Salvador Ruiz de Maya salvruiz@um.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 03 November 2015 Accepted: 02 December 2015 Published: 05 January 2016

#### Citation:

Ruiz de Maya S, Lardín-Zambudio R and López-López I (2016) I Will Do It If I Enjoy It! The Moderating Effect of Seeking Sensory Pleasure When Exposed to Participatory CSR Campaigns. Front. Psychol. 6:1940. doi: 10.3389/fpsyg.2015.01940

Keywords: CSR associations, participation, sensory pleasure, skepticism, attitude

## INTRODUCTION

In the current marketplace, companies must attain differentiation and credibility to develop strong and long-term relationships with consumers. To achieve these goals, a growing number of firms are allocating resources to corporate social responsibility (CSR; Lee et al., 2012).

Both the way consumers perceive the information on CSR and the level of stimulation this information generates influence attitudes and behaviors (Brown and Dacin, 1997; Sen and Bhattacharya, 2001). For example, inferences drawn from a company's prosocial actions can change even product evaluations (products are perceived as performing better), regardless of whether consumers are observing or experiencing the product (Chernev and Blair, 2015). A large body of research has empirically established that consumers' perceptions of firms' motives for engaging in CSR influence their evaluations of and responsiveness to CSR (Ellen et al., 2006). In general, consumers are aware that CSR can contribute to company image formation, and thus their interests in CSR activities continue to rise (Lee et al., 2012; Schmeltz, 2012). However, to some extent current approaches to CSR are still disconnected from companies' global strategy, thus masking their opportunities to benefit society (Porter and Kramer, 2006). This flaw highlights the need to shed light on the connection between CSR actions and other mechanisms in order to assist consumer persuasion.

In the past two decades, consumers have begun taking more active roles in companies' efforts to compete for and create value (Prahalad and Ramaswamy, 2000). That is, consumers are no longer passive audiences but active coproducers of value (Dong et al., 2008). Bendapudi and Leone (2003) linked high levels of consumer participation to competitive effectiveness. In support of this, extant literature in marketing has found that consumer participation has a positive effect on consumer behavior (Dellande et al., 2004; Chan et al., 2010). However, while the impact of CSR and participation on consumer behavior has been widely demonstrated in the literature, whether consumer participation in CSR activities increases the effectiveness of the latter still remains unexplored. In addition, the notion that consumers seek out pleasurable products and experiences (Hirschman and Holbrook, 1982) must be taking into account. Because participation may be associated with use of a product or going through an experience, the possibility of experiencing sensory pleasure may influence consumers' perceptions of the CSR activities in which they participate.

With the aim of shedding light on this issue, the goal of this paper is 2-fold. First, we aim to demonstrate that the participatory nature of CSR campaigns influences consumer perceptions. Second, we assess whether the dispositional trait of sensory pleasure seeking moderates this effect. The structure of the paper is as follows: We begin with a review of the relevant literature and present the theoretical background. Then, we develop a set of hypotheses and describe the method. Finally, we report the main results and discuss conclusions.

### CONCEPTUAL FRAMEWORK AND HYPOTHESES

In the past decade, researchers have shown interest in understanding how CSR activities influence consumer behavior (Marin and Ruiz, 2007; Boulouta and Pitelis, 2014). By engaging in CSR and signaling this engagement to consumers, companies can improve consumer-related outcomes (Luo and Bhattacharya, 2006). Companies can use CSR as an instrument to enhance firm image through its effects on consumers' intentions and attitudes (Brown and Dacin, 1997; Sen and Bhattacharya, 2001; Bhattacharya and Sen, 2004). That is, CSR initiatives can be central, distinctive, and enduring, thus contributing to more positive consumer evaluations of the company (Marin et al., 2009).

Consumers attribute many corporate motives to CSR engagement related mainly to company contributions to society (Ellen et al., 2006). Attribution theory states that people attribute causes to events and that their cognitive perceptions influence their subsequent attitudes and behavior (Kelley and Michela, 1980). In addition, and according to the persuasion knowledge model (Friestad and Wright, 1994), consumers accumulate knowledge on persuasive motives and tactics (Groza et al., 2011) and then use such knowledge to make inferences about firm ultimate motives. Thus, what consumers know about a company influences their associations. CSR associations reflect an organization's status and activities with respect to its perceived obligations to society and can exert different effects on consumer responses (Brown and Dacin, 1997). In summary, consumers' associations with CSR activities influence their evaluations of and responsiveness to CSR (Becker-Olsen et al., 2006; Ellen et al., 2006). Using the persuasion knowledge model and attribution theory as theoretical foundations, we posit that a CSR campaign is a persuasive attempt to create positive consumer perceptions.

### Consumer Participation

Being consumer oriented is not enough for firms to successfully compete in today's marketplaces. Firms must learn from and collaborate with consumers to create value that meets their individual and dynamic needs (Prahalad and Ramaswamy, 2000). Ulrich (1989) argues that involving consumers is a powerful way to increase consumer loyalty and commitment. The service literature lends further support to this claim, finding a positive and significant relationship between consumer participation and commitment (Bettencourt, 1997).

Previous research claims that as consumers' involvement with a firm increases, the company gains more opportunity to shape consumer perceptions (Bowen, 1986). Thus, consumers with high levels of involvement may have perceptions of quality and levels of satisfaction that differ from those who are less involved (Kelley et al., 1990). In line with this, research in marketing has underscored the importance of consumer participation (Bettencourt, 1997), or "the degree to which the consumer is involved in producing and delivering the service" (Dabholkar, 1990, p. 484). Participation can include tasks such as spending time interacting, responding to questions, or providing information on product specifications, brand preferences, and price range (Dabholkar and Sheng, 2012).

The potential of consumer participation has attracted research attention because of the assumption that when consumers participate actively, organizations can gain competitive advantage through greater sales volume, enhanced operating efficiencies, positive word of mouth, reduced marketing expenses, and enhanced consumer loyalty (Reichheld and Sasser, 1990). One stream of research focuses on the reasons consumers should engage in the service provision process and deals with the economic benefits of consumer participation (Bendapudi and Leone, 2003). A second stream also considers consumer motivations to cocreate a service, analyzing the motivation of self-service consumers (Bateson, 1985) and exploring key factors that influence initial trial decisions, consumer traits, and situational factors on technology adoption (Meuter et al., 2005). A third stream focuses on managing consumers as partial employees (Bendapudi and Leone, 2003), assuming that consumers' active participation in service provision leads to greater perceived service quality and enhanced consumer satisfaction (Dabholkar, 1990; Claycomb et al., 2001).

In addition, extant literature in marketing has found that consumer participation has a positive effect on consumer behavior (Dellande et al., 2004; Chan et al., 2010). Thus, research has shown the positive effect of participation in the areas of consumer decision making (Cermak et al., 1994; Matzler et al., 2005), brand loyalty (Bagozzi and Dholakia, 2006), commitment to the brand (Casaló et al., 2007), quality perceptions (Dabholkar, 1990), word of mouth (Kim and Jung, 2007), trust (Ouschan et al., 2006), affective commitment to the product (Atakan et al., 2014) and sensory perceptions (Troye and Supphellen, 2012). Bendapudi and Leone (2003) show that when the service outcome is better than expected, participating consumers are more satisfied than non-participating consumers. Matzler et al. (2005) report that in contexts characterized by high consumer participation, consumer satisfaction and other postpurchase responses (e.g., positive word of mouth, loyalty) are more favorable. Additionally, participation has been related to higher employees' satisfaction and performance (Yi et al., 2011). Therefore, as consumers' participation increases, subsequent outcomes become more positive. Encouraging consumer participation, then, may represent a good opportunity to gain competitive effectiveness and should deliver value to both customers and firms (Bendapudi and Leone, 2003).

### Consumer Participation and CSR Associations

Prior research has shown that firms can generate more favorable attitudinal responses from consumers when they are proactively engaged in CSR activities rather than acting reactively (Becker-Olsen et al., 2006; Wagner et al., 2009). This effect finds support in the employee participation literature, which shows that participation influences perceptions of, for example, service quality (Dabholkar, 1990). In the same vein, Bowen (1986) suggests that as consumers increase their level of involvement with a firm, the firm gains the opportunity to shape their perceptions, and Kelley et al. (1990) report that consumers with high levels of service involvement have perceptions of service quality and levels of satisfaction that differ from consumers not highly involved in the participatory role. Furthermore, Claycomb et al. (2001) demonstrate that consumer participation results in more positive perceptions of the organization and that higher levels of consumer participation in the service delivery process are associated with positive perceptions of service encounter performance. In this context, consumer participation in a CSR campaign reflects the degree to which the consumer is involved in CSR activities.

The findings on consumer participation related to a company's main activity can also apply to other activities developed by the company, such as those related to CSR. Therefore, we propose that the participatory nature of the CSR campaign will have a positive effect on consumer perceptions of CSR. We contend that consumers' participation in CSR activities will result in greater involvement, greater understanding, and deeper knowledge, which in turn will lead to perceptions of more CSR effort and, therefore, greater CSR associations (Stanaland et al., 2011). CSR associations influenced by corporate efforts depend to some degree on effective firm communication with external audiences and represent consumers' perceptions. Therefore, we propose the following:

H1: Consumers exposed to a participatory CSR campaign will have greater CSR associations than consumers exposed to a non-participatory campaign.

### Motivation for Sensory Pleasure and CSR Associations

Prior research has documented that consumers seek out pleasurable products and experiences (Hirschman and Holbrook, 1982) and show motivational differences in pursuing favorable experiences and avoiding unpleasant ones (Chapman and Chapman, 1985). The motive for sensory pleasure (MSP) describes the individual drive to seek out pleasant auditory, visual, tactile, olfactory, and taste experiences and to similarly avoid unpleasant sensory experiences (Eisenberger et al., 2010). Recently, Eisenberger et al. (2010) noted that high MSP individuals engage in greater pursuit of favorable experiences.

Moreover, personality theorists have examined dispositional differences in the enjoyment of sensory experiences (Chapman and Chapman, 1985). Thus, some individuals are high sensory pleasure seekers and others are less biased in relation to this pursuit of pleasure. According to Jackson (1984, p. 7), the highly sentient person "notices smells, sounds,sights, tastes, and the way things feel; remembers these sensations and believes they are an important part of life; is sensitive to many forms of experience; [and] may maintain an essentially hedonistic or aesthetic view of life." As a result, consumers can serve as "moderators" of pleasure through their idiosyncratic reactions to product experiences (Alba and Williams, 2013).

Prior research has shown the importance of pleasure in consumer behavior, demonstrating that emotional states (pleasure and arousal) are important determinants of purchase behavior (Sherman et al., 1997; López López and Ruiz de Maya, 2012). Fiore (2002) finds that sensory pleasure from a catalog page positively affected approach responses of global attitude. In addition, theoretical support exists for the link between pleasure and satisfaction. As Bigné et al. (2005) note, consumers who derive pleasure from an experience are more likely to exhibit positive behavioral intentions, such as positive word of mouth, satisfaction, and intention to return to the store.

However, motivation for sensory pleasure may work in the opposite direction. If consumers motivated for sensory pleasure do not experience what they are looking for (sensory pleasure), their interest in the stimulus may be low, which will imply a lower processing too (Petty and Cacioppo, 1986). In addition, while motivation for sensory pleasure is clearly related to emotional involvement (Nurse et al., 2010), Eisenberger et al. (2010) point out to the uniqueness of this personality trait as separated from need for cognition (Cacioppo et al., 1996) and, as such, subjects highly motivated to seek for sensory pleasure will base their behavior on emotions associated to the activity rather than cognitions (that require processing) related to how good the company is doing with they CSR activity it is developing. The application of this reasoning to CSR activities, therefore, leads us to propose that those who are highly motivated to seek sensory pleasure will process much less the campaign and will show less CSR associations than those who are less motivated to search for sensory pleasure. Formally,

H2: The higher the consumers' motivation for sensory pleasure, the lesser their CSR associations will be when exposed to a CSR campaign.

### The Moderating Role of Motivation for Sensory Pleasure

Personal relevance theory holds that individuals have a level of interest in and give particular importance to a cause (Antil, 1984). Therefore, when a cause is important to consumers, they will feel more interested and involved in the action. Previous research has shown that involvement significantly moderates how stimulus cues influence brand evaluation and communication effectiveness (Maoz and Tybout, 2002). More important, involvement is positively related to information processing (Leigh and Menon, 1987). Therefore, because CSR campaigns influence consumers' cognitive responses, those more involved will process the information of the CSR campaign more thoroughly and will value the social nature of the campaign more than those less involved (Gupta and Pirsch, 2006).

Consumer participation is a behavior that reflects a state of involvement (Cermak et al., 1994) that can be increased by other sources of motivation. As Eisenberger et al. (2010) argue, "high MSP individuals' enhanced motivation [will] produce greater pursuit of favorable nature experience." Accordingly, the nature of the campaign (participatory or non-participatory) should generate different responses, depending on the participants' additional involvement (i.e., the level of consumers' motivation to seek sensory pleasure). From these arguments, we propose that in a participatory campaign, consumers who are sensory pleasure seekers will be more involved with the campaign and, consequently, will have more CSR associations (perceive greater CSR). Therefore, we propose the following:

H3: When exposed to a participatory CSR campaign, the effect on CSR associations will be stronger for consumers with high motivation for sensory pleasure than for those with low motivation for sensory pleasure.

#### Consumer Skepticism of CSR Associations

Skepticism refers to a person's tendency to doubt, disbelieve, and question (Forehand and Grier, 2003). Research in the field of economics and business views skepticism as a potential consumer response to the actions of companies (Skarmeas and Leonidou, 2013) and defines it as consumers' distrust of or disbelief in companies (Webb and Mohr, 1998). The limited research on this topic notes that skepticism toward a company (negative assessment) occurs when consumers attribute selfish motives to the company actions (Webb and Mohr, 1998; Ellen et al., 2006). Thus, skepticism predisposes consumers to doubt the veracity of the communication activities of the company (Obermiller and Spangenberg, 1998). Indeed, consumers show a natural tendency to be skeptical of advertising (Obermiller and Spangenberg, 1998), though the extent to which they are skeptical varies from consumer to consumer.

The cognitive approach provides an explanation for consumer skepticism of persuasive communication (Hovland et al., 1953). Within this approach, the persuasion knowledge model (Friestad and Wright, 1994) states that consumers learn to interpret and evaluate the persuasion agents' goals and tactics and use this knowledge to cope with persuasion attempts. Consumers use the resulting knowledge to identify situations that motivate skepticism. Research on skepticism has been developed in different contexts, such as corporate social marketing (Forehand and Grier, 2003), environmental claims (Mohr et al., 1998), communication of CSR (Vanhamme and Grobben, 2009), and CSR programs (Pirsch et al., 2007). As a result, communicating CSR initiatives may be problematic (Pomering and Dolnicar, 2009) because consumer frequently perceive these initiatives as marketing actions that companies engage in out of their own self-interests (Haniffa and Cooke, 2005). Therefore, inferred motivations determine the level of consumer skepticism toward CSR messages and the credibility of social actions. If consumers perceive a company's motivation as selfish, they will be more skeptical about the campaign and will give less credibility to the company communication activities (Groza et al., 2011). Because prior research has established that CSR campaigns include tactics that can raise suspicion of firm motives (Ellen et al., 2006), consumer skepticism can bias the perception of CSR engagement (Groza et al., 2011). Suspicion about CSR activities will be stronger for skeptical consumers than for nonskeptical consumers, with the subsequent negative impact on CSR associations. Thus:

H4: The more skeptical consumers are, the lesser their CSR associations will be.

### The Relationship between CSR Associations and Attitudes

What consumers know about a company can influence their overall evaluations of and attitudes toward it (Luo and Bhattacharya, 2006). As part of their knowledge about the firm, consumers' perceptions of CSR are likely to influence their attitudes toward the firm and its social initiatives (Brown and Dacin, 1997). Attribution theory provides an appropriate framework for explaining how people attribute causes to events and how this cognitive perception affects their subsequent attitudes and behavior (Kelley and Michela, 1980). In this sense, CSR associations play an important role in consumers' responses to the company because they create a general context for evaluations (Sen and Bhattacharya, 2001). Consumers evaluate companies as well as their products in terms of CSR, and their perceptions of the motives for engaging in CSR influence their evaluations of and responsiveness to CSR (Becker-Olsen et al., 2006; Ellen et al., 2006).

Prior research has shown that consumers who are aware of a CSR initiative view the company as socially responsible (Brown and Dacin, 1997; Bhattacharya and Sen, 2004). As a result, the CSR activity has the potential to increase CSR associations and attitudes (Sen et al., 2006). If consumers believe that a company is concerned with the well-being of society and is committed to "doing good," they are more likely to have favorable attitudes toward the company (Stanaland et al., 2011). More specifically, consumers who are aware of CSR initiatives report more positive attitudes and behavioral intentions (Öberseder et al., 2013). Accordingly, we expect that these positive associations with CSR actions lead to more positive attitudes toward the company.

H5: CSR associations have a positive direct effect on consumers' attitudes toward the company.

**Figure 1** illustrates our conceptual model.

### METHOD

### Sample and Procedure

We ran a field study in which 196 people were randomly selected on the street of a medium- sized European city. Upon arrival at the lab, participants were randomly assigned to one of two experimental conditions: (1) a participatory CSR campaign condition or (2) a non-participatory CSR campaign condition. They were exposed to an advertisement of a fictitious CSR campaign developed by a local brewer. The campaign was related to reforestation. In the participatory condition, the participants were invited to take part in the campaign by planting a tree, while the non-participatory scenario was purely informative and indicated that the company performed reforestation activities. After that, participants reported their attitudes toward the company, CSR associations, search for sensory pleasure, and skepticism. Participants took approximately 10 min to complete the questionnaire.

The sample included 94 men (47.96%) and 102 women (52.04%), ranging from 18 to 35 years of age (M = 24.35), with 66.33% between 18 and 25 years. Graduate respondents accounted for 6.12%, undergraduates represented 30.10%, and respondents with a high school education (61.73%) or less (2.04%) accounted for the remainder. All research activities were performed in accordance with University of Murcia's institutional review board policies concerning research with human subjects. Prior to participation in the study, we gave participants an information sheet and told them the activity was part of an experiment. After completing the questionnaire, they were thanked and debriefed.

The questionnaire comprised four scales adapted from previous research. We used three seven-point semantic differential scale items adapted from Lafferty and Goldsmith (2005) to measure attitude. We assessed CSR associations with a four-item Likert scale adapted from Dean (2002). We measured motivation for sensory pleasure with five seven-point scale items (1 = strongly disagree; 7 = strongly agree) adapted from Eisenberger et al. (2010). Finally, we used Skarmeas and Leonidou's (2013) six seven-point semantic differential scale items for skepticism and two items for the manipulation check (did company X ask for your participation in the reforestation campaign? yes/no; did the campaign ask you to do something specific? yes/no).

### Measurement Assessment

Preliminary versions of the questionnaire were administered to a convenience sample of 20 consumers. We used the pretest results to improve the measures and design an appropriate structure for the questionnaire. Regarding the manipulation check, the 98 participants exposed to the participatory campaign confirmed that the campaign was participatory as required, while the 98 participants exposed to the non-participatory campaign confirmed the opposite.

We performed a validation check for the resulting measurement scales to assess their reliability, validity, and unidimensionality. We evaluated the reliability of the constructs using Cronbach's alpha coefficients (see **Table 1**). Cronbach alphas for the four constructs were above 0.70.

Confirmatory factor analysis tested the measurement model and obtained acceptable overall model fit statistics [χ 2 (59) = 118.08, p ≈ 0.00; RMSEA = 0.07; SRMR = 0.06; NNFI = 0.95; CFI = 0.96]. We assessed reliability using the composite reliability index and the average variance extracted (AVE) index. For all the measures, both indices were higher than the evaluation criteria of 0.60 and 0.50, respectively (Bagozzi and Yi, 1988), as **Table 1** shows. In line with Fornell and Larcker's (1981) suggested procedures, the scales showed acceptable convergent and discriminant validity. We assessed convergent validity by verifying that all indicators had statistically significant loadings on their respective latent constructs. The robust standard errors resulting from the use of the asymptotic covariance matrix were substantially larger (and the t-values smaller) than those produced by a model using the standard covariance matrix as input, validating the need for revised structural equation modeling (SEM) procedures in the face of strong non-normality in the data set. We also have evidence of discriminant validity. First, the phi matrix and associated robust standard errors presented in **Table 2** ensured that unit correlation among latent variables was extremely unlikely (Bagozzi and Yi, 1988). Second, for all the pairwise relationships in the phi matrix, the AVE

#### TABLE 1 | Construct and measures.


#### TABLE 2 | Phi matrix of latent constructs for full sample.


Standard errors in parentheses.

for each latent variable exceeded the square of the correlation between the variables.

To provide a further check of discriminant validity, for each pair of the latent variables, we compared the scaled difference chi-square statistic of the hypothesized measurement model with a second model that constrained the correlation between those two latent variables to unity. The corrected chi-square difference tests using the Satorra–Bentler scaled chi-square values (Satorra and Bentler, 2001) indicated that the hypothesized measurement model was always superior to the constrained models. As a result, we are confident that each of the latent variables in our model exhibits discriminant validity with all other latent variables. Internal consistency and discriminant validity results enabled us to proceed with the estimation of the structural model.

Assessment of potential common method bias was analyzed following the recommendations of Lindell and Whitney (2001). We used the smallest positive value within the correlation matrix as a conservative estimate of bias. This happens to be

#### TABLE 3 | Model testing.


\*\*\*p < 0.001.

the correlation between the motivation for sensory pleasure and CSR associations (r = 0.02). When we determined the statistical significance of the adjusted correlations, none of the correlations which were significant before the adjustment lost significance after the adjustment, indicating that the hypothesized relationships were not impacted by CMV.

#### RESULTS

**Table 3** reports the results of the SEM applied to test the hypotheses proposed in the theoretical model. We again used the asymptotic covariance matrix and robust maximum likelihood in model estimation. The model fit the data acceptably, as evidenced by the goodness-of-fit measures [χ 2 (119) = 228.98, p = 0.00; RMSEA = 0.069; SRMR = 0.077; NNFI = 0.95; CFI = 0.96].

We tested the effect of the participatory nature of the CSR campaign on CSR associations. While the main effect of the participatory nature of the CSR campaign did not significantly influence CSR associations (β = 0.13; SE = 0.15), thus rejecting H1, its interaction with sensory pleasure did (β = 0.41; SE = 0.15), in support of H3. Therefore, on the basis of these results, we can affirm that sensory pleasure seeking moderates the effects of the participatory nature of the CSR campaign on CSR associations. The negative coefficient of the main effect of sensory pleasure seeking (β = −0.35; SE = 0.13) confirms the direct effect of sensory pleasure, as proposed in H2. In addition, in H4 we predicted that the more skeptical the consumers, the lesser their CSR associations would be, and our statistical test also found support for this relationship. That is, less skeptical consumers have greater CSR associations (β = 0.37; SE = 0.07). Accordingly, if consumers are less suspicious about the real motives of a CSR campaigns, their associations with CSR actions are more positive.

Finally, the results confirm that CSR associations exert a significant and positive influence on consumers' attitudes toward the company (β = 0.31; SE = 0.07), as predicted in H5. Thus, as consumers generate more positive CSR associations, their attitudes toward the company become more favorable.

The significant interaction effect was further analyzed through floodlight analysis using Johnson-Neyman's approach (Spiller et al., 2013), in order to calculate the range of values of sensory pleasure seeking for which the participatory nature of the CSR campaign has an effect on CSR associations different from zero. Results, obtained with the probemod R package (Tang, 2015), show that the effect of the participatory nature of the CSR campaign on CSR associations is positive and significant (i.e., confidence interval does not contain zero at p = 0.01) for values of sensory pleasure seeking above 5.02. With this information, we divide the sample into two groups, low sensory pleasure seeking subjects (with scores in this variable below 5.02) and high sensory pleasure seeking subjects (with scores above 5.02), with subsample sizes of 133 and 63, respectively.

While the variable sensory pleasure seeking has been hypothesized in our study to interact only with the participatory nature of the CSR campaign, for the multi-group analysis we also considere its potential effect on the other relationships, as a way to check whether these interactions should have been included in the original model. Therefore, a model that imposed equality constraints on the three parameters (participatory nature of the CSR campaign—CSR associations, skepticism— CSR associations, and CSR associations—attitude toward the company) and a general model that allowed those parameters to vary freely across subgroups were compared. A chi-square difference test revealed that the unconstrained model represented a significant improvement in fit over the constrained model

A further series of tests identified there is only one path moderated by sensory pleasure seeking (**Table 4**). More specifically, the results showed a significant moderating effect consistent with H3. For high sensory pleasure seekers, being exposed to a participatory CSR campaign (compared to a nonparticipatory one) has a significant influence on CSR associations (β = 1.10; SE = 0.24). However, this effect is not significant for low sensory pleasure seekers (β = −0.23; SE = 0.19). The significant change in chi-square (1χ2 = 7.49; 1DF = 1, p < 0.01) indicates that this coefficient is different for the two groups. Additionally, the coefficients for the other two relationships displayed in **Table 4** are significant for the two groups and the changes in chi-square indicate that they are not significantly different between the two groups. In other words, seeking sensory pleasure does not moderate the effect of skepticism on CSR associations nor the effect of the latter variable on consumer attitudes toward the company. In summary, these results fully confirm H3.

The negative effect of seeking sensory pleasure on CSR associations can be related to how the company CSR activities have been described. As Eisenberger et al. (2010) demonstrate, high motivation for sensory pleasure individuals show increased interest in high- but no low-detail contextual information about the pleasantness character of the campaign. In other words, these subjects preference for very detailed information about possibilities of experiencing pleasure could have provoked a sense of frustration and lack of interest in the experiment stimuli as they were not very detailed when describing the company CSR activities (but this is a common characteristic to many ads). This lack of interest may have favor lower processing of the information and, therefore, less CSR associations.

We ran an additional study to provide further support to the scenario we used. All research activities were performed in accordance with University of Murcia's institutional review board policies concerning human subjects research. We collected 41 questionnaires. Twenty-one participants were assigned to the participatory campaign whereas the remaining 20 were assigned to the non-participatory campaign. Through items ranging from 0 to 10, individuals rated the campaign in terms of credibility, realism, level of participation required and level of involvement required. They also rated their ability to imagine themselves immersed in the situation described. Results showed that both groups perceived the scenario they were exposed to as highly credible [Mnp = 7, 15; M<sup>p</sup> = 7, 62; F(1, 39) = 1.43; p > 0.10]

#### TABLE 4 | Model testing for the multi-group analysis.


\*\*\*p < 0.01; \*\*p < 0.05.

and realistic [Mnp = 7, 40; M<sup>p</sup> = 7, 86; F(1, 39) = 1.63; p > 0.10], regardless of the type of campaign. They were also able to imagine themselves in the scenario [Mnp = 7, 35; M<sup>p</sup> = 7, 95; F(1, 39) = 2.03; p > 0.10]. As expected, there were significant differences between the two conditions in terms of level of participation [Mnp = 4, 10; M<sup>p</sup> = 7, 05; F(1, 39) = 20.57; p < 0.01] and involvement [Mnp = 4, 60; M<sup>p</sup> = 7, 62; F(1, 39) = 20.26; p < 0.05], with higher scores for those in the participatory condition, which lends further support to our methodology. In summary, despite our manipulation was based on scenarios instead of real participation, subjects immersed themselves in those scenarios and those assigned to the participative condition indicated they perceive the CSR campaign as more participative than participants in the regular CSR campaign.

#### GENERAL DISCUSSION

In the current competitive marketplace, companies intensely seek differentiation and credibility. One mechanism to reach such goals is consumer participation in CSR campaigns. However, while the effect of participation on consumer behavior has received considerable attention (Dellande et al., 2004; Chan et al., 2010), whether consumer participation in CSR activities increases the effectiveness of these activities remains unknown. The current research lends support to the contention that by proactively engaging consumers in CSR initiatives, firms can generate more favorable attitudinal responses than by acting in a reactionary manner (Becker-Olsen et al., 2006; Wagner et al., 2009).

Our findings on the effect of the participatory nature of a CSR campaign constitute a significant contribution both to the theory of consumer behavior and to business management. Thus, from a theoretical perspective, this research contributes to a better understanding of the effects of the participatory nature of CSR campaigns on consumer behavior. Specifically, although we did not find a general effect of the participatory nature of the campaign on consumers' perceptions of CSR activities, this does not mean that this effect does not exist. As the interaction shows, this effect is associated to consumers who are highly motivated to seek sensory pleasure. When the campaign is participatory, sensory pleasure seekers have greater perceptions of CSR activities than when the campaign is non-participatory. However, when consumers do not seek sensory pleasure, the fact that the campaign can offer possibilities to interact with the senses does not contribute to increase their CSR associations concerning the company. These results are in line with prior research suggesting that the motivation for sensory pleasure plays an important role in consumer responses (Eisenberger et al., 2010; Nurse et al., 2010) as well as those suggesting that enabling consumers participation leads to more positive outcomes (Chan et al., 2010; Yi et al., 2011; Troye and Supphellen, 2012; Mustak et al., 2013; Olsen and Mai, 2013; Atakan et al., 2014). In summary, our results show that motivation and involvement positively moderate the effect of a message on the valence of consumes' associations. A positive message (CSR activities undertaken by the company) contributes to more positive associations (CSR associations) when the consumer is more motivated (seek sensations) and involved (participates in the production of the CSR activity).

In addition, in light of the negative attributions consumers may attach to companies' real motives for conducting CSR actions, our research demonstrates that skeptical consumers show less CSR associations because they may perceive the campaign as manipulative. This, in turn, results in less favorable attitudes toward the company. This result is in line with research that posits that consumers are often skeptical of advertising claims related to a company's participation in social or environmental issues (Obermiller and Spangenberg, 1998; Chan et al., 2010).

From a managerial perspective, this research shows that, by itself, participation may not be enough to gain consumers' involvement and, in turn, generate greater CSR associations and favorable attitudes. To obtain this effect, companies should emphasize the possibility of finding pleasure during the CSR campaign. Firms could try to activate sensory pleasure, beyond the personal predisposition of each individual, by promoting CSR as an action that will produce an enjoyable experience as involvement increases. Therefore, marketing managers should not only engage their customers in their CSR actions, so that they actively take part in their implementation (as opposed to adopting a passive role and trusting that the company will do what is promised), but also design participatory CSR activities that stimulate and match the consumers' hedonic motivations for searching for pleasure.

In addition to participation and sensory pleasure seeking, marketers should pay attention to skepticism of the campaign. Some consumers tend to discredit CSR actions, as they interpret them as an attempt to manipulate their perceptions. They believe that the real motives behind CSR actions are not social oriented but benefit oriented. To minimize the impact of such skepticism, managers should provide consumers with clues that lend the campaigns more credibility. For example, they could report the results of previous initiatives to prove that the company is socially oriented and concerned about societal well-being.

Despite the findings and implications, this research has some limitations. First, although participants were able to imagine themselves in the proposed scenarios, we must acknowledge that they do not allow consumers to really participate in the CSR campaign. Therefore, further research should assess whether our findings remain the same in such a real context. Second, we use only one company, which limits the generalizability of the results. Thus, future research should analyze whether the implementation of participatory CSR activities for different products can generate different results for different company sectors. Third, other variables related to consumer–company interactions, such as the relationship with company workers, may also affect the results. Finally, another avenue for research pertains to consumers' personality traits, which may moderate the effects. For example, proenvironmental attitudes or prosocial behavior may boost the positive influence of participation.

### AUTHOR CONTRIBUTIONS

RL collected the data and the three authors have equally participated in literature review, data analysis and writing of the paper.

### FUNDING

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the grant ECO2012-35766 from

### REFERENCES


the Spanish Ministry of Economics and Competitiveness and by the Fundación Séneca-Agencia de Ciencia y Tecnología de la Región de Murcia (Spain), under the II PCTRM 2007- 2010. Authors also thank the support provided by Fundación Cajamurcia.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Ruiz de Maya, Lardín-Zambudio and López-López. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# The Effects of Technology Entrepreneurship on Customers and Society: A Case Study of a Spanish Pharmaceutical Distribution Company

Rosa M. Muñoz\*, Jesús D. Sánchez de Pablo, Isidro Peña and Yolanda Salinero

Business Management Department, Castilla-La Mancha University, Ciudad Real, Spain

The main purpose of this paper is to provide an understanding, within the field of corporate entrepreneurship, of the various factors that enable technology entrepreneurship in established firms and its principal effects on customers and society. The paper reports on a case study regarding technology entrepreneurship in a Spanish company whose activity is pharmaceutical distribution. This company has been able to overcome the consequences of the worldwide crisis and start an innovative process which includes the installation of new information technology (IT) and an investment of 6 million Euros. It is, in this respect, a model to imitate and the objective of this paper is therefore to discover the managers' entrepreneurial orientation (EO) characteristics which have made this possible, along with the organizational and social effects resulting from the process. We verify that EO is present in this company and that the development of new IT has important effects on customers and the population.

#### Edited by:

Alicia Izquierdo-Yusta, University of Burgos, Spain

#### Reviewed by:

Francisco José Pinto, Escola Superior de Gestão, Hotelaria e Turismo – Universidade do Algarve, Portugal Rocco Agrifoglio, University Parthenope of Naples, Italy

#### \*Correspondence:

Rosa M. Muñoz rosamaria.munoz@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 19 May 2016 Accepted: 13 June 2016 Published: 28 June 2016

#### Citation:

Muñoz RM, Sánchez de Pablo JD, Peña I and Salinero Y (2016) The Effects of Technology Entrepreneurship on Customers and Society: A Case Study of a Spanish Pharmaceutical Distribution Company. Front. Psychol. 7:978. doi: 10.3389/fpsyg.2016.00978 Keywords: corporate entrepreneurship, technology entrepreneurship, case study, pharmaceutical distribution, effects on customers and society

### INTRODUCTION

Entrepreneurial behavior is one of the key strategies of organizations that are seeking to acquire and sustain a competitive advantage in global markets. Researchers have coined various terms to describe this issue at the firm level: intrapreneuring, intrapreneurship, intracorporate entrepreneurship, corporate venturing, internal corporate entrepreneurship, and entrepreneurial strategy (Antoncic and Hisrich, 2004). However, for the purpose of this study, the general term corporate entrepreneurship (CE) will be used.

Although many companies state that the entrepreneurial spirit is part of their organizational cultures, it is not common to see organizations that have gained a competitive edge through the use of solid CE strategies (Demirci, 2013). CE can be defined as "a multidimensional process with many forces acting in concert that lead to the implementation of an innovative idea" (Hornsby et al., 1993, p. 30).

This process is initiated by the top management and includes a redefinition of the business concept, reorganization, and the introduction of system-wide changes for innovation (Tajeddini and Mueller, 2012). CE encompasses innovativeness, risk taking, and proactiveness (Zahra and Covin, 1995) and is an important determinant of firm, regional, and national economic

performance (Gupta et al., 2004; Wiklund and Shepherd, 2005). Bearing these characteristics in mind, the introduction of new information technology (IT) can be considered as a part of a company's CE. That is, opportunities to use new products and processes which stem from the development of new technology and/or the commercialization of technologies developed by others.

The need to pursue CE has arisen as the result of a variety of pressing problems including: required changes, innovations and improvements in the marketplace to avoid decline, perceived weaknesses in the traditional methods of corporate management and the turnover of innovative-minded employees who are disenchanted with bureaucratic organizations (Kuratko et al., 1990). Entrepreneurship is not limited to the start-up domain and a parallel strand in literature has been developed that stresses the importance of entrepreneurship for and within existing corporations. Entrepreneurial activities in existing organizations result in two possible types of corporate activities: strategic renewal- changes in organizational business processes and new business venturing which is related to the creation of new business units (Gómez et al., 2011).

The focus of this study is to explore the supporting factors in IT entrepreneurship and its consequences. The paper reports on a case study regarding technology entrepreneurship in a Spanish company whose activity is pharmaceutical distribution. It is a mature firm which, during the course of the last few years, has been able to adopt IT in innovative ways thanks to the entrepreneurial spirit of its managers. This rapid and necessary adaptation to the requirements of the new economy is positioning the company within its sector, and it is now one of the leaders. The firm has been able to overcome the consequences of the worldwide crisis and continue its innovative process, which includes the installation of new IT and an investment of 6 million Euros. It is, in this respect, a model to imitate, and the objective of this paper is therefore to discover the managers' entrepreneurial orientation (EO) characteristics which have made this possible, along with the effects on the organization, customers and society resulting from the process. The research question of this study is, therefore, to investigate "how" IT entrepreneurship occurs in the company and "what" factors enable it.

Three concepts are linked throughout the paper: CE, technology entrepreneurship and IT. Schumpeter (1942) defined the entrepreneur as a person who introduces new technologies into the production process. This author argued that entrepreneurship boosts innovation, the introduction of new products, or processes. With regard to IT, this factor favors competitiveness and innovation. IT is therefore an important part of technology entrepreneurship and technology entrepreneurship is an important part of CE.

This paper is organized as follows. Section "Corporate Entrepreneurship and Information technology entrepreneurship: the influence of consumer behavior and other factors" draw on literature in the field. Section "Methodology" presents the methodology. Section "Case Study" describes the history of the organization, its main characteristics and its type of business, explaining the technology entrepreneurship of the company and paying special attention to its organizational, strategic and social implications. Finally, Section "Conclusion" presents the main conclusion and implications extracted from the case study.

### CORPORATE ENTREPRENEURSHIP

The ideas behind CE can be traced back to the 70s, but it was not until the appearance Pinchott's (1985) book that it became a separate topic (Christensen, 2005). Entrepreneurship is considered to be a vital component in the process of economic growth and development. Organizational performance, growth and development may depend on entrepreneurship to a considerable extent (Antoncic and Antoncic, 2011).

This phenomenon can be studied from an individual perspective by analyzing the characteristics and functions of the individual entrepreneur (Bygrave and Hofer, 1991), differences between individual entrepreneurs and non-entrepreneurs (Gartner, 1990), or the collective process of the discovery, evaluation and exploitation of opportunities (Shane and Venkataraman, 2000).

Entrepreneurship is a process that consists of revitalizing existing companies, revenue growth, profitability enhancement and pioneering the development of new products, services and processes (Bailetti, 2002; Miles and Covin, 2002). We can reflect upon the definition proposed by Davidsson (2005), which states that it: "is about the processes of discovery and exploitation of opportunities to create future goods and services." A significant proportion of innovations emerge from within existing organizations (Hayton et al., 2013). Established organizations possess significant resource advantages over new start-ups: capabilities for the production, distribution and marketing of their services and products and legitimacy in their strategic fields and among stakeholders, particularly potential customers and suppliers (Miller, 1983; Rauch et al., 2009).

Innovation is at the heart of the entrepreneurial spirit (Lumpkin and Dess, 2001). Covin and Miles (1999) define innovation as "the introduction of a new product, process, technology, system, technique, resource or capability to the firm or its markets." This is conceptualized as new products or processes that significantly improve customer benefit and technological delivery over existing products (Chandy and Tellis, 2000). Durand (2004) considers innovation as a process which "goes beyond the limits of technologies to address the larger scope of change in general. Innovation can indeed deal with the technological side of human activities, thus with product design and manufacturing processes, but it may also deal with the organizational and social side, e.g., external interactions with suppliers, clients or partners, internal processes which became routines in the way the firm operates." In this case, we shall pay particular attention to external interactions and organizational and social effects.

Castrogiovanni et al. (2011), found that the creation of personal relationships and the development of an open

communication between owner-managers and employees, and among employees themselves, can help to explain the dynamics of entrepreneurial behaviors within small firms. Openness in communication is important as regards both promoting CE activities and creating the most appropriate work environment in which to carry out other resource management practices that stimulate entrepreneurial behaviors (DeNisi, 2015).

Entrepreneurial activities carried out by the enterprise to sustain or improve its competitive position have several consequences as regards processes, structures and capabilities (Srivastava and Agrawal, 2010; Ozdemir et al., 2016). Given the importance of CE, various scholars have focused on identifying the factors that contributing to or enhance CE (see **Table 1**).

Finally, Lumpkin and Dess (1996) define EO as an organizational decision-making proclivity that favors entrepreneurial activities. There is an assumption that EO represents a continuous variable in which all organizations can be positioned. Covin and Wales (2012) explore how the concept has been portrayed and assessed in prior research. They claim that researchers are free to choose whichever measurement approach best serves their research purposes. In this respect, some authors have conceived of EO as a construct composed of three sub-dimensions: innovativeness, risk taking, and proactiveness (Miller, 1983; Covin and Slevin, 1989). Some other authors have later expanded the number of dimensions that characterize EO by adding autonomy (Dess and Lumpkin, 2001; Hughes and Morgan, 2007), and they consider that this dimension is an important characteristic of EO. The level of autonomy that managers give to employees can drive the innovations, creativity and changes usually linked to EO.

### INFORMATION TECHNOLOGY ENTREPRENEURSHIP: THE INFLUENCE OF CONSUMER BEHAVIOR AND OTHER FACTORS

Within the field of CE, we shall pay attention to the process of discovering and applying new IT systems. Scholars began to analyze the creation of technology-based firms from the 1990s, since these firms contribute to job creation and play a crucial role in renewing the economic system. Scholars have adopted diverse conceptualizations of IT, and have extended it beyond hardware and software to include a range of contextual factors associated with its application within organizations.

Information technology has been one of the most important drivers of economic and social value in the last 50 years, transforming organizations, markets, industries, societies, and the lives of individuals (Lucas et al., 2013). Understanding the economic impact of IT is a critical issue for researchers, and there is a rich body of literature concerning IT value (e.g., Wade and Hulland, 2004). Many papers have stressed the strategic significance of IT as regards supporting competitive strategies and improving firm performance (Powell and Micallef, 1997; Kohli and Devaraj, 2003; Melville et al., 2004; Pavlou and Ei Sawy, 2006; Chae et al., 2014). Others have stressed the IT-Productivity relationship at national economy level (Dedrick et al., 2013). These works are based on several theoretical paradigms, including microeconomics, industrial organization theory, and sociological and socio-political paradigms showing the complex problem of linking IT to organizational performance (Melville et al., 2004).

For many firms, the most common reasons for adopting IT are to provide a means to enhance survival and growth, thus staying competitive and enhancing innovation abilities (Nguyen, 2009).


IT can add value to an organization via the functionality, usability and information structure, which in turn affect the quality, efficiency and innovations of IT users (Gustafsson et al., 2009). Adopting new IT is also a means to enhance the way in which people capture and distribute information, lower production and labor costs, add value to products and services and increase the company's competitive advantage (Nguyen, 2009). These IT systems are widely applied in business such that not only production but also administration processes can be technologyintensive.

Information technology systems require quality interactive interfaces and compatible dynamic knowledge systems. If they are not compatible with these systems, then either the systems should be improved or replaced, or people should be better trained or replaced (Wang, 2003). If the cost of hiring new staff is high, firms must provide the current staff with training. Training and effective communication must be provided for existing employees if there is to be a substantial change in the IT (Nguyen, 2009).

Information technology systems can reduce and automate repetitive works, and reduce the time needed to search for copy, collect, and format information, and they can enable team members to focus on critical and inventive activities (Wang, 2003). Moreover, innovations in one area have important implications in other areas, and distribution impacts on the concepts of product and service (Franklin et al., 2013).

Information technology influences the organizational structure via the company's organizational design parameters (Mintzberg, 1979): (a) job design, thereby reducing specialization since IT assumes the routine tasks. This increases autonomy within the organization. When employees have the capacity to make decisions regarding their work, this is directly related to their attitudes toward engaging in entrepreneurial activities; (b) design of the superstructure, thus cutting down the number of hierarchical levels since IT simplifies communication, coordination and control functions and therefore increases the degree of decision making authority in the possession of individuals; (c) design of the lateral linkages, thereby improving analytical and design capabilities, increasing access to information and making it easy to access the results; and (d) design of the decision-making system, thus allowing organizations to simultaneously exploit the advantages of both centralization and decentralization. In this respect, Lau et al. (2001) prove that IT has significant impacts on: (a) formalization, moving organizations toward less formalized network structures; (b) specialization, thus facilitating outsourcing process; (c) promoting the use of lateral communication; (d) team work, as this is basic to a flatter organization; and (e) the learning of organization based on open communication, coordination improvements, and training.

### METHODOLOGY

The paper adopts an exploratory perspective and employs a qualitative approach. A case study was used to gain deeper insights into a contemporary and complex issue within its reallife context (Yin, 1994). Many studies are based on qualitative methodologies and can reveal the reality of entrepreneurship in organizations (e.g., Brennan and McGowan, 2006; Gómez et al., 2011) and concretely the implantation of IT in pharmaceutical distribution (Clemons and Row, 1988; Bruque et al., 2004).

Clemons and Row (1988) analyzed McKesson's order entry and distribution system, Economost. This system allows almost 100% of McKesson's orders to be entered electronically by customers. The impact of Economost on McKesson's system has been favorable, affecting: (a) the competitive position, including its profitability and market share relative to its competitors; (b) the industry as a whole, including profitability and concentration; (c) suppliers; and (d) customers, because sales personnel are no longer principally order takers and they can be used actively, as business consultants to the retailers.

Bruque et al. (2004) carried out an analysis of 16 cases in the pharmaceutical distribution sector in Spain. The results indicate frank and fluid communication between departments and members of the organization, low levels of conflict and the explicit support of top management as regards the introduction and development of IT.

The single case setting limits the applicability of the research to other institutions. However, the framework and model that are developed, along with the overall approach, are valuable contributions to an important and emerging research area. Casebased research aims to generate refined theory based on an in-depth understanding of a particular context. According to Yin (1994), research must identify some situations in which all research strategies might be relevant. The "how" and "what" questions are asked about a contemporary set of events over which the investigator has little or no control. This study investigates "how" EO occurs in the company and "what" factors enable it. As the paper seeks to address research questions, this suggests the adoption of an exploratory approach (Yin, 1994). The identification of EO enablers is essentially exploratory, in the sense that the main objective is to refine a research idea in order to facilitate further research (Kervin, 1992).

The qualitative approach and exploratory nature of the research question influenced the data-collection method. Research conducted within the qualitative paradigm is characterized by its commitment to collecting data from the context in which social phenomena naturally occur and to generating an understanding that is grounded on the perspectives of research participants (Marshall and Rossman, 1995). Data collection was consequently developed using desk and field research. Desk research, based on internal company documents, served to provide a detailed understanding of the innovation. Moreover, the field research included in-depth semi-structured interviews with the company's Managing Director, Financial Director and IT professionals. Interviewing multiple respondents in the company provided diverse sources of evidence and served as a means to validate and replicate the findings. The interviews took place during June–July 2013.

Bearing in mind the principal opinions of the scholars reviewed above, we have included the following dimensions in our research: EO, organizational effects, company strategy and



Source: Cofarcir's Annual Reports.

Cooperativa Farmacéutica de Ciudad Real (Cofarcir) is a Spanish cooperative located in the city of Ciudad Real (Spain; it also has another warehouse in Alcázar de San Juan, a town in the province of Ciudad Real), whose business is the acquisition and distribution of pharmaceutical specialties for exclusive use by its partners (the members of the cooperative society) and, in general terms, every product related to the practice of the pharmaceutical profession. Its organization chart is shown in **Figure 1**.

The company started its activity in 1931, and during the course of the last 84 years of its existence has been numbered amongst the six largest companies in the province and the twenty largest companies in the country. Some data related to the company is shown in **Table 2**.

### Technology and Innovation at Cofarcir

The pharmaceutical distribution sector, and particularly the company under analysis, has always been a pioneer in relation to the adoption of continuous IT developments. In the 80s, Cofarcir gave its pharmacies dataphones for the electronic transmission of medical orders, thus preceding the Internet and the official definition of e-business by several years. That initial step in the use of IT added to the robotic transformation of the company's main warehouse, which took place in 1994 and which was an important step that allowed Cofarcir to fulfill two daily orders to each pharmacy in the region, thereby consolidating its position as regional leader. The automation of the other warehouse took place in 2005 and permitted the application of the most innovative technology in the sector worldwide.

The pharmaceutical sector in general, and the distribution of drugs in particular, is immersed in a continuous process of change in Spain. The worldwide crisis we have been confronting since 2008, which is more serious in this country, has resulted in constant decreases in the prices of medication, in subsidies for patients as regards acquiring them and in many other measures such as the promotion of generic products or the setting up of public pharmacy services inside hospitals. Government sanitary agencies meanwhile continue to push the ideal goal of total traceability of the whole medication chain, from the laboratory that produces it to the final patient who consumes the drug. This is an extremely difficult objective to achieve, and the most recent theoretical technologies are intended for this purpose. Thirdly, the great pressure that the huge wholesalers, which control the pharmaceutical distribution in other European countries, are exerting on Spain with the aim of changing the

social effects. The Covin and Slevin (1989) and the Hughes and Morgan (2007) scales were chosen in order to analyze the EO in our case study. The first scale considers that EO is a composite of risk taking, innovativeness, and participativeness. We have also introduced the autonomy items from the latter because, like Hughes and Morgan, we consider the fact that the employees are permitted to act and think without interference and perform jobs that allow them to make and instigate changes to be important for EO. Furthermore, we analyze the organizational effects based on Mintzberg (1979). We have also added more questions in relation to organizational effects, company strategy and customers and social effects in order to include some important issues that affect this specific case (see **Appendix**).

First, in the following section, we describe the company, its activity and its most recent innovations. We then explain the principal ideas extracted from the interviews.

## CASE STUDY

### The Company and the Pharmaceutical Distribution

The principal activity of the firm under analysis is that of pharmaceutical distribution. The pharmaceutical industry in Europe is defined as being composed of commercial enterprises engaged in the design, creation and development of medicines to prevent or cure disease and relieve human suffering. Distribution can be defined as a variety of processes linked with delivering the pharmaceutical products needed to the right place at the right time (such as express delivery to hospitals or daily shipments to pharmacies).

One of the most important characteristics of pharmaceutical distribution in Spain is its wide public regulation. Customers can access medicines in three main ways, which implies, respectively, 1, 21, and 78%. The national health programs fall into the first of these, which, through specific programs (concerning, for example, vaccinations), supply medicines directly to users. Hospitals make up the second, through the medication administered to hospitalized patients. The third is direct purchase by the general public, and it is in this process that the wholesalers (like Cofarcir) and the retailers (the pharmacies) participate.

successful Mediterranean model. Finally, some large cooperatives in this sector have opted for expansion outside their natural territories.

Cofarcir has decided to confront these major challenges in two ways: firstly, it continues to rely on the current regional cooperative model, convinced that this is the optimal means to maintain the high standards of the Mediterranean model. To accomplish this goal Cofarcir has strengthened its ties with several other Spanish cooperatives that are united in this vision in order to create the group Unnefar, the fourth biggest of its kind in Spain. This group's intention is to provide its members with the size needed for them to continue being a relevant piece of the system in defense of their principles. Secondly, Cofarcir wishes to improve the service it provides to its customers/owners, whilst enhancing its productivity. The company has therefore made the decision to renew the technology at its main warehouse in Ciudad Real. The automation level of this storehouse, although a pioneer when installed back in 1995, was becoming outdated and was virtually unable to keep pace with the new regulations commented on above. The company selected for this change in technology has again been Knapp, an Austrian company and a worldwide leader in this sector.

The core of the project is to install a chaotic storage system of 16 m-high towers – called OSR35b – in which the system automatically chooses the place where the goods are stored, allowing the separation of a good by its batch number and its expiration date. The user does not know where the products are kept inside the robot. They only ask the system for an amount of good and the system retrieves it to them, following certain premises (FIFO, for instance) while keeping track of which batch is sent to which pharmacy.

There are several options by which to 'feed' that robot with the exact batch numbers and expiry dates. The best appear to rely on the information the laboratories send with the products when delivered to Cofarcir. However, this information is not always sent in the way-bill and worse, when it is, it is very difficult to introduce into the computer system as it is printed on paper. For the time being, Cofarcir will continue to use the AECOC standard which uses the GS1 bar-code system to describe much important information about the products delivered. This information is registered in Cofarcir with bar-code readers and special software developed internally, and is linked with the product income via radio-frequency devices when available.

The increase in productivity will principally occur in two ways. Firstly, although around 75% of the products will be dispatched automatically by new and existing robots (TDA, SDA, and HDA Knapp technologies), the remaining 30% will, owing to their fragile nature, great volume or high demand, have to be dispatched by hand. A new technology is now being used and improves the speed of manual dispatching by up to four times, while decreasing the error rate.

The last principal way in which the new technology will improve the service is with the use of a new robot – called OSR35ds – that will efficiently order the client packages into routes, thus saving the delivery van drivers' time in addition to minimizing the possibility of mistakes as regards mixing up boxes. There is also a new positioning system, which has been developed by Cofarcir employees and which aims to use the Internet to inform clients where their order is and how long it will take them to receive it.

The total investment will be up to 6 million Euros. This amount includes not only the cost of the technology explained above but also the building of a new warehouse next the current one to contain it.

### Technology Entrepreneurship at Cofarcir and Its Main Effects

We shall now present the main ideas extracted from the interviews (**Appendix Figures A1** and **A2**).

#### Entrepreneurial Orientation

Following the Covin and Slevin (1989) and the Hughes and Morgan (2007) scales we can consider that the company has an EO, i.e., it exhibits high levels of risk taking, innovativeness and proactiveness. In general terms, we can summarize that the organization has, over the last few years, marketed very many new lines of products and has placed a strong emphasis on R&D, technological leadership and innovation; in dealing with its competitors, it is very often the first business to introduce new techniques and operating technologies, adopting a very competitive posture; the company's top managers have a strong proclivity for high-risk projects, consider that owing to the nature of the environment, wide-ranging acts are necessary to achieve the firm's objectives and, when confronted with decision-making situations involving uncertainty, the firm typically adopts a bold, aggressive posture in order to maximize the probability of exploiting potential opportunities. However, upon considering the autonomy items proposed by Hughes and Morgan (2007), although employees are given freedom to communicate without interference and have access to all vital information, they are not given freedom to decide for themselves how to go about doing their work or to make and instigate changes as regards the way in which they perform their work tasks.

#### Organizational Effects

In order to analyze the new technologies' organizational effects we consider Mintzberg's (1979) structuring of organizations. IT influences the organizational structure via the company's organizational design parameters: job design, design of the superstructure, design of the lateral linkages and design of the decision-making system.

In relation to job design, the IT in Cofarcir allows its workers to know how they are performing their jobs, thus allowing them to improve their skills in order to make their tasks more efficient. This may lead to increased motivation since the workers can evaluate their performance. It also makes the workers more multifunctional because they have to cope with the instructions, guides, checking, warnings, etc. implemented in their workloads, thus allowing them to learn new tasks faster. However, IT's role as regards making work more creative is not so clear, since one of its main goals is standardization. In the design of the superstructure, there is no need to cut the number of managers required because,

considering the firm's innovative trajectory, it is already adjusted and is as flat as is necessary. In any case, IT helps workers to auto-organize their work whilst simultaneously providing management with powerful means of control, from the overall performance of the production system to the most detailed performance of an individual worker. This leads to a direct and efficient control of the company, eliminating intermediaries in the process.

In terms of jobs, owing to the considerable changes in the way that the work will be done, the company considers that there is no need to decrease the amount of warehouse workers. New specialized jobs such as an industrial engineer in charge of the new installation in order to control the new robot, and the reconversion of some warehouse workers into administrative workers are considered sufficient to offer better/new services to the customers. The automation of the firm has not therefore created or eliminated departments but has simply redistributed the employees to control, administration, maintenance and IT jobs since fewer workers are now directly needed in the production process. It is therefore necessary to make changes as regards the employees' training not only in the mechanical area but also in those of management, maintaining, etc.

With regard to lateral linkages, tools like Business Intelligence Software provide the management with a better understanding of the process and allow them to determine the main variables that affect it, in addition to keeping track of the critical points involved. This leads to an improvement in the planning phase. Improved control is one of the main reasons for investing in IT. In the design of the decision-making system, IT eventually supports both centralization and decentralization. On one hand, as stated previously, managers need IT to be able to make decisions based on a better understanding of the development of the process. On the other hand, the exact information at the right time is vital for those processes that are directly related workers in order to prevent errors, correct deviations and assign more resources to those activities which are more demanding in each case.

Finally, the IT developed by Cofarcir has been strategically planned in such a way that it will have a lifecycle of about 20 years. This is not only a leading technology but also a change of paradigm with regard to the existing processes, and it will therefore have a long trajectory. Here we are referring to a vertical occupation of the space, the chaotic storage of goods, the intensive use of radiofrequency and the virtual elimination of the use of paper. The company is also preparing for new business lines. The technology installed should be improved in the future. In this respect, various complementary innovations have been planned, such as the installation of a massive new automatic dispatching robot (SDA), fed by the chaotic storing system currently installed (OSR), or the transformation of part of the manual dispatching of goods (picking) to robotic devices (currently in a prototype state).

This is the third technological "wave" developed by Cofarcir. The first took place 20 years ago, in the company's main installations in Ciudad Real, and was the more disruptive innovation owing to the replacement of manual processes with robotic ones. The second technological change, which was similar to the first, took place in the Alcázar de San Juan installations. These prior experiences have permitted the company to maintain its leadership and to guarantee the success of the third "wave" development.

#### Company Strategy

Although part of the system is already in other companies, the use of the latest innovations has led the installer firm to consider Cofarcir as a model for other organizations (show room). The main reasons for the changes have been the obsolescence of some parts of the system and the need to adapt to the present requirements of pharmaceutical distribution. In this respect, the company is following a differentiation competitive strategy.

The entire process has been developed by Cofarcir itself, although its excellent relationship with other cooperatives in the sector has permitted it to share experiences and visit their installations. This benchmarking process is therefore a source of new ideas for the company, and these external partners are important sources of knowledge for Cofarcir, thus allowing it to keep up to date with the latest technological developments (Brettel and Cleven, 2012).

### Effects on Customers and Society

The present technological development at Cofarcir is having an important effect on pharmacies and, as a consequence, on customers and the population. All of them are able to receive two daily orders in no more than 3–4 h, so a just in time system has in fact been developed. Moreover, the improvements in the conservation and transportation of the medicines have led to a better delivery of the goods to the pharmacies. The social impact is also important, bearing in mind the special characteristics of the pharmaceutical sector. It is a sensitive sector for the population as all the activities are related to health.

All these characteristics make it possible to consider Cofarcir as a "market pioneer" or "first-mover." The mechanism that produce a pioneering advantage is in some way able to slows the natural forces of competition, thus making it difficult for later entrants to attain a pioneer's advantage. The process by which consumers learn about new products or services and form preferences for them play an important role in creating an advantage for pioneers. The first-mover can become strongly associated with the product category as a whole and, as a result, become the "standard" against which all later entrants are judged (Carpenter and Nakamoto, 1989).

### CONCLUSION

This study shows the organizational and social effects of IT entrepreneurship. We consider that it is a potential source of sustainable competitive advantage through the technical skills in technology and the ability to manage these new technologies. Technological entrepreneurs are more motivated than other entrepreneurs as regards starting projects and putting their innovative ideas into practice. They therefore tend to be driven by the need for achievement and self-realization and the desire to implement their projects. One area that is increasingly seeking

ways in which to add value through innovation is that of the logistics function (Soosay and Hyland, 2004). Cofarcir is a pharmaceutical distribution leader that has been able to follow a successful technological strategy during its more than 75 years of life. The IT developed by Cofarcir is expected to have:


The recent history of Cofarcir shows an important EO, this being principally a technology entrepreneurship which contributes toward improving its productivity. Here, we present a summary of the main improvements the company is expecting from its new technology development:


Although, the managers of Cofarcir have a relevant amount of EO (this new technology is only one new phase in the innovation strategy), they maintain the traditional work system (Mediterranean model). The company also attempts to improve its knowledge by reaching cooperative agreements with other cooperatives. Cofarcir is therefore adapting its traditional strategy to changes in the environment. In this case we consider the managers' attitudes to innovation to be particularly determinant. Fry (1987, p. 5) argued that:"...if managers aren't innovative, if they don't provide the climate for creativity, if they can't set aside their carefully laid plans to take advantage of a new opportunity, then intrapreneurs have little encouragement."

The research question we set out at the beginning of the paper concerned investigating "how" EO occurs in the company and "what" factors enable it. In this respect, this study makes several contributions to the literature on technological entrepreneurship. Our analysis of this company's evolution has allowed us to extract the following key dimensions as determinants and consequences of its technology entrepreneurship (see **Table 3**).

With regard to "how" and "what," the results of the study therefore support the previous research developed by the authors referred to in the theory review. Risk taking, innovativeness and proactiveness are the entrepreneurial enablers for a company's EO and for the successful result of the activities linked to it.

### Research Implications

This study has important implications for companies. Planning is necessary in order to obtain a successful strategy. This is typically analyzed in academic and practitioner communities as a fundamental managerial activity. Unfortunately, the contributions of planning are often difficult to quantify in practice (Segars and Grover, 1998) with an important gap between research and practice. It is generally accepted that one of the key factors for successful information systems planning and implementation is the close linkage with business strategy (Baets, 1992). Planning activities require substantial resources in terms of managerial time and budget. The process must therefore provide benefits beyond the resources needed to sustain it in order to make a positive contribution to organizational effectiveness (Segars and Grover, 1998). In this respect, Cofarcir's managers have communicated a clear vision and plan to employees for a long term horizon, and this is one of the most important factors as regards explaining the company's success and long



Source: Authors.

life. We can also see how the entrepreneurs' innovating spirit does not expire with the creation of the project, and allows them to constantly redefine the business. For already established companies, this case illustrates the need for and usefulness of maintaining entrepreneurship tension throughout project development in order to favor the emergence of new and potentially complementary projects.

The improvement made to the whole process of product delivery is one of the most important results of Cofarcir's IT entrepreneurship. Consumer service is now better in terms of: a decrease in the error rate, the speed of order dispatching, software to inform customers where their order is and how long it will take them to receive it, etc. As stated previously, Cofarcir can be considered a "market pioneer" but the consumers play an important role in making this pioneering advantage persistent. Our analysis therefore contains some suggestions as to how a successful relationship with consumers can be attained, i.e., achieving a competitive advantage by influencing consumer tastes rather than responding to them. Recent research in marketing and psychology suggests that consumers often use information already contained in existing product or service categories to learn about new ones. Consumers will learn about these innovations more quickly and with fewer mistakes if entrepreneurs delineate the appropriate information that should be transferred from each domain (Moreau et al., 2001). That is, it is not sufficient for a company to be able to follow a successful technological strategy, and the process of educating consumers about an innovation is necessary in order to influence how consumers will structure their representations of it. There are many ways in which consumers can learn about new products or services. The learning theories most frequently cited in marketing literature include category-based

#### REFERENCES


learning, analogies, and mental simulation (Hoeffler, 2003) and entrepreneurs should attain some knowledge of them in order to improve their relationships with consumers and obtain mutual benefits.

This article has several limitations that reveal possible avenues for further research. First, this study has applied an in-depth case study to a firm. This implies an understanding of a complex issue or object extending experience or adding strength to what is already known thanks to previous research. Case studies emphasize a detailed contextual analysis of a limited number of events or conditions and their relationships. However, the findings should therefore be simultaneously treated with caution and should be verified and validated in other firms and other sectors. The focus of this case study was limited to a few concepts (EO and organizational and social effects) which are important for technology entrepreneurship. Various future research opportunities may therefore be possible. In order to form a more comprehensive and integrative case study, some other variables that are possibly important to CE and technological entrepreneurship could also be included, such as the characteristics of cooperative agreement or the managers' personalities. It will also be necessary to repeat this study within a few years in order to observe the long-term effects of this technology entrepreneurship and the innovations developed by the company.

#### AUTHOR CONTRIBUTIONS

RM: has developed the interviews with the company's CEO; JSP: has made the theoretical review; IP: has made the empirical study; YS: has made the empirical study.



inform new audiences. MIS Q. 37, 371–382. doi: 10.1098/rsif.2009. 0564


future research. MIS Q. 28, 107–142. doi: 10.1016/j.socscimed.2011. 11.042


Yin, R. (1994). Case Study Research. Beverly Hills, CA: Sage.

Zahra, S. A., and Covin, J. G. (1995). Contextual influences on the corporate entrepreneurship-performance relationship: a longitudinal analysis. J. Bus. Ventur. 10, 43–58. doi: 10.1016/0883-9026(94)00004-E

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Muñoz, Sánchez de Pablo, Peña and Salinero. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## APPENDIX

### Interviews Items Entrepreneurial Orientation

fpsyg-07-00978 June 25, 2016 Time: 13:12 # 12


FIGURE A1 | The Covin and Slevin (1989) EO Scale. Source: Covin and Wales (2012).


FIGURE A2 | Autonomy items of Hughes and Morgan (2007). Source: Covin and Wales (2012).

#### Organizational Effects



#### Company Strategy


#### Effects on Customers and Society

• Has the entire process had any effect on customers and society? How will the pharmacies be affected by it?

# Assessing the Moderating Effect of the End User in Consumer Behavior: The Acceptance of Technological Implants to Increase Innate Human Capacities

Jorge Pelegrín-Borondo<sup>1</sup> \*, Eva Reinares-Lara<sup>2</sup> , Cristina Olarte-Pascual <sup>1</sup> and Marta Garcia-Sierra1, 3

<sup>1</sup> Departamento de Economía y Empresa, Universidad de La Rioja, Logroño, Spain, <sup>2</sup> Departamento de Economía y Empresa, Universidad Rey Juan Carlos, Madrid, Spain, <sup>3</sup> Institute of Environmental Science and Technology, Universitat Autònoma de Barcelona, Barcelona, Spain

#### Edited by:

Ana I. Jiménez-Zarco, Universitat Oberta de Catalunya, Spain

#### Reviewed by:

Kiyoshi Murata, Meiji University, Japan Maria Dolores Reina, Universidad Nacional de Educación a Distancia, Spain

> \*Correspondence: Jorge Pelegrín-Borondo jorge.pelegrin@unirioja.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 19 December 2015 Accepted: 25 January 2016 Published: 22 February 2016

#### Citation:

Pelegrín-Borondo J, Reinares-Lara E, Olarte-Pascual C and Garcia-Sierra M (2016) Assessing the Moderating Effect of the End User in Consumer Behavior: The Acceptance of Technological Implants to Increase Innate Human Capacities. Front. Psychol. 7:132. doi: 10.3389/fpsyg.2016.00132 Today, technological implants are being developed to increase innate human capacities, such as memory or calculation speed, and to endow us with new ones, such as the remote control of machines. This study's aim was two-fold: first, to introduce a Cognitive-Affective-Normative (CAN) model of technology acceptance to explain the intention to use this technology in the field of consumer behavior; and second, to analyze the differences in the intention to use it based on whether the intended implant recipient is oneself or one's child (i.e., the moderating effect of the end user). A multi-group analysis was performed to compare the results between the two groups: implant "for me" (Group 1) and implant "for my child" (Group 2). The model largely explains the intention to use the insideable technology for the specified groups [variance explained (R 2 ) of over 0.70 in both cases]. The most important variables were found to be "positive emotions" and (positive) "subjective norm." This underscores the need to broaden the range of factors considered to be decisive in technology acceptance to include variables related to consumers' emotions. Moreover, statistically significant differences were found between the "for me" and "for my child" models for "perceived ease of use (PEU)" and "subjective norm." These findings confirm the moderating effect of the end user on new insideable technology acceptance.

Keywords: consumer behavior, technology acceptance, technological implants, insideables, cognitive factors, affective factors, subjective norm

## INTRODUCTION

Companies and research institutions are currently developing technological implants (insideables) both to increase innate human capacities (Technological Implants to Increase Innate Capacities, T3ICs), such as memory (Berger et al., 2011; Cohen, 2013), and to endow us with new ones, such as the remote control of machines (Regalado, 2015). The fact that this technology could a priori be implanted in healthy people for the sole purpose of enhancing their senses is controversial, and many people even have ambivalent opinions (Olarte-Pascual et al., 2015). For some, the integration of technological implants is considered a "quantum leap" for the species that will allow reasonable people to enhance their capacities to the extent that technology allows (Selinger and Engström, 2008). For others, this technology triggers fear of dehumanization (Lai, 2012). Although not yet quite a reality, T3ICs will most likely be available for future generations. The biomedical engineer Ted Berger, developer of the prosthesis for restoring memory, has noted that such implants will be available for our children in the near future (Berger et al., 2011). With regard to their use in children, Apple cofounder Steve Wozniak has noted that the intention to use T3ICs on one's children may be greater than that to use them on oneself (Jáuregui, 2014) and that, while he "would like to remain natural himself," he would want his children to have them "if in a few years other kids (. . . ) will have certain advantages thanks to technology." Nevertheless, public acceptance of this new insideable technology, the subject of the current paper, has not yet been investigated in academic research, whereas the ethical and moral implications of T3ICs have (Schermer, 2009).

With this in mind, the aim of this study was two-fold: first, to introduce a Cognitive-Affective-Normative (CAN) model of technology acceptance to explain the intention to use of this insideable technology in the field of consumer behavior; and second, to analyze the differences in the intention to use it based on whether the intended implant recipient is oneself or one's child (i.e., the moderating effect of the end user). The CAN model was tested on two groups: "T3ICs for me" (Group 1) and "T3ICs for my child" (Group 2).

In conducting research on the acceptance of new technologies, many researchers build on variables from previous models that have proved influential to technology acceptance (e.g., Hameed et al., 2012). In this vein, the Technological Acceptance Model (TAM) variables "perceived usefulness (PU)," "perceived ease of use (PEU)," and "(positive) social norms" positively affects the intention to use a new technology (Davis, 1989; Davis et al., 1989; Venkatesh and Davis, 2000). These variables consistently explain a substantial part (∼40%) of the variance in the intention to use innovative technologies, as demonstrated in several studies (Venkatesh and Davis, 2000). The CAN model is based on these previous models of technology acceptance, namely the TAM (Davis, 1989; Davis et al., 1989), the TAM2 (Venkatesh and Davis, 2000), and their extensions via the Unified Theory of Acceptance and Use of Technology (UTAUT and UTAUT2), which include the effect of social influence (Venkatesh et al., 2003, 2012). The CAN model includes the cognitive variables "PU" and "PEU," as well as the normative variable "subjective (or social) norm." The literature has recognized the influence of normative factors on people's attitude, intention, and behavior (Fishbein and Ajzen, 1975; Bagozzi, 2000; Venkatesh et al., 2012). The latter may also play a key role in assessing implantation, especially in children, by capturing parents' moral concerns. Indeed, in the field of pediatric surgery, meta-analytic results show that cognitive (i.e., personal factors, preferences), affective and normative factors, namely the opinions of other community members, do influence parent's consent to implantation (Lipstein et al., 2012).

However, the CAN model introduces a novel extension with respect to the TAM and UTAUT models: the inclusion of the affective variables "positive emotions," "negative emotions," and "anxiety." The benefits of including both cognitive and affective factors in order to better understand subjects' assessments of products has been widely acknowledged in the literature (Holbrook and Hirschman, 1982; Shiv and Fedorikhin, 1999; Campbell, 2007; Bigné et al., 2008; Levav and McGraw, 2009; Zielke, 2011).

There are no previous references specific to the acceptance of T3ICs. The closest background literature are studies contrasting the acceptance of implantable medical technology and physical implants for medical or cosmetic reasons. These contexts offer some evidence that can be interpreted as an indication of the possible acceptance of T3ICs. Technological implants to compensate for physical impairments, e.g., peacemakers or cochlear implants to assist children with hearing disabilities, are widely accepted and their use is widespread (Hill and Sawaya, 2004; Rosahl, 2004; Schermer, 2009; Pray and Jordan, 2010). Likewise, the use of physical implants for reasons other than improving one's health status, such as the incorporation of physical implants for breast augmentation (i.e., augmentation mammoplasty), seems to be accepted as well, at least in adults. Many people have already chosen to modify their body to match socially-accepted beauty standards (Adams, 2010) and increase their seductive capacities (Lawton, 2004). As for physical implants for cosmetic reasons, in the US, 4% of all cosmetic surgeries performed in 2014 were performed on patients between the ages of 13 and 19 (American Society of Plastic Surgeons, 2014). In Spain, 10% of all cosmetic surgeries where performed on patients under the age of 18 (Sahuquillo, 2008). Some interventions are due to true pathologies, but not all. Indeed, the Spanish Society of Plastic Reconstructive and Aesthetic Surgery (SECPRE in Spanish) advises against unnecessary cosmetic interventions in minors (Sahuquillo, 2008). It is noteworthy that in most countries minors need the consent of their parents to undergo surgery, although in specific cases, minors over the age of 16 are allowed to decide for themselves. For some people implantation is only considered desirable if it addresses medical issues (impairments due to accidents or illness), not when it is performed for the purpose of beautification (Schaar and Ziefle, 2011).

### MODEL VARIABLES, HYPOTHESES, AND MULTI-GROUP COMPARISON

The CAN model accounts for the influence of cognitive factors ("PU," and "ease of use"), affective factors ("positive emotions," "negative emotions," and "anxiety") and normative factors ["subjective (or social) norm"] on consumer behavior. The following subsections describe the model variables and underlying hypotheses, including the hypothesis that the decision to undergo implantation is moderated by whether the T3IC is "for me" or "for my child."

### Models of Technology Acceptance: Cognitive and Normative Variables

The cognitive variables "PU" and "PEU" are more or less selfexplanatory. The former refers to how the technology is perceived to help the user enhance his or her performance, whereas the latter refers to how its use is perceived to be "free of effort" (Davis, 1989). Their influence on the intention to use implants in adults has been proven with regard to physical implants for cosmetic surgery (Adams, 2010), as well as technological implants to address health issues, such as submammary defibrillators, cardiac resynchronization therapy devices, cardioverter defibrillators, or pacemakers (Giudici et al., 2010). "PU" also plays a key role in the intention to use cochlear implants in children (Christiansen and Leigh, 2004; Christie and Bloustien, 2010). These studies support the idea that most parents have high expectations that cochlear implants will help their children improve their verbal communication, educational options, and emotional well-being, often beyond the standard level among the deaf population (Li et al., 2003).

In addition to utilitarian factors, other factors also influence the decision to undergo implantation. Most et al. (2007) highlighted the importance of family environment in attitudes toward cochlear implants. In the same vein, Hyde et al. (2010) noted that parents often found the information provided by professionals insufficient to judge the implications of cochlear implants. In making such a decision, most parents consult other families with implanted children, and children with implants themselves, and highly value their support and the information they provide about their own personal experiences (see also Fitzpatrick et al., 2008).

With regard to body modification for strictly aesthetic purposes, Adams (2010) and Javo and Sørlie (2010) also established the influence of family and friends on the decision to undergo cosmetic surgery in adults. Additionally, for this area of surgery in particular, the social pressure to maintain a youthful and attractive image seems to be crucial (von Soest et al., 2006; Dorneles de Andrade, 2010). The variable "subjective (or social) norm" captures the influence of others' opinion on one's choices. Social norms are expected to play a key role in assessments of potential implantation, especially implantation in children.

Based on the aforementioned observations regarding TAM models, and findings in the therapeutic field, the following hypotheses were proposed with regard to T3ICs:


#### Affective Variables

A novel extension of models is proposed here in order to capture the effect of emotions on the acceptance of new technologies. Both medical advances in transplant technology, such as organ transplants from animals, and the integration of technological devices have led to the perception that the body is modifiable (Christie and Bloustien, 2010), which, in turn, can generate apprehension and anxiety (Buchanan-Oliver and Cruz, 2011) and fear of dehumanization (Lai, 2012). With regard to young people's perceptions, Schaar and Ziefle (2011) analyzed, through qualitative methods, benefits, and fears regarding four implantable medical devices known to different degrees and entailing different levels of surgical risk for the patient: pacemakers, cochlear implants, medical chips, and deep brain stimulation (or "brain pacemakers") for the treatment of Parkinson tremor and paralysis. The results showed that, when deciding, young people also make a trade-off between the perceived benefits and risk or fears. Moreover, there was a negative relationship between self-reported technical literacy and risk perception triggered by unspecified concerns and fears resulting from the lack of knowledge of these implants and of technology in general.

With regard to implanting one's children, numerous studies have shown that the process of deciding on cochlear implantation is stressful for the parents of deaf children (Richter et al., 2000; Spahn et al., 2001; Weisel et al., 2006; Zaidman-Zait, 2008). Moreover, parental stress during the pre-examination stage seems to be relatively higher for those with children who are still verbally competent (i.e., borderline cases for cochlear implantation), and for whom this option is therefore not selfevident (Burger et al., 2005). Nevertheless, marked relief is experienced after the initial fitting of the cochlear implant. That is, over time, parents adapt to the new situation and begin to perceive the benefits and adjust their expectations accordingly.

All in all, this evidence supports the inclusion of affective variables in the model of technology acceptance. Doing so enables it to distinguish between emotions that stimulate action, namely implantation, and emotions that inhibit or change the course of action (Oliver et al., 1997; O'Neill and Lambert, 2001; White and Yu, 2005; Cohen et al., 2006). In general, actions associated with positive emotions are evaluated favorably, whereas actions triggering negative emotions are evaluated unfavorably (Bagozzi et al., 1999; Mano, 2004). There is also a natural tendency to avoid decisions that generate bad feelings (Elliott, 1998; Schwarz, 2000; Han et al., 2007).

Using the multidimensional structure of affect (Watson et al., 1988) as a reference, the following hypotheses were proposed:


### The User Moderating Effect: T3ICs "for me" vs. T3ICs "for my child"

According to Chorney et al. (2015), there is a considerable level of decisional conflict when making a decision about surgical treatment for one's child. Hyde et al. (2010) showed how parents found it very difficult to come to a decision on cochlear implantation for children with substantial residual hearing, due to uncertainty regarding improvement. These parents are under greater pressure to make decisions and take on responsibility, and are thus more severely affected by nervous stress (affective

factors; Burger et al., 2005). Nevertheless, Li et al. (2004) found that two thirds of the parents of children who were eligible for cochlear implantation were actually considering it. Parents whose children were eligible but who did not consider implantation prioritized bilingual success (verbal and sign language). Indeed, parents might encounter great social pressure in favor of cochlear implantation as opposed to the alternative: letting their child live a "deaf life" (Li et al., 2003; Hyde and Power, 2005; Fitzpatrick et al., 2006). A parallel can be drawn between consenting for children to increase their innate capacities or letting them develop naturally. All in all, the evidence suggests that the factors influencing the decision to undergo implantation and the intensity of their effect may differ depending on whether the intended implant recipient is oneself or one's child.

Notably, affect and normative factors can be relatively more important when deciding for one's children. In a meta-analysis, Lipstein et al. (2012) concluded that a variety of factors influence parents' decisions in the field of pediatric surgery, including personal factors, emotions, and the opinions of other community members. In the same vein, Li et al. (2004) found that aside from medical recommendations, parents' values, beliefs, and expectations about the outcomes of implantation influence the decision to allow their children to undergo cochlear implantation. Moreover, these factors are particularly relevant when parents are considering an early intervention, that is, when there is not enough information confirming that implants are the option that will yield the best outcomes. Indeed, despite the considerable achievements of cochlear implants, this technology still raises questions and poses conflicts and difficulties among parents of children with hearing impairments. The refusal of some parents to have their children implanted may be due to the possibility that children's hearing expectations will remain unfulfilled, which, in turn, could affect their selfesteem (Most et al., 2007). That is, they are seeking to protect their children from disappointment. In addition, Most et al. also demonstrated the existence of a social group identity among some adult deaf people. These people fear that cochlear implantation will lead to a loss of this identity in young deaf people, with no guarantee that the implant will even work as expected. At the opposite extreme, cosmetic surgery in children is sometimes permitted by parents. In many cases, the aim is to achieve not a normal appearance, but an outstanding one (i.e., to look better than the average) (Gilbert, 2009). Accordingly, the following hypothesis was proposed, under the assumption that it can affect all key relationships specified in the model:

H7. The intention to use T3ICs is moderated by whether the T3ICs are "for me" or "for my child," which involves differences in the explanatory variables affecting the intention to use T3ICs and the intensity of their effect.

### The Conceptual Model

The formulated hypotheses define a proposal for a comprehensive theoretical model of variables influencing the intention to use T3ICs, namely, the CAN model shown in **Figure 1**.

## METHODOLOGY

### Data Collection and Sample Characteristics

Data were retrieved from a self-administered, online survey. More than 3500 invitations to participate in the study were sent out. Only individuals over the age of 16 and residing in Spain could participate. The actual sample selected for the study consisted of 600 randomly selected individuals, proportionally

#### TABLE 1 | Technical details of the data collection and sample description.


distributed according to gender and age quotas. It is worth noting that the classification variable "implanted" and "non-implanted" participants was collected in the survey by the double question: Do you have any implants?(due to cosmetic/health-related reasons) Of which type? The percentage of implanted participants in the selected sample was however very low, some 8.7%. Full details are given in **Table 1**.

This study was approved by the Ethics Responsible at the Faculty of Business Administration of the University of La Rioja, and according to ICC/ESOMAR International Chamber of Commerce/ESOMAR (2008). Each participant provided informed consent.

### Statistical Analysis

In order to test the working hypotheses, a sequential process was followed consisting of the following steps, which are summarized in **Figure 2**. Partial least squares structural equation modeling (PLS-SEM) was chosen to test the CAN model, as it is less sensitive to violations of data normality (Chin, 1998). The software used was SmartPLS 3.0.


## RESULTS

### "Intention to Use" and "Predicted Use" of "T3ICs for me" and "T3ICs for my Child"

**Table 2** shows the descriptive mean, standard deviation, and median for the variables "intention to use" and "predicted use" and the groups T3ICs "for me" and T3ICs "for my child." There are significant differences in the mean values for the "for me" and "for my child" groups. The mean values for "intention to use" and "predicted use" for the "T3ICs for me" group were around 4.6 (on a scale of 0–10). These mean values were lower for the "T3ICs for my child" group (around 3.9). However, there was a high dispersion of mean values for "intention to use" and "predicted use" for both groups (close to 3.3 in both cases). Median values were therefore estimated, which differ by one point.

Together, the results show that the acceptance of T3ICs is higher when they are for oneself than when they are for one's child. Moreover, the high dispersion of mean values for both groups justifies testing the proposed model for explaining the acceptance of T3ICs and the differences found for this acceptance between the "T3ICs for me" and "T3ICs for my child" groups.

### Exploratory Factor Analysis and Validation of the Factors Formed from the Observable Variables

Exploratory factor analysis was carried out to test the factors naturally formed from the observable variables, i.e., the measurement scales. Based on the results from the exploratory factor analysis, "PU", "PEU", "subjective norm" (SN), and "intention to use" (IU) were all formed of a single factor with high variance explained: PU = 91.90% (KMO = 0.845), PEU = 91.97% (KMO = 0.875), SN = 97.71% (KMO = 0.500), and IU = 96.46% (KMO = 0.500). Bartlett's sphericity tests were significant for all the aforementioned scales (p < 0.001).

TABLE 2 | "Intention to use" and "predicted use" of T3ICs "for me" and T3ICs "for my child."


Mean, standard deviation (SD), and median. SD stands for standard deviation.

The affective scale, however, was more complex. This scale was formed of three differentiated factors, "positive emotions," "negative emotions," and "anxiety," which together explained 73.15% of the variance in the intention to use T3ICs. The KMO index was 0.941, and the Bartlett's sphericity test was significant (p < 0.001). "Positive emotions" included feeling enthusiastic, determined, proud, inspired, strong, active, interested, and excited, that is, positive feelings toward the use of T3ICs. "Negative emotions" included feeling hostile, upset, irritable, ashamed, and guilty. Finally, "anxiety" involved being afraid, scared, jittery, alert, nervous, distressed, and attentive.

#### Assessment of the Measurement Model

The assessment of the measurement model was carried out in two steps. First, item validity was examined. This was assessed in terms of the standardized loadings (>0.70) and t-values (>1.96) (Hair et al., 2013). The latter indicates the significant contribution of a variable to the content validity of the corresponding factor. Exceptionally, a significant variable can be kept in the model to the detriment of the standardized loading (Hair et al., 2013). Based on these criteria, it was decided to remove the variables "alert" and "attentive" (not significant) and to preserve the variable "ashamed" (standardized loading of 0.690 and t-value of 12.448 for the "T3ICs for me" group, and of 0.744 and 13.384, respectively, for the "T3ICs for my child" group). The standardized loadings and t-values of all the variables included in the final model are shown in the Appendix.

Second, the measurement model was verified in terms of construct reliability (i.e., composite reliability and Cronbach's Alpha), convergent validity, and discriminant validity. The composite reliability and Cronbach's Alpha values were all above 0.70. The convergent validity of the constructs was also satisfied, with an average variance explained (AVE) above 0.5 in all cases. The discriminant validity of the constructs was measured through the comparison of the square root of AVE vs. the correlations among constructs (Roldán and Sánchez-Franco, 2012). The square root of AVE (diagonal elements in bold in **Table 3**) has to be larger than the corresponding inter-construct correlations (off-diagonal elements in **Table 3**). This criterion was also met in all cases.

#### Assessment of the Structural Model

The CAN model greatly explained the intention to use T3ICs. The R <sup>2</sup> was 73.8% for T3ICs "for me" and 75.9% for T3ICs "for


#### TABLE 3 | Construct reliability, convergent validity, and discriminant validity of T3ICs "for me" and T3ICs "for my child."

CR stands for composite reliability. AVE stands for average variance explained. Diagonal elements (in bold) are the square root of the AVE. Off-diagonal elements are the inter-construct correlations.

#### TABLE 4 | Effect on endogenous variables.


my child" (**Table 4**). Stone-Geisser's cross-validated redundancy Q<sup>2</sup> was >0 in both cases, specifically, 0.718 for T3ICs "for me" and 0.740 for T3ICs "for my child." These results further confirmed the predictive relevance of the CAN model (see Hair et al., 2011a,b). The variance explained by each factor and for each group are also shown in **Table 4**.

The sign, magnitude, and significance of the path coefficients and the R 2 are shown in **Figure 3** and **Table 5**. Three hypotheses were fully supported by the results: H3 (regarding the influence of the "subjective norm"), H4 (regarding "positive emotions"), and H6 (regarding "negative emotions") (**Table 5**). These relationships were significant in both models, and the direction set coincided with that hypothesized. Two hypotheses were partly supported: H1 and H2. Cognitive factors affected the intention to use the technology on oneself but not on one's child. The relationship was only significant, and only coincided with the direction set, in Model 1 ("T3ICs for me"). H5 (regarding "anxiety") was rejected, as the relationship was not significant in either group.

#### Multi-Group Analysis

A multi-group analysis was performed to compare the results of the models for each group. Two parametric and two non-parametric tests were used to analyze differences in the key relationships between the models and to further assess possible moderating effects (**Tables 6, 7**).

Column PEV in **Table 6** shows the p-values obtained applying the method proposed by Chin (2000). This method assumes that

the data is normally distributed and/or that the variances of the two samples are similar (Afonso et al., 2012). Column PW−<sup>S</sup> shows the p-values obtained applying the Welch–Satterthwaite test in the cases where the variances of the two samples were different. The results of these two parametric tests were similar.

As for the non-parametric tests (Sarstedt et al., 2011), column P<sup>H</sup> in **Table 6** shows the p-value obtained applying the Henseler test. **Table 7** shows the results of the test of confidence intervals, the second non-parametric test used. The criteria establish that when the parameters estimated through confidence intervals for the two groups overlap, a significant difference can be established between the two group-specific path coefficients.

In three out of the four tests performed, a significant difference was found between the two groups regarding the key relationship between "PEU" and "intention to use"; the exception was the result of the confidence intervals test. However, because the confidence intervals test is relatively more conservative than the other three tests (Sarstedt et al., 2011), the significance between the groups of the difference in the relationship between "PEU" and "intention to use" was accepted. The relationship between "subjective norm" and "intention to use" was significantly different according to the four tests performed. No significant differences were found in other key relationships.

#### TABLE 5 | Path coefficients, t-values, and support for the hypotheses.


\*p < 0.05 => t > 1.65; \*\*p < 0.01 => t > 2.33; \*\*\*p < 0.001 => t > 3.09; n.s. = not significant [based on t(4.999), one-tailed test]. Please note that the levels of significance (p-values) are either "non significant" or lower than 0.01.

#### TABLE 6 | Multi-group comparison.


Levels of significance based on Student t(4.999) distribution with two tails. PEV = p-value equivalent variances test. PW−<sup>S</sup> = p-value Welch–Satterthwaite test. P<sup>H</sup> = p-value Henseler test.

### DISCUSSION, IMPLICATIONS, LIMITATIONS, AND FUTURE RESEARCH

This study introduces a model that combines cognitive, affective, and normative factors to explain the acceptance of a new insideable technology, namely, technological implants to increase innate human capacities (T3ICs). The CAN model largely explains the intention to use the technology for the specified groups, with a variance explained (R 2 ) of 73.8% for Group 1 ("T3ICs for me") and of 75.9% for Group 2 ("T3ICs for my child"). The variables contributing the most were found to be "positive emotions" and (positive) "subjective norm."

The CAN model is based on previous models of technology acceptance, specifically, TAM models (Davis, 1989; Davis et al., 1989; Venkatesh and Davis, 2000) and UTAUT models (Venkatesh et al., 2003, 2012). These models merely include cognitive and normative variables. Venkatesh et al. (2012) have contributed significantly to the application of these models, developing the UTAUT model, which obtained a PLS R <sup>2</sup> of 44% for direct effects. This value compares to the substantially higher R <sup>2</sup> obtained through the inclusion of the emotional dimension in the CAN model: R 2 above 70% for both of the specified groups. These results thus confirm the benefits of extending the factors determining the acceptance of a new technology to include the emotional dimension of consumer behavior. Affective factors greatly contribute to explaining underlying motives influencing subjects' assessment of products (Pieters and van Raaij, 1988; van Waterschoot et al., 2008; Levav and McGraw, 2009; Zielke, 2011) through variables such as "positive emotions," "negative emotions," and "anxiety."

The results showed that the acceptance of T3ICs was higher when the intended recipient was oneself than when it was one's child. Moreover, statistically significant differences were found between the two models/groups—"T3ICs for me" and "T3ICs for my child"—when applying the multi-group comparison to the three dimensions specified in the CAN model, namely, the cognitive, affective, and normative dimensions.

The cognitive variables "PU" and "ease of use" influenced the intention to use T3ICs on oneself, but not on one's child. That is, neither "PU" nor "ease of use" had a (positive) significant effect on the intention to have one's child implanted (H1 and H2 were partly supported). The between-groups comparison yielded statistically significant differences regarding the variable "PEU" in at least three of the four multi-group tests applied. The moderating effect of the end user in this relationship was thus accepted. It is worth noting, however, that significant differences between the groups were not found with regard to the variable "PU." The reason for this is two-fold: first, the scant contribution of this variable to overall variance explained (only 5.13% in the reference group "T3ICs for me"), and second, the equally low difference in standardized loadings between the two groups (only 0.044 points).

In conclusion, the cognitive variables included in the CAN model have only a limited influence when the end user is oneself, and no influence when the end user is one's child. The results slightly modify those of previous studies, such as Li et al. (2003), Christiansen and Leigh (2004), and Christie and Bloustien (2010), that have shown the importance of "PU" in the decision to implant one's child for health-related reasons. In the current study, implantation is not performed for medical reasons,



Sig. denotes a significant difference at 0.05; n.s. denotes a non-significant difference at 0.05.

and this variable ceases to have an effect. Likewise, it can be concluded that "PU" is a relevant factor in the decision of whether to get a T3IC, but is not as essential a factor as it is in the decision of whether to undergo cosmetic surgery (see Adams, 2010).

Regarding the normative dimension, in both groups the variable "subjective (or social) norm" positively influenced the intention to use T3ICs (H3 was supported). Moreover, the four multi-group analyses revealed significant differences between the two groups for this normative variable. This variable largely explains the intention to use T3ICs in the "for my child" group (42.45%, the highest variance explained), while it is the second variable in terms of the percentage of variance explained (25.19%) for the "for me" group. These results are consistent with those of von Soest et al. (2006), Most et al. (2007), Hyde et al. (2010), Adams (2010), Javo and Sørlie (2010), and Dorneles de Andrade (2010), who established that family, friends, and society influence the decision to undergo changes in the body. However, the current study has shown that this influence is stronger when the decision affects one's child than when it affects oneself.

As for the emotional dimension, both positive and negative emotions affected the intention to use T3ICs in the direction established in the CAN model. H4 and H5 were thus supported. In both groups, positive emotions explained the intention to use T3ICs (31.88% for the "T3ICs for me" group and 26.78% for the "T3ICs for my child" group). Negative emotions explained the intention to use them in both groups to a lesser extent (4.44% and 3.26%, respectively). The multi-group analysis showed no significant differences between the two groups regarding the influence of these two variables on the intention to use T3ICs. These findings support studies establishing that positive emotions promote a positive assessment of a technology (Bagozzi, 1997; Shiv and Fedorikhin, 1999). However, the proposal of a natural tendency to make decisions that minimize the probability of negative emotions occurring (Elliott, 1998; Schwarz, 2000; Han et al., 2007) was found to have little bearing, at least in this case.

The influence of "anxiety" on the intention to use T3ICs was non-significant in the models of both groups (H6 was rejected). The results of the multi-group analysis showed no differences between the two groups. This lack of influence of "anxiety" is contrary to the research of Buchanan-Oliver and Cruz (2011), which determined the anxiety produced by the idea of the dissolution of the limits of what is human due to the introduction of implants. However, it is consistent with what Venkatesh et al. (2003) demonstrated when they developed the UTAUT model.

Finally, as detailed before, differences were found between the two groups for some of the key relationships specified in the CAN model. H7 can thus be partly accepted. Another contribution derived from the results of this study is related to the demonstration of the moderating effect of the end user on the acceptance of a new technology. When one is considering the decision to get T3ICs for one's child, social influence is the principal factor. At a considerable distance from the "social norm," emotions, especially "positive emotions," have a secondary level of importance. In this case, it did not matter whether the T3ICs were considered to be useful or easy-to-use; neither of these variables significantly explained the intention to use T3ICs. When the T3ICs were for oneself, "positive emotions" were what most greatly explained the intention to use T3ICs, followed closely by "social norm."

The CAN model proved useful for explaining the intention to use a technology in the early stages of adoption; however, the importance of the variables in explaining such an intention to use varied depending on the moderating variable "for me" vs. "for my child." When studying the acceptance of a technology, one should thus distinguish between when the recipient is oneself and when it is another user. Lastly, the low explanatory power of cognitive variables may be due precisely to the early stage of development in which the studied technology currently finds itself. Consumers know very little about the usability of the products and are more concerned about what they feel and what others will think of them (i.e., social norm).

### Implications of the Results

This work opens a new line of research on the acceptance of a technology integrated into the human body with psychosociological implications for the evolution toward a human with superior capacities. Challenges for companies selling T3ICs involve two main aspects, convincing society of the goodness of T3ICs and generating positive emotions toward this type of product. When parents are deciding on getting T3ICs for their children, the battle must be won by convincing society of the goodness of these implants, since the social norm is the most important aspect. On the other hand, when the decision is about whether or not to implant oneself, the development of marketing communications generating positive emotions toward T3ICs would be the most effective way forward. Encouraging the idea that T3ICs are useful and easy-to-use, while important, may be of secondary concern. Nonetheless, sooner or later companies will have to address these barriers and opportunities.

### Limitations and Future Research

So far, the CAN model has been applied to the general idea of T3ICs. However, the results could vary if it were applied to a particular type of T3IC. "PU" and "ease of use" could, for instance, acquire more relevance. Therefore, as a future line of research, the CAN model could be applied to specific types of T3ICs to observe consumers' reactions. In addition, the model was tested on an emerging product, and it was not possible to ascertain whether its explanatory power would be similarly high for more widely used technology products. The CAN model should thus be tested on these and other products with the same degree of diffusion, as well-known types of implants would most likely be assessed differently, a factor that might also depend on respondents' technological literacy (see Schaar and Ziefle, 2011). In this regard, cultural factors might be also relevant, since different countries have different degrees of technological proneness and literacy (Alagöz et al., 2011). Other factors worth considering are gender and age differences in attitudes and acceptance of technological implants, also in relation to general attitudes toward technology (self-reported technological interest, literacy, handling competence, and distrust in technology) (Ziefle and Schaar, 2011). Moreover, future research could focus on analyzing the differences on the acceptability of T3ICs depending on whether individuals have already undergone implantation due to cosmetic/health-related reasons. This would, however, involve a new sample including a larger percentage of implanted participants.

With regard to the decision to implant a third party, the current study was limited to comparing the acceptance of implants "for me" vs. "for my child." However, several companies in the US already use insideable tech to identify their workers. Moreover, one proposed public health system in the US also contemplates the possibility of requiring people to get ID implants. Such events have generated public

### REFERENCES


controversy. In this regard, another possible line of research would be to examine the reactions arising from having the decision to undergo implantation be imposed by an external authority.

Finally, the current study does not take the ethical component into account. The authors believe that these types of products could greatly exacerbate social differences. A society could emerge made up of an implanted elite alongside non-implanted children who would be unable to compete to reach the same levels of development as their implanted counterparts. It is thus essential for future research dealing with this issue to return as much information as possible to society in order to enable the type of informed decision-making that will be essential to our progress as social human beings.

### AUTHOR CONTRIBUTIONS

The directors of research have been the professors JP, ER, and CO. The three co-authors have participated in all stages of work, including the conception and design of the research, the revision of intellectual content and drafting the work. MG has participated in the revision of intellectual content and the drafting and translation of the article.

### ACKNOWLEDGMENTS

This work has been funded by The Ministry of Economy and Competitivity (Spain), Research Project with reference: ECO2014-59688-R, Programa Estatal de Investigación, Desarrollo e Innovación Orientada a los Retos de la Sociedad, Plan Estatal de Investigación Científica y Técnica y de Innovación 2013-2016.

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fpsyg. 2016.00132


Software Systems Engineering and Information Systems, eds M. Mora, O. A. Gelman, and M. Steenkamp (Hershey, PA: Raisinghan Information Science Reference), 193–221.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Pelegrín-Borondo, Reinares-Lara, Olarte-Pascual and Garcia-Sierra. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# How Can We Improve Patient Satisfaction As a Consumer of Public Health Services? The Case of Psychiatric Patients Undergoing Electroconvulsive Therapy

Carmen Selva-Sevilla<sup>1</sup> \*, Patricia Romero-Rodenas <sup>2</sup> and Marta Lucas-Perez-Romero<sup>2</sup>

<sup>1</sup> Department of Applied Economics, University of Castilla-La Mancha, Albacete, Spain, <sup>2</sup> Department of Psychiatry, University Hospital Complex of Albacete, Albacete, Spain

Keywords: mental disorders, electroconvulsive therapy/psychology, patient satisfaction, family/psychology, costs of health care

#### Edited by:

Ana I. Jiménez-Zarco, Universitat Oberta de Catalunya, Spain

#### Reviewed by:

Antoni Olive-Tomas, IQS - Universitat Ramon Llull, Spain Clara Viñas-Bardolet, Universitat Oberta de Catalunya, Spain

#### \*Correspondence:

Carmen Selva-Sevilla carmen.selva@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 26 March 2016 Accepted: 12 May 2016 Published: 26 May 2016

#### Citation:

Selva-Sevilla C, Romero-Rodenas P and Lucas-Perez-Romero M (2016) How Can We Improve Patient Satisfaction As a Consumer of Public Health Services? The Case of Psychiatric Patients Undergoing Electroconvulsive Therapy. Front. Psychol. 7:801. doi: 10.3389/fpsyg.2016.00801 Electroconvulsive therapy (ECT) involves the induction of seizures for therapeutic purposes in cases of serious psychiatric disorders. Since the early twentieth century, ECT was one of the most important therapeutic alternatives to treat patients suffering from severe psychiatric disorders, and ECT maintained its therapeutic role despite the emergence of psychoactive drugs by mid-century. However, in the 1970's it fell into disuse due to the emergence of a psychiatric paradigm shift that considered it repressive and unsafe. The introduction of anesthesia as a fundamental part of the ECT has significantly improved the safety of the therapy, which, today, continues to be an effective therapy for its precise indications (Lopez Villaescusa et al., 2011; Rodriguez-Jimenez, 2015).

### SHOULD WE TAKE INTO ACCOUNT THE OPINION OF THE PSYCHIATRIC PATIENT UNDERGOING ECT?

The concept of "patient-centered care" consists of giving the patient an important role in making decisions about their health and, therefore, supporting the idea that clinical practice decisions should consider the patient's perspective. Consideration of the concept of patient satisfaction as an outcome of health care involves changes in the evaluation and improvement of the clinical practice (Mira and Aranaz, 2000).

For two decades, the so-called evidence-based medicine has emerged as a new paradigm in medicine. In this conception of medicine, the patient is considered to have an active part in the process of making decisions regarding their disease (Sackett et al., 1996). However, the real clinical application of the methods of evidence-based medicine has been distorted, especially with regard to the inclusion of patients in decision-making processes. There is currently a movement that calls for re-assessing the actual application that it is made of evidence-based medicine with the aim of returning to its original basis. In this sense, it is considered a key point the fact that patients will be taken into account in decision-making processes (Greenhalgh, 2014).

According with the proposal above mentioned, the mental health field has proposed that there is a need to evaluate the decision-making ability of psychiatric patients to ascertain if it is feasible to consider them in decision-making (Villagran et al., 2014). Even patients undergoing ECT have been considered for their degree of satisfaction with this therapy (Chakrabarti et al., 2010). In this sense, the success of healthcare would be related to patient satisfaction, perceived by them not only as the health outcomes achieved but also for the means used to achieve such outcomes (Mira and Aranaz, 2000).

In short, we believe that the outcome of healthcare should be measured in terms of not only effectiveness, safety, and efficiency but also the patient's perception of their autonomy and their sense of physical or mental well-being.

### WHAT DOES THE PSYCHIATRIC PATIENT UNDERGOING ECT NEED IN ORDER TO IMPROVE THEIR SATISFACTION WITH THE ASSISTANCE RECEIVED?

The inclusion of patient satisfaction as a criterion to develop clinical practice guidelines (Greenhalgh, 2014) is a highly recommended practice in the particular case of the psychiatric patient (Lasalvia and Ruggery, 2007; Khöler et al., 2015), and this approach is recommended because the rate of success of the chosen therapy may be much more dependent on psychosocial factors than in cases of non-psychiatric disorders.

The first step in learning the concerns of patients is asking for their opinion. Research about psychiatric patients' satisfaction usually involves the employment of questionnaires which are administered to patients and relatives. By doing so it has been identified that their perception about the health care was most influenced by four items: the competence of the care professionals who serve them, being given complete information about their diagnosis and treatment, being involved with their own treatment plan, and having the usefulness of their hospitalization explained to them (Berghofer et al., 2001; Blenkiron and Hammill, 2003; Gani et al., 2011; Fernández-Carbonell et al., 2012). ECT exhibits significant differential nuances, and therefore, specific questionnaires have been developed to ascertain the concerns of these patients, such as whether they face ECT with fear or if they believe that ECT is dangerous or painful, among others (Lopez Villaescusa et al., 2011; Rajagopal et al., 2013; Rodriguez-Jimenez, 2015).

### WHAT SHOULD BE DONE FROM THE POINT OF VIEW OF CLINICAL PRACTICE TO IMPROVE SATISFACTION OF THE PATIENTS UNDERGOING ECT?

At present, ECT and psychotropic drugs coexist for the treatment of certain psychiatric disorders, but in some clinical conditions ECT is considered first-line treatment. Some of these indications arise from medical issues, such as when teratogenic psychotropic drugs must be avoided because of pregnancy, for instance, or when an outcome is required as rapidly as possible for patients suffering with severe life-threatening conditions. One of the main indications for ECT is, paradoxically, the patients' preference to receive ECT (Rodriguez-Jimenez, 2015). In addition, based on performance criteria, ECT has been more effective than drug treatment in some specific indications, such as severe depressive disorders (The UK ECT Review Group., 2003). This improved effectiveness affects the length of hospital stay for patients, which is shorter in duration (Markowitz et al., 1987) and which then leads to lower health care direct costs.

Considering the above information, our proposal for patients to undergo ECT earlier than is currently done seems feasible. To achieve this, a first step would be for physicians to lose their fear of ECT and consider applying it, in certain indications, without or before exhausting all of the possibilities of psychotropic drugs.

A second proposal would be to promote the use of ECT in continuation/maintenance cycles, which has been well appreciated by patients and physicians responsible for the therapy. ECT administered in continuation/maintenance cycles has been proven to have clinical efficacy and economic advantages by reducing the number of total days of hospitalization, the number of emergency room visits, and the number of emergency admissions for recurrences. In this regard, direct costs associated with the above mentioned concepts were lower after a continuation/maintenance program was implemented, as compared with the period before such program was implemented (American Psychiatric Association Committee on Electroconvulsive Therapy, 2001; Perestelo-Perez et al., 2013; Rodriguez-Jimenez et al., 2015).

A third proposal would be to act on those aspects related to ECT methodology by modifying both anesthetic and psychiatric protocols. Currently, there is a great disparity in the way in which ECT is performed between centers with regard to specific aspects, such as the employed hypnotic drug or the lower limit for an adequate electrical convulsion (Lihua et al., 2014). It is possible that the development of clinical guidelines based on the best available evidence and its widespread use afterwards would help to improve the quality of ECT sessions. This could then shorten the duration of the therapeutic process, which in turn would result in lower health care costs and, of course, in increased satisfaction of patients and relatives.

### WHY IS OUTPATIENT ECT NOT USED?

The most effective way to shorten a hospital stay is by avoiding admittance. Currently, in most of the Mental Health Departments, patients remain hospitalized while undergoing all of the prescribed ECT sessions for the cycle of therapy. However, there is a real possibility to perform ECT in an outpatient way. These programs consist of admitting the patients to the hospital only for the duration of each ECT session, so that they remain at home for the duration of the prescribed ECT cycle (Kramer, 1990; Chan et al., 2006; Rodriguez-Jimenez, 2015). Of course, ECT outpatient programs should not be applicable for all psychiatric patients, but only to those patients meeting specific criteria, such as those that do not require continued psychiatric care and have good support from their relatives. These selected patients would benefit from the general advantages associated with any outpatient treatment, such as staying at home with their family for as long as possible; in our opinion, this course of action would yield clear psychological benefits for the psychiatric patients. It must be remembered that admission to a Unit of Acute Mental Disorders itself carries a burden of psychosocial stress that is linked to the feeling of a loss of freedom, which may be associated with an unfavorable outcome in this type of patient population.

In short, we believe that performing ECT in an outpatient way has two theoretical advantages: from a clinical point of view it can positively influence the way that the patients perceive their treatment process, and from an economic point of view it would reduce the costs derived from hospital stay.

### HOW WOULD THESE ACTIONS THAT ARE AIMED TO IMPROVE CLINICAL PRACTICE IMPACT THE SATISFACTION OF PSYCHIATRIC PATIENTS UNDERGOING ECT?

In our opinion, it would improve their perception of the quality of received health care in three areas: clinical, psychological, and economic.

The clinical improvement would be because the quality of every single ECT session and the quality and efficacy of the therapy as a whole would be improved, as would the perception of the process by the patients.

The psychological improvement would be because improving the efficacy of clinical treatment and reducing the hospital stay would increase patients' comfort, as they could stay at home with their relatives, and would improve the psychological attitude of the patient toward ECT, as it would not be experienced as a negative element in their life.

#### REFERENCES


The economic improvement would be because shortening the hospital stay would yield economic savings that could improve opportunity cost, and resources could be returned to the patients in other useful ways such as the acquisition of better ECT equipment and specialized medical staff.

Our final conclusion is that using patients' satisfaction as an indicator of the quality of health care processes should not be neglected when establishing programs to improve health care. In the particular case of psychiatric patients undergoing ECT, achieving the best possible patient satisfaction with the mental health care received is very important because it would have great influence on treatment adherence and, therefore, on clinical outcome.

#### AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct, and intellectual contribution to the work, and approved it for publication.

### ACKNOWLEDGMENTS

We would like to express our sincere gratitude to our good friend Manuel Geronimo-Pardo for his patience, generosity, enthusiasm, and inmense knowledge. We greatly appreciate his assistance in research and writing, and his valuable comments and suggestions to improve the quality of the manuscript.


Health System public hospital: a case series. Rev. Psiquiatr. Salud. Ment. 8, 75–82. doi: 10.1016/j.rpsm.2014.10.002


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Selva-Sevilla, Romero-Rodenas and Lucas-Perez-Romero. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# The Psychiatric Patient as a Health Resource Consumer: Costs Associated with Electroconvulsive Therapy

Carmen Selva-Sevilla<sup>1</sup> \*, Maria Luisa Gonzalez-Moral<sup>2</sup> and Maria Teresa Tolosa-Perez<sup>3</sup>

<sup>1</sup> Department of Applied Economics, University of Castilla-La Mancha, Albacete, Spain, <sup>2</sup> Department of Clinical Analysis, University Hospital Complex of Albacete, Albacete, Spain, <sup>3</sup> Department of Psychiatry, University Hospital Complex of Albacete, Albacete, Spain

Background: Clinical practice protocols should consider both the psychological criteria related to a patient's satisfaction as a consumer of health services and the economic criteria to allocate resources efficiently. An electroconvulsive therapy (ECT) program was implemented in our hospital to treat psychiatric patients. The main objective of this study was to determine the cost associated with the ECT sessions implemented in our hospital between 2008 and 2014. A secondary objective was to calculate the cost of sessions that were considered ineffective, defined as those sessions in which electrical convulsion did not reach the preset threshold duration, in order to identify possible ways of saving money and improving satisfaction among psychiatric patients receiving ECT.

#### Edited by:

Ana I. Jiménez-Zarco, Open University of Catalonia, Spain

#### Reviewed by:

Imran Ali, King Abdulaziz University, Saudi Arabia Francesc Saigí Rubió, Open University of Catalonia – Internet Interdisciplinary Institute, Spain

> \*Correspondence: Carmen Selva-Sevilla carmen.selva@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 26 March 2016 Accepted: 10 May 2016 Published: 27 May 2016

#### Citation:

Selva-Sevilla C, Gonzalez-Moral ML and Tolosa-Perez MT (2016) The Psychiatric Patient as a Health Resource Consumer: Costs Associated with Electroconvulsive Therapy. Front. Psychol. 7:790. doi: 10.3389/fpsyg.2016.00790 Methods: A descriptive analysis of the direct health costs related to ECT from the perspective of the public health system between 2008 and 2014 was performed using a retrospective chart review. All of the costs are in euros (2011) and were discounted at a rate of 3%. Based on the base case, a sensitivity analysis of the changes of those variables showing the greatest uncertainty was performed.

Results: Seventy-six patients received 853 sessions of ECT. The cumulative cost of these sessions was €1409528.63, and 92.9% of this cost corresponded to the hospital stay. A total of €420732.57 (29.8%) was inefficiently spent on 269 ineffective sessions. A sensitivity analysis of the economic data showed stable results to changes in the variables of uncertainty.

Conclusion: The efficiency of ECT in the context outlined here could be increased by discerning a way to shorten the associated hospital stay and by reducing the number of ineffective sessions performed.

Keywords: mental disorders, electroconvulsive therapy, patient satisfaction, costs and cost analysis, costs of health care, direct service costs

## INTRODUCTION

An economic evaluation of health care is needed because financial resources are limited and must therefore be allocated in the most efficient way. In the field of health, such evaluations would prove useful because they would help decision-makers identify the best ways to allocate resources, which would be immensely helpful for making decisions both at the general level of health by Health Administration and at a more specific level by health staff (Drummond et al., 2015).

In recent years, there has been an ongoing discussion regarding the concept of "patient-centered care"; this concept considers allowing patients to play an important role in making decisions about their own health. Furthermore, clinical practice decisions should consider not only factors related to the effectiveness and efficiency of a treatment but also the perspective of the patient, specifically his/her satisfaction with the treatment process (Mira and Aranaz, 2000).

The subject of this work is the more specific case of the psychiatric patient, for whom satisfaction with the care process is evaluated by questionnaires fulfilled by both the patient and his/her relatives. These questionnaires have identified the most important items that influence a positive perception of health service by the patient. Those items include (1) the competence of care professionals, (2) being given complete information about the diagnosis and treatment, (3) patient involvement in the treatment plan, and (4) providing an explanation for the usefulness of the hospitalization (Berghofer et al., 2001; Blenkiron and Hammill, 2003; Gani et al., 2011; Fernandez-Carbonell et al., 2012).

Electroconvulsive therapy (ECT) is currently considered an effective treatment for patients with severe psychiatric disorders, especially for severe depressive disorders (mainly psychotic depression), some cases of acute mania and schizophrenic disorders (Bernardo, 1999; American Psychiatric Association Committee on Electroconvulsive Therapy, 2001; Rodriguez-Jimenez, 2015a). ECT consists of the application of electrical stimulation to the brain in order to trigger a generalized seizure. It is accepted that the optimal response to ECT is related to the duration of the triggered seizures, although there is no consensus for what the limit should be. ECT is administered two or three times per week for the acute phase of the disease, until between 6 and 12 sessions have been administered; usually, acute phase treatment is performed on an in-patient basis (Sackeim, 1991; Bernardo, 1999; American Psychiatric Association Committee on Electroconvulsive Therapy, 2001; Ding and White, 2002; Gonzalez et al., 2007; Perestelo-Perez et al., 2013; Rodriguez-Jimenez, 2015a). After completing the acute phase, the patients may possibly continue to receive additional ECT sessions as part of a continuation cycle or a maintenance cycle, depending on whether these sessions are applied during the first 6 months after the acute phase (continuation cycle to prevent relapses) or after 6 months (maintenance cycle to prevent recurrence; American Psychiatric Association Committee on Electroconvulsive Therapy, 2001). However, it is a more common practice to indiscriminately call them continuation/maintenance (Rodriguez-Jimenez, 2015a).

An ECT program was implemented in our hospital to expand treatment alternatives beyond drug treatment. Regarding the Spanish health care system, economic evaluations focused on ECT are scarce, and to the best of our knowledge, there has been no evaluation of the economic impact of the so-called ineffective sessions. The aim of this work was to identify possible ways of optimizing the use of health and economic resources, which in turn would improve patient perception of the treatment process.

Specifically, the main objective of this study was to determine the cost associated with the sessions of ECT that were implemented in our hospital between 2008 and 2014. A secondary objective was to calculate the cost of sessions that were considered ineffective, defined as those sessions in which the electrical convulsion did not reach the preset threshold duration.

### MATERIALS AND METHODS

A retrospective chart review of all patients who had received at least one ECT session from April 2008 to December 31st, 2014 while admitted at the Mental Health Service in a 752-beds tertiary hospital was conducted. No exclusion criteria were considered.

A partial economic evaluation consisting of a description of costs was done from the perspective of the public health system. Only direct health costs were considered for our analysis, which include those costs incurred by the health system as a direct result of the disease (Oliva et al., 2015).

The cost of a standard ECT session was calculated as the product of health resources employed (measured in natural units and obtained from the routine records of each ECT session) for their unitary prices (measured in monetary units of the 2011 base year, discounted at a rate of 3%). The cost of an ECT session for each patient and each given cycle (the acute phase versus the continuation/maintenance phase) was also calculated. For this purpose, the cost of the hospital stay associated with ECT, the cost of the time spent by the medical staff, the consumables and drugs employed, and the cost of the purchase of the ECT equipment were considered, with the following observations:

### Hospital Stay

Information about the costs of a 24-h hospital stay at the Department of Mental Health was expressed in euros per day and was provided by the Control Service Management of our hospital. The total hospital stay was obtained from patients' charts. For the purpose of this study, three different time periods of hospitalization were defined: pre-ECT, ECT, and post-ECT. The total days of admittance to the Mental Health Department were calculated by subtracting the date of hospital discharge from the date of hospital admittance. The days attributable to the ECT period were calculated by subtracting the date of the last ECT from the date of the first ECT session and then adding 2 days to account for anesthetic evaluation. If the patient had been transiently discharged during the weekend holidays, then these days were subtracted from the days attributable to the ECT period.

### Medical Staff and Nursing

Information about the salary (expressed in euros per minute of time spent by the professionals involved) was provided by our Service Management Control. To calculate the cost per staff caused by this activity, a typical ECT session was considered to be 30 min of time consuming activity for the following professionals: three physicians (two from the Mental Health Department and

one from the Anesthesiology Department) and three nurses (two from the Anesthesiology Department and one from the Mental Health Department). The time spent by other professionals, such as assistants and orderlies, was not considered for this analysis.

### Drugs and Consumables

Information concerning the acquisition cost of drugs and consumables used in an ECT session according to our protocol was provided by the Hospital Pharmacy and Management Control Services.

### Equipment

The equipment with which the ECT was applied was ABBOT SPECTRUM brand (model 5000Q, EQ-ELMED class), and it was acquired in December, 2007 for €23500. This information was provided by the Management Control Service. The equivalent annual cost (at an interest rate of 3% and considering a useful life of 10 years) was calculated since its acquisition, and this cost was divided by the number of annual sessions of ECT conducted in the study period.

The cost of all of the ECT sessions was calculated by adding the cost of each session received by all of the patients. The cost of the ineffective sessions was calculated in the same way, i.e., by adding the cost of each ineffective session of all patients. Following the established protocol of the Mental Health Department, ECT sessions were considered ineffective when an electrical convulsion lasted less than 20 s. A more conservative approach was taken to assess the cost of ineffective sessions, considering those sessions lasting 20 or more seconds after re-stimulation as not ineffective, and by also not considering titration sessions because titration sessions are used to determine the minimum threshold electrical charge for each patient in each acute phase cycle and are, therefore, associated with a higher percentage of ineffective sessions.

In addition, these cost calculations were repeated considering 25 s, not 20 s, as the minimal duration for an acceptable electrical convulsion because 25 s is the most frequently cited limit in the psychiatric literature.

A sensitivity analysis was conducted, considering variations from the base case, in terms of personnel costs, the discount rate and the average cost per hospital stay.

#### Statistical Plan

This is a descriptive study in which continuous variables were described using means and standard deviations, and the qualitative variables were described using absolute and relative frequencies.

#### Ethics Statement

This chart review study was approved by the Ethics Committee and Clinical Research of our hospital.

### RESULTS

During the implementation of the ECT program at our hospital from April 2008 to December 2014, 76 patients received at least one ECT session, accounting for 853 sessions. **Table 1** shows the characteristics of the patients and the ECT sessions. For 13 sessions (1.5%), the information about seizure duration was not available, and the sessions were classified as effective to avoid bias favoring the economic burden of ineffective sessions (our second objective).

The cost of a standard ECT session was €1667.35. The distribution of the costs derived from a standard ECT session is shown in **Table 2**, where it can be seen that the hospital stay represented the main economic burden.

The cumulative cost of all of the 853 ECT sessions administered was €1409528.63 (**Table 3**). Of those, €420732.57 (29.8%) was inefficiently employed for performing 269 ineffective sessions, defined as when the electrical seizure did not reach 20 s. Following our conservative approach, 31 titration sessions lasting less than 20 s with an associated cost of €52165.90 were excluded from the ineffective group, and there were also 62 sessions lasting more than 20 s after re-stimulation with an associated cost of €98671.55. Therefore, using a more conservative approach €269865.42 (19.1%) were determined to be used inefficiently.

By pure logic, setting 25 s as the lower limit for an adequate electrical convulsion would increase the number of inefficient sessions (426, **Table 1**) and the cost associated to inefficient sessions (48.3%, **Table 3**).

As stated above, the hospital stay represented the largest portion of the total cost (92.9%; **Table 3**). Before it had been decided to treat the patients with ECT (pre-ECT period), they had been admitted for a median of 15.5 days (interquartile range 5–25.25), representing a cost of €891564.38. Considering this time course, the cost of treating psychiatric patients would have increased by more than 60%, specifically from €1409528.63

TABLE 1 | Clinical and demographic characteristics of patients undergoing electroconvulsive therapy.


ECT, electroconvulsive therapy; GCI-I, Global Clinic Impression-Improvement scale.



<sup>a</sup>All costs are in euros of 2011 and are discounted at a rate of 3%. <sup>b</sup>Cost per hospital stay is expressed in euro/day. <sup>c</sup>Staff costs are expressed in euros/minute. <sup>d</sup>These calculations correspond to the first session of each cycle because the nasal cannula is re-used in every electroconvulsive therapy session for the same patient. <sup>e</sup>Cost of equipment has been obtained by calculating the Annual Equivalent Cost (at an interest rate of 3% and useful life of 10 years) and dividing it by the number of annual sessions of electroconvulsive therapy.

to €2301093.01. In contrast, during the ECT period, patients had 149 days of weekend discharges, resulting in a savings of €89102.57 (5.9%).

The sensitivity analysis of the economic data shows stable results for variations of the uncertainty variables (**Table 4**).

#### DISCUSSION

The main results of our work are, first, that the hospital stay is responsible for more than 90% of the total costs attributed to ECT and, second, that there are a large number of ineffective ECT sessions, which implies that almost 30% of the economic resources had been inefficiently used.

Regarding the economic evaluation of ECT, there have been studies that have reported a cost-effectiveness analysis or cost description analysis in which ECT is compared with pharmacotherapy (Aziz et al., 2005; Greenhalgh et al., 2005) or other techniques such as repetitive transcranial magnetic stimulation (Kozel et al., 2004; McLoughlin et al., 2007; Knapp et al., 2008; Vallejo-Torres et al., 2014), as well as studies focusing on continuation/maintenance programs (Bonds et al., 1998; McDonald et al., 1998; Aziz et al., 2005; Odeberg et al., 2008; Rodriguez-Jimenez et al., 2015b). In general, these studies have concluded that ECT is cost-effective.

Based on this premise, we have conducted a purely descriptive study of the direct costs that our hospital has incurred after implementing an ECT program. According to our results, the cost of a standard ECT session was €1667.35. In our environment, Vallejo-Torres et al. (2014) found a cost of €737, but they only considered the costs of staff, materials, equipment, tests, and drugs in their analysis, and they did not include the cost of the hospital stay. If they had included this cost, then the cost of a standard ECT session would rise to €1475 (referred to in 2013 euros and ascribing 2 days of stay for each ECT session), a figure very similar to what we found.

#### TABLE 3 | Total cost of electroconvulsive therapy sessions (2008–2014).


ECT, electroconvulsive therapy. All costs are expressed in constant euros of 2011 and discounted at a rate of 3%.

#### TABLE 4 | Sensitivity analysis.


Total cost of electroconvulsive therapy sessions (2008–2014). All costs are expressed in constant euros of 2011.

In view of our results, we propose two ways to improve the efficiency of the ECT: reducing the number of ineffective sessions and shortening the hospital stay of the patients. By attaining these objectives, two advantages would be expected: first, an improvement in the efficiency of the treatment itself and, second, a consideration of the patients' preferences in line with the above mentioned "patient-centered care." By implementing these actions, patients would be allowed to return to their homes and resume their activities earlier, which, in turn, would be perceived with great satisfaction both for the patients and their relatives.

Regarding the objective of reducing the number of ineffective sessions, we found that the percentage of funds invested in ineffective sessions (those lasting less than 20 s; Bertolin-Guillen et al., 2006; Ramirez-Segura and Ruiz-Chow, 2013) ranged between 29.8 and 19.1%, as assessed either in a less or more conservative approach, respectively. A clear limitation of this assessment is the assumption that the clinical efficacy of ECT depends solely on the electrical length of the evoked seizure, when there is no evidence that this assumption is true (Fear et al., 1994; Geretsegger et al., 1998; Sackeim, 1999; Azuma et al., 2007; Bauer et al., 2009; Stewart, 2012; Martinez-Amoros et al., 2014). In fact, different values for the lower limit of the electrical convulsion can be found in the literature: 15 s (Gonzalez et al., 2007), 20 s (Bertolin-Guillen et al., 2006; Ramirez-Segura and Ruiz-Chow, 2013), 25 s (Nguyen et al., 1997; Dogan et al., 2011; Gombar et al., 2011; Hizli Sayar et al., 2014; Martinez-Amoros et al., 2014), and 26 s (Wang et al., 2011). These data preclude us from directly comparing the percentage of sessions that were considered ineffective. To perform an indirect comparison with other authors who used propofol as a hypnotic and to consider the lower limit of efficacy at 25 s, we re-analyzed our data after establishing the 25 s limit. Using this approach, our percentage of ineffective sessions would rise to 50%, a value that is much lower than that found by Martinez-Amoros et al. (2014; 82.9%) but much higher than that found by Nguyen et al. (1997; 4.2%). This huge disparity in the percentage of ineffective sessions could be explained by the small number of sessions held by these authors: 35 sessions (Martinez-Amoros et al., 2014) and 22 sessions (Nguyen et al., 1997).

However, despite the considerable number of so-called ineffective sessions that have taken place, the efficacy of ECT in our patients measured by the CGI-I was 61.4%. This finding was obtained as a result of the sum of cycles of treatment whose clinical development had been "very much improved" and "much improved," as shown in **Table 1**.

Our percentage of ineffective sessions (19.1%, conservative estimate) suggests that there is room for improvement in the performance of the ECT sessions. This improvement would include modification of the guidelines of psychiatry and anesthesiology, which is beyond the scope of this work.

Regarding the second aspect of our investigation, we did a thorough analysis of the costs associated with ECT and found that micro-costs (i.e., those costs associated with consumables, equipment and drugs) could be ignored in future studies because they represent a practically negligible percentage of the total cost (less than 4%). Conversely, the hospital stay represented the highest percentage of economic burden; therefore, any strategy to save costs should be aimed at reducing the hospital stay of the patients.

We found an average delay of 2 weeks from hospital admittance to inclusion in the ECT program. It was not the aim of this work to study the costs associated to this pre-ECT period, but they were definitely significant costs assumed by the Health System and a clear target for some cost-saving measures. This delay occurs because it is routine clinical practice of our Mental Health Department physicians to exhaust drug therapy before entering the patients into an ECT program.

To reduce the hospital stay, one possibility could be to exhaust drug therapy in an out-patient setting for those patients whose clinical condition would permit out-patient management; therefore, in the event that they must be admitted to receive ECT, it could be performed soon after admittance. For those patients whose clinical condition would not allow for out-patient management before drug alternatives had been exhausted, ECT could be considered earlier than when it is currently indicated. In either case, the length of the hospital stay would be shortened (Kramer, 1990; Chan et al., 2006).

Another way to reduce the hospital stay would be by increasing weekend discharges. Until now, this option has been scarcely adopted. We did not review the full clinical aspects of the patients; thus, based on our data, it is impossible to determine whether more weekend discharges were possible for clinical reasons, which represents a clear limitation to our statement. However, weekend discharges in our Mental Health Department are not a standard protocol; therefore, we do believe that there is a potential to save costs in this way.

Another way to reduce the duration of hospital stays would be to assess the performance of ambulatory ECT (Kramer, 1990; Chan et al., 2006; Rodriguez-Jimenez, 2015a) instead of the traditional custom of performing ECT in an in-patient manner for the entire duration of the corresponding cycle. Selected patients, such as those not requiring ongoing psychiatric care, could benefit the obvious advantages of ambulatory treatment, such as staying at home with their relatives for as long as possible, which, in turn, could significantly contribute to improving their clinical condition and perhaps would also improve the patients' perception of and satisfaction with their psychiatric treatment. Ambulatory ECT is already a real option in our context (Lopez Villaescusa et al., 2011; Rodriguez-Jimenez, 2015a).

This work has several limitations. Some of the limitations have been exposed above, for example, the assumption that the clinical efficacy of ECT depends solely on the electrical length of the seizure, not having analyzed the causes behind the delay between the hospital admission and the decision to indicate ECT, and not explaining the limited number of weekend hospital discharges of the patients. In addition to these limitations, this work has the general limitations of any retrospective review (i.e., a lack of information in medical records) and a more specific economic limitation, that is, that the average costs of the hospital stay (cost per diem) were used to assess the cost of the stay (Drummond et al., 2015).

### CONCLUSION

fpsyg-07-00790 May 26, 2016 Time: 15:10 # 6

Electroconvulsive therapy efficiency could increase by reducing the number of ineffective sessions and by reducing hospital stay. Whether a reduction in the number of ineffective sessions could be attained by modifying the anesthesiologist-psychiatric protocol deserves future research, as well as a reduction in hospital stay by encouraging hospital permissions and ambulatory ECT. On the basis of these assumptions were true, the patients' perceptions of their treatment process would be improved for three main reasons: clinically, because their treatment efficacy would be increased; psychologically, because their hospital admittance would be shortened in time or even avoided; and economically, because by applying the concept of opportunity cost, the economic savings could be employed to benefit the patients even if they are unaware of these savings.

### REFERENCES


### AUTHOR CONTRIBUTIONS

CS-S has done the economical evaluation and has contributed to statistical analysis. MG-M has done the main part of statistical analysis. MT-P has reviewed the clinical charts. All authors have made substantial, direct and intellectual contribution to the work, and approved it for publication.

### ACKNOWLEDGMENTS

We would like to express our sincere gratitude to our good friend Manuel Geronimo-Pardo for his patience, generosity, enthusiasm, and immense knowledge. We greatly appreciate his assistance in research and writing, and his valuable comments and suggestions to improve the quality of the manuscript.



continuation/maintenance electroconvulsive therapy in a Spanish National Health System public hospital: a case series. Rev. Psiquiatr. Salud. Ment. 8, 75–82. doi: 10.1016/j.rpsmen.2014.10.005

Sackeim, H. A. (1991). Are ECT devices underpowered? Convuls. Ther. 7, 233–236.

Sackeim, H. A. (1999). The anticonvulsant hypothesis of the mechanisms of action of ECT: current status. J. ECT 15, 5–26. doi: 10.1097/00124509-199903000- 00003


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer FSR and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.

Copyright © 2016 Selva-Sevilla, Gonzalez-Moral and Tolosa-Perez. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Influence of Customer Quality Perception on the Effectiveness of Commercial Stimuli for Electronic Products

#### Álvaro Garrido-Morgado\*, Óscar González-Benito and Mercedes Martos-Partal

Departamento de Administración y Economía de la Empresa, Universidad de Salamanca, Salamanca, Spain

#### Edited by:

Maria Pilar Martinez-Ruiz, University of Castilla-La Mancha, Spain

#### Reviewed by:

Noemi Martinez-Caraballo, Centro Universitario de la Defensa de Zaragoza/University of Zaragoza, Spain Doreen Pick, Freie Universitaet Berlin, Germany

> \*Correspondence: Álvaro Garrido-Morgado algamo@usal.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 16 October 2015 Accepted: 23 February 2016 Published: 14 March 2016

#### Citation:

Garrido-Morgado Á, González-Benito Ó and Martos-Partal M (2016) Influence of Customer Quality Perception on the Effectiveness of Commercial Stimuli for Electronic Products. Front. Psychol. 7:336. doi: 10.3389/fpsyg.2016.00336 Creating and maintaining customer loyalty are strategic requirements for modern business. In the current competitive context, product quality, and brand experience are crucial in building and maintaining customer loyalty. Consumer loyalty, which may be classified into cognitive loyalty and affective loyalty, is related to customers' quality perception. Cue utilization theory distinguishes two dimensions for perceived quality, extrinsic quality—linked to the brand—and intrinsic quality—related with internal product characteristics. We propose that (i) cognitive loyalty is more influenced by intrinsic product quality whereas extrinsic product quality (brand name) is more salient for affective loyalty and, (ii) different commercial stimuli have a differential effectiveness on intrinsic and extrinsic perceived quality. In fact, in this study, we analyze how perceived quality dimensions may influence the effectiveness of two different commercial stimuli: displays and advertising flyers. While displays work within the point of sale under time-constrained conditions where consumers are more likely to use heuristics to simplify their decisions, advertising flyers work outside of the point of sale under low time-constrained conditions, and therefore favor a more reasoned purchase decision where systematic processing will be more likely. We analyze the role of quality perception in determining the effectiveness of both these commercial stimuli for selling products that induce high purchase involvement and perceived risk. The empirical analysis focuses on computer products sold by one of Europe's largest computer retailers and it combines scanner, observational, and survey data. The results show that both dimensions of quality perceptions moderate the influence of displays and advertising flyers on sales, but their impact is different on each commercial stimuli. Extrinsic quality perception increases to a greater extent the effect of displays due to the use of a brand name heuristic. However, intrinsic quality perception improves to a greater extent the effect of advertising flyers, which in turn are more closely related to systematic decision processing.

Keywords: quality perceptions, customer loyalty, commercial stimuli, displays, advertising flyers, cue utilization theory, consumer information processing, time theory

## INTRODUCTION

Today the current economic situation is reflected in the slowdown in household consumption expenditure in Europe in recent years. In fact, some Euro area countries, such as France, Germany, Italy or Spain, have presented negative year-on-year average rates in 2013 or 2014 (The World Bank, 2015). Therefore, the economic situation causes increasingly more competition for both retailers and manufacturers that are struggling to attract consumers to their stores and products (Lin et al., 2013). Because of this, retailers and manufacturers try to analyze consumers' needs and preferences in order to be able to adapt the offer to increase customer loyalty to their stores/brands. Therefore, they adjust their prices, improve their product quality and increase their communication effort to make their new offers more noticeable in order to improve their sales (Cant and Hefer, 2014).

In fact, retailers use different commercial stimuli that take place both inside and outside the store. On one hand, instore stimuli, such as merchandising tools or special displays, are noticeable by customers inside the store at the moment in which they carry out their purchase decision (Bava et al., 2009; Bezawada et al., 2009; Phillips et al., 2015). Therefore, instore stimuli trigger unrecognized needs and desires or trigger memories for forgotten needs, leading to in-store decision making, or unplanned purchasing (Inman et al., 2009), such as the purchase of one particular substitutive product or one particular brand instead of another. These stimuli usually trigger unplanned purchases, which now represent about 70% of total purchases (Stilley et al., 2010; Bell et al., 2011). On the other hand, out-of-store commercial stimuli, such as advertising flyers, are mainly used by retailers to increase traffic to the store or to publicize other promotions (Gijsbrechts et al., 2003; Schmidt and Bjerre, 2003; Ziliani and Ieva, 2015). Advertising flyers are sent to potential consumers' homes, facilitating their purchase planning (Schmidt and Bjerre, 2003), as they may carry out a more reasoned purchase decision under low time-constrained conditions.

The academic literature on commercial stimuli effectiveness has analyzed the moderating role of specific attributes of the stimuli on sales (e.g., redemption time in coupon effectiveness or placement of the display or number of pages and size of the advertising flyers; e.g., Bawa, 1996; Gijsbrechts et al., 2003; Bezawada et al., 2009; Barat and Ye, 2012; Luceri et al., 2014) or the effect of more general characteristics, such as consumer characteristics (e.g., price sensitiveness, value consciousness, coupon properness, variety seeking, impulsive personality, or planned and unplanned consumer; e.g., Bawa, 1996; Laroche et al., 2003; Wakefield and Inman, 2003; Swaminathan and Bawa, 2005; Govindasamy et al., 2007; Haans and Gijsbrechts, 2011; Lin et al., 2013; Gázquez-Abad et al., 2014; Gázquez-Abad and Martínez-López, 2016). Moreover, previous studies have shown the moderating role of product and brand characteristics (storage and perishable conditions, impulse purchase, hedonic/utilitarian nature, interpurchase cycle, brand tiers) on the effectiveness of commercial stimuli (e.g., Narasimhan et al., 1996; Lemon and Nowlis, 2002; Pauwels et al., 2002; Wakefield and Inman, 2003; Inman et al., 2009; Castro et al., 2013).

One very important product characteristic is customers' perceived product quality, as it is a key determinant in building and maintaining customer loyalty (Brakus et al., 2009; Pan et al., 2012; Akdeniz et al., 2014). Despite this, the research literature on the influence of perceived quality on commercial stimuli is limited. Furthermore, previous studies such as Lemon and Nowlis (2002) use price tiers as proxy of perceived quality and do not distinguish between different quality attributes. Product quality comprises two types of attributes: extrinsic and intrinsic attributes (Bava et al., 2009; Gooner and Nadler, 2012; Akdeniz et al., 2014). Whereas, extrinsic attributes (e.g., brand name) are more related to affective loyalty (customers build affect toward the brand on the basis of cumulatively satisfying usage occasions), intrinsic attributes have a more objective nature (e.g., the technical characteristics of electronic products), which can be compared easily by seeking out information about the product (Szibillo and Jacoby, 1974; Richardson et al., 1994) and, therefore, are more related to cognitive loyalty. Consequently, in this paper we attempt to fill in this gap by analyzing whether perceived product quality (extrinsic and intrinsic attributes) has a moderating role in the effectiveness of different commercial stimuli (product displays and advertising flyers).

The objective of this paper is to extend the research on commercial stimuli effects on sales by (i) examining and comparing the direct effect of one in-store stimuli (special displays) and one out-store stimuli (advertising flyers) on the sales of infrequently purchased products with high perceived risk and high involvement and (ii) analyzing the moderating role of perceived quality (intrinsic and extrinsic attributes) on the effectiveness of both commercial stimuli on sales.

Our paper contributes to the marketing, retail and consumer behavior literature by providing a theoretical framework based on cue utilization theory and consumer processing information and by making an empirical analysis to ascertain the differences in effectiveness of commercial stimuli regarding both dimensions of perceived quality on sales of technological products. Specifically our paper contributes to previous research in four ways.

First, we consider the two dimensions of perceived quality, intrinsic (product characteristic) and extrinsic (brand). Previous research has not analyzed the moderating role of intrinsic perceived quality and brand perceived quality on the effectiveness of commercial stimuli. The present study seeks to extend and improve the insights of the few previous studies that have attempted to capture the quality effect on effective commercial stimuli. For example, Lemon and Nowlis (2002) use price tiers as proxy of perceived quality of the brand. Liu-Thompkins and Tam (2013) have researched the differential effectiveness of cross-selling promotion on attitudinal loyal customer and spurious loyalty, but they ignore the differences between intrinsic perceived quality-cognitive loyalty and extrinsic perceived quality-affective loyalty which may undermine the effectiveness of commercial stimuli. Therefore, this study aims to analyze the effectiveness of different commercial stimuli depending on the intrinsic perceived quality of product and brand.

Second, we combine three data sources in our analysis: scanner data and observational data and customer surveys to conduct the empirical analyses. Previous studies assume the brand quality or use price tiers as proxy of perceived quality and do not distinguish between different quality attributes (e.g., Lemon and Nowlis, 2002). By contrast, this study obtains customer perceived quality by collecting a consumer survey (extrinsic perceived quality) or analyzing the product characteristics (intrinsic perceived quality).

Third, we study a high involvement product. Most prior studies about the effectiveness of commercial stimuli analyze frequently purchased products with low perceived risk and low involvement (e.g., Burton et al., 1999; Lemon and Nowlis, 2002; Gijsbrechts et al., 2003; Van Heerde et al., 2004; Bezawada et al., 2009). We study a high involvement product to generalize previous findings about the effect of commercial stimuli on frequently purchased products or detecting different effects due to differences in this type of product, such as the greater effect of perceived quality (extrinsic and intrinsic) compared to price. In this paper we study infrequently purchased products (computers), which involve high customer perceived risk and involvement. Previous studies on sales of technological products have analyzed the direct effect of the brand and objective quality on product sales (Neelamegham and Chintagunta, 2004; Sriram et al., 2006); however, the effect of different commercial stimuli have not been previously analyzed. Identifying the drivers of customer loyalty can help us attain a better understanding of consumer behavior and also allow resources to be allocated among marketing tactics.

Four, we analyze the effect of two different commercial stimuli, one in-store stimulus (product displays which represent a time-constrained condition) and one out-of-store stimulus (advertising flyers which represent a low time-constrained condition) on the sale of computer products. Some studies have used different tools at the same time (Narasimhan et al., 1996; Van Heerde et al., 2004; Ailawadi et al., 2006; Haans and Gijsbrechts, 2011); however, most of them use these commercial stimuli as control variables when they analyze promotion effectiveness on frequently purchased products and, therefore, do not compare and explain their effect on product sales depending on the customer perceived quality as we do in this study.

The next section offers a conceptual framework that leads directly into the study hypotheses. After a description of the methodology, we provide details on the empirical analysis and results, along with their interpretations. The final section summarizes the conclusions and implications of the present study.

### CONCEPTUAL FRAMEWORK: PROPOSED HYPOTHESES

#### Customer Perceived Quality

Quality is a central element in business strategy and academic research. Firms compete on quality, customers search for quality, and markets are transformed by quality (Golder et al., 2012). In the marketing literature, some researchers distinguish between objective and subjective quality (e.g., Mitra and Golder, 2006) whereas others, such as Zeithaml (1988), notes that "objective quality may not exist because all quality is perceived by someone." In this paper we focus on customer perceived quality.

Perceived quality is defined as a buyer's estimate of a product's cumulative excellence (Zeithaml, 1988). Perceived product quality is a key variable in the consumer decision process (Steenkamp, 1990) and it is considered a pivotal determinant of shopping behavior and product choice (e.g., Zeithaml, 1988; Grewal et al., 1998; Wang, 2013; Akdeniz et al., 2014).

According to information economics (Nelson, 1970, 1974), consumers have uncertainty about the quality attributes and benefits of the products they aim to purchase because of the imperfect, asymmetric information that characterizes most product markets. Companies are more informed about their products than customers, so firms can behave opportunistically. To overcome that uncertainty, companies must inform consumers and give them cues about their credibility (Erdem and Swait, 1998; Akdeniz et al., 2014). Cue signals mostly serve as heuristics in assessing product quality when (1) there is a need to reduce the perceived risk of purchase, (2) the consumer lacks expertise and consequently the ability to assess quality, (3) consumer involvement is low, (4) objective quality is too complex to assess or the consumer is not in the habit of spending time objectively assessing quality, or (5) there is an information search preference and need for information (Dawar and Parker, 1994). Therefore, in technological products (e.g., computers) the use of cues will be very useful to simplify customer decision process.

Cue utilization theory considers that products consist of a set of cues that serve as surrogate indicators of quality to consumers and specifies that, when consumers make inferences about perceptions (Olson, 1978). Cues are represented by the set of attributes related to the product they are assessing. The particular cues are evoked according to their predictive and confidence values. The predictive value of a cue is the degree to which consumers associate a given cue with product quality. The confidence value of a cue is the degree to which consumers have confidence in their ability to use and judge that cue accurately (Richardson et al., 1994).

These cues can be either extrinsic or intrinsic. The former relate less closely to the product, such that changes to the extrinsic cue do not necessarily entail changes in product attributes (e.g., brand names, packaging, and product communication).The latter include attributes whose modification would involve a change in the physical properties of the product (e.g., ingredients in food products or technical characteristics in computers). Research evidence suggests that consumers tend to use both extrinsic and intrinsic cues concurrently when evaluating product quality (Szibillo and Jacoby, 1974; Richardson et al., 1994; Gooner and Nadler, 2012; Akdeniz et al., 2014).

Dawar and Parker (1994) propose that the relative importance of these cue signals generally follow their specificity, or the extent to which a particular signal is not shared across competitive products. A brand name, for example, is typically shared by only a few products within a competitive line of products and is therefore a very specific signal. Physical features, on the other hand, can be shared to a greater extent across competing products and are therefore less specific. The more specific a signal, all else being equal, the more likely it will provide information that is useful in an assessment of product quality. This distinction is consistent with the belief that cue signals are relied on as a function of their predictive value.

Consumers use cues to develop beliefs about products and that task response (i.e., choice or evaluation) may be a direct function of these mediating beliefs (Olson, 1978; Dawar and Parker, 1994; Richardson et al., 1994; Aqueveque, 2006, 2008; Akdeniz et al., 2014).

### Customer Brand Loyalty

Customer loyalty remains a topic of great interest for firms (Kotler and Keller, 2009; Garnefeld et al., 2013; Wieseke et al., 2014). Consumers exhibit behavioral loyalty when they repeatedly patronize a business, often to the exclusion of competing offers. Behavioral loyalty is desirable from a financial perspective, for example, because superior brand performance outcomes such as greater market share and price premiums relate to customer brand loyalty (Chaudhuri and Holbrook, 2001).

In a comprehensive discussion of customer brand loyalty, Oliver (1999) defines loyalty as "a deeply held commitment to rebuy or repatronize a preferred product/service consistently in the future, thereby causing repetitive same-brand or same brandset purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior." According to his framework, attitudinal loyalty addresses the psychological component of a consumer's commitment to a brand and may encompass beliefs of product/service superiority as well as positive and accessible reactions toward the brand. This definition emphasizes the two different aspects of brand loyalty: behavioral and attitudinal (Dick and Basu, 1994; Chaudhuri and Holbrook, 2001; Garnefeld et al., 2013).

Oliver's (1999) framework proposes a cognition-affectconation-action framework with four loyalty phases. In the first loyalty phase, the brand attribute information available to the consumer indicates that one brand is preferable to its alternatives. This stage is referred to as cognitive loyalty, or loyalty based on brand belief only. Cognition can be based on prior knowledge or on recent experience-based information. Loyalty at this phase is directed toward the brand because of this information (attribute performance levels).

In the second phase of loyalty development, a liking or attitude toward the brand has been developed on the basis of cumulatively satisfying usage occasions. Commitment at this phase is referred to as affective loyalty and is encoded in the consumer's mind as cognition and affect.

The next phase of loyalty development is the conative (behavioral intention) stage, influenced by repeated episodes of positive affect toward the brand. Conation, by definition, implies a brand-specific commitment to repurchase.

In the action control sequence, the motivated intention in the previous loyalty state is transformed into readiness to act. The action control paradigm proposes that this is accompanied by an additional desire to overcome obstacles that might prevent the act. Action is perceived as a necessary result of engaging both these states. If this engagement is repeated, an action inertia develops, thereby facilitating repurchase.

In short, cognitive loyalty focuses on the brand's performance aspects, affective loyalty is directed toward the brand's likeableness, conative loyalty is experienced when the consumer focuses on wanting to rebuy the brand, and action loyalty is commitment to the action of rebuying (Oliver, 1999). Dick and Basu (1994) refer to behavioral loyalty without attitudinal loyalty as "spurious loyalty."

In the current competitive context product quality and brand experience are crucial in building and maintaining customer loyalty (Brakus et al., 2009; Pan et al., 2012; Schmitt et al., 2014; Wieseke et al., 2014; Lin et al., 2015).

Brand experience is defined as subjective, internal consumer responses (sensations, feelings, and cognitions) and behavioral responses evoked by brand-related stimuli that are part of a brand's design and identity, packaging, communications, and environments (Brakus et al., 2009). If a brand evokes an experience, this alone leads to loyalty. In addition, an experience may be the basis for more elaborative information processing and inference making that results in brand-related associations. In turn, these associations affect loyalty (Brakus et al., 2009).

In this paper, we focus on cognitive and affective loyalty. We propose that in the first loyalty phase (cognitive loyalty) the customer is more focused on utilitarian and performance product attributes, therefore intrinsic quality will be a more relevant signal in this phase. However, in the second phase (affective loyalty), because affect toward the brand has been developed on the basis of cumulatively satisfying experience occasions, the customer can be driven by the positive reaction toward a brand signal (extrinsic quality).

### Commercial Stimuli

Commercial stimuli are informative cues (Steenkamp, 1990) that aim to attract attention to and raise interest in a product, by featuring it inside the store (special location or displays) or outside the store (coupons and advertising flyers; e.g., Bawa, 1996; Gijsbrechts et al., 2003; Bezawada et al., 2009; Inman et al., 2009; Bava et al., 2009; Haans and Gijsbrechts, 2011; Luceri et al., 2014). These stimuli trigger a cognitive or emotional response to the featured product and, therefore, can become part of the set of considered options in the evaluation phase or choice, by enhancing its purchase probability (Yeung and Wyer, 2004; Chandon et al., 2009; Inman et al., 2009). This paper focus on two of the most used commercial stimuli: displays (as in-store stimuli) and advertising flyers (as out-of-store stimuli).

On one hand, in-store stimuli (special displays) can be considered as any special presentation in the store aimed at drawing attention to a product and increasing its sales (e.g., islands, end-of-aisle displays, shelf signages, etc.; Bezawada et al., 2009; Garrido-Morgado and González-Benito, 2015; Phillips et al., 2015). These special displays are noticeable to customers inside the store, i.e., at the moment in which they decide and carry out their purchase decision (Bava et al., 2009; Bezawada et al., 2009; Inman et al., 2009). Therefore, displays may trigger unrecognized needs and desires or trigger memories for forgotten needs, leading to in-store decision making, or unplanned purchasing (Inman et al., 2009). What is more, consumers tend to view them as special bargains and often buy from a displayed product, which they had no previous intention of buying (Lin et al., 2013). Thus, they are usually related to unplanned purchases, which now represent around 70% of total purchases (Stilley et al., 2010; Bell et al., 2011).

Several studies confirm that displays or merchandising techniques exert direct effects on sales (e.g., Lemon and Nowlis, 2002; Chandon et al., 2009; Bezawada et al., 2009; Inman et al., 2009), usually by adopting displays as a control variable in their analyses of the effects of promotions (e.g., Narasimhan et al., 1996; Van Heerde et al., 2004).

On the other hand, out-of-store stimuli (advertising flyers) are mass communication techniques that are mainly used by retailers to increase the store traffic as well as to publicize promotions or to increase purchases (Burton et al., 1999; Haans and Gijsbrechts, 2011; Luceri et al., 2014). Advertising flyers are sent to potential consumers' homes in order to remind consumers about the existence of a product or to inform them about any promotions to enhance the impact of those deals (Gijsbrechts et al., 2003; Schmidt and Bjerre, 2003; Luceri et al., 2014). Thus, advertising flyers work differently from displays because they may facilitate the purchase planning (Burton et al., 1999; Govindasamy et al., 2007).

Previous studies confirm that advertising flyers, in particular, exert direct effects on store traffic and store sales as well as product sales (Burton et al., 1999; Gijsbrechts et al., 2003; Schmidt and Bjerre, 2003; Haans and Gijsbrechts, 2011; Luceri et al., 2014). However, most of the previous literature tends to include advertising flyers in a "feature" control variable that refers to any external technique designed to increase sales of a product or attract consumers to the store when they focus on other issues such as promotions (Narasimhan et al., 1996; Van Heerde et al., 2004; Ailawadi et al., 2006).

Based on this previous literature, we expect a positive effect of both commercial stimuli, displays and advertising flyers, on sales. Therefore, we propose the following hypotheses:

H1: Product displays engender a positive effect on sales H2: Advertising flyers engender a positive effect on sales

### The Moderating Role of Perceived Product Quality on Commercial Stimuli Effectiveness

Studies about commercial stimuli usually take into account different typical characteristics of the analyzed stimuli and thus control for various particular aspects that can moderate their results (see **Table 1**). For example, studies that analyze special presentations at the point of sale control for the distance between the new position of the displayed product from its usual position or whether this new position is in proximity to a complementary product or another commercial stimulus (Bezawada et al., 2009; Chandon et al., 2009; Phillips et al., 2015). Studies focused on advertising flyers also account for their characteristics, such as the number of pages, geographic area in which they are launched, temporal frequency, or average discounts (Gijsbrechts et al., 2003; Luceri et al., 2014). Studies about promotions frequently control for their characteristics, such as whether they induce immediate benefits, for example a price discount, a gift, or additional amounts of the stimulated product (Hardesty and Bearden, 2003; Palazón and Delgado-Ballester, 2009; Yi and Yoo, 2011).

Furthermore, other studies analyze the key moderating role of consumers' profile on commercial stimuli effectiveness. For example, consumer characteristics such as coupon proneness, brand loyalty, store loyalty, value consciousness, price consciousness, or income (Bawa, 1996; Laroche et al., 2003; Wakefield and Inman, 2003; Swaminathan and Bawa, 2005) have an impact on the effectiveness of different promotions or stimuli.

Other important moderating issues of commercial stimuli effectiveness are product characteristics, such as its storage conditions, impulsive buying, hedonic/utilitarian benefit (Narasimhan et al., 1996; Wakefield and Inman, 2003), or brand characteristics (Nowlis and Simonson, 2000; Lemon and Nowlis, 2002). In this line, Lemon and Nowlis (2002) examine synergies between different types of promotions (price promotions, displays, and feature advertising) and characteristics of the brands that offer the promotions. They consider the pricequality tier of the brand as a moderator of the effectiveness of promotions. Specifically they consider leading national brands as the high quality tier and private labels or small share brands as the low-quality tiers by assuming price tiers as quality tiers. They found that high tier brands benefit more than low-tier brands from price promotions, display, or feature advertising when the promotional tools are used by themselves.

However, Lemon and Nowlis (2002), as in the majority of studies, performed their research on frequently purchased products and did not take into account that the product perceived quality is formed through both types of characteristics, product characteristics—intrinsic attributes—and brand characteristics extrinsic attributes—(Szibillo and Jacoby, 1974; Richardson et al., 1994). Their results may be moderated if the study focuses on infrequently purchased products (e.g., computers). This product features substantial technological components and greater perceived risk and high involvement, because of their complexity and dynamic evolution (Laurent and Kapferer, 1985; Neelamegham and Chintagunta, 2004; Sriram et al., 2006). For these reasons, consumers may be motivated to develop a more reasoned and planned process for purchasing this type of product and, therefore, the degree to which alternatives can be compared directly or must be considered individually may become a very important aspect of the decision task that can affect a wide variety of decisions (Hsee and Leclerc, 1998; Ritov, 2000) as well as the effectiveness of different commercial stimuli depending on the degree to which they favor a more reasoned purchase and allow products to be directly compared or considered separately.

Previous research shows that time constraints influence consumers' perceptions of quality (Suri and Monroe, 2003). Dhar and Nowlis (1999) found that under time pressure, consumers are more likely to consider unique features among choices and less likely to consider common features. In addition, their subjects recalled more features (unique and common) in the no time pressure condition than in the time pressure condition. Under time-constrained conditions, customers are more likely to use heuristics, such as the brand name heuristic, to simplify the cognitive task (Kaplan et al., 1993; Nowlis, 1995). Whereas, in low time-constrained conditions the opportunity to process

#### TABLE 1 | Summary of previous studies about commercial stimuli.


<sup>a</sup>This study focuses on promotions, but it adds display and feature as control variables.

information is high and then systematic processing will be more likely (Suri and Monroe, 2003). Moreover, an increase in time pressure led to a greater use of heuristics when the motivation to process information was high relative to when it was low (Sanbonmatsu and Fazio, 1990; Suri and Monroe, 2003), for example with a high involvement product like computers.

Displays usually isolate a displayed product from the rest of competing alternatives. These stimuli take place within the store where consumers have less time to reason the purchase; therefore, they trigger a more unplanned purchase and are more likely to use heuristics to simplify decisions. Displays are used for highlighting a particular product and consumers consider this product individually. In this context, the brand is a unique feature among choices and brand name could be used as a heuristic by the consumer as a risk reduction strategy (Fischer et al., 2010; Gooner and Nadler, 2012). Brands identify the source or maker of a product. Consumers recognize a brand and activate their knowledge about it (Zhang and Sood, 2002). Using what they know about the brand in terms of overall quality and specific characteristics, consumers can form reasonable expectations about the functional and other benefits of the brand. Consequently, brands contribute to reducing the consumer's (subjective) risk of making a purchase mistake (e.g., Kapferer, 2008; Keller, 2008) and may be used for avoiding a more complicated purchase process. Moreover, in the case of display, brand is a more salient cue because it is a more specific signal which provides more information about the quality of the product (Dawar and Parker, 1994).

Advertising flyers are out-of-store stimuli because they are usually sent to potential consumers' homes in order to facilitate their purchase planning and, ideally, to attract them to the store (Schmidt and Bjerre, 2003; Haans and Gijsbrechts, 2011). Advertising flyers allow consumers to have time to overthink the purchase decision, giving them time to search for information and evaluate the perceived quality of a product in relation to all its attributes before they even enter the point of sale (Gijsbrechts et al., 2003). Although advertising flyers also encourage separate evaluations of products, in this context in low time-constrained conditions the consumers have more time to reason the purchase. Thus, they have the opportunity to process more information and therefore it is more likely that consumers will carry out more systematic information processing in which they analyze the intrinsic characteristics of the product. They may easily compare information about different product alternatives by using other available communication channels like the Internet.

Thus, we propose that the perceived quality attributed to the brand is more salient, and thus engenders greater sales, with displays than with advertising flyers. However, the perceived quality attributed to the intrinsic product characteristics is more salient, and thus engenders greater sales, with advertising flyers than with displays. Therefore, in line with this argument, we present the following hypotheses:

H3: The perceived quality attributed to the brand enhances the effectiveness of displays more than it does the effectiveness of advertising flyers for sales.

H4: The perceived quality attributed to the intrinsic product characteristics enhances the effectiveness of advertising flyers more than it does the effectiveness of displays for sales.

### METHODS

### Study Context

For this study, we analyzed computer products instead of the more frequently analyzed grocery products or FMCG (Fast-Moving Consumer Goods). We considered this type of product to be the most appropriate for analyzing the influence of the quality attributed to the brand—very often used as a risk reduction strategy for this product category—and the intrinsic product quality characteristics—very easy to compare through objective measures like MegaBites or GigaBites—on the effectiveness of different commercial stimuli.

Although consumers usually plan their purchase of computers rather than buy them on impulse, they usually have product experience and high motivation to process information (Suri and Monroe, 2003); therefore, commercial stimuli can trigger the purchase of one particular SKU (Stock-Keeping Unit) or one brand instead of another. Also, this type of infrequently purchased product with substantial technological components in dynamic evolution is more complex and, therefore, is usually linked to lower knowledge, greater perceived risk and higher involvement (Neelamegham and Chintagunta, 2004; Sriram et al., 2006).

In fact, consumers often have a purchase intention about the product category they need or want to buy, but they do not know exactly what particular SKU or brand to buy. In these situations, they may decide what SKU or brand to acquire by planning their purchase through the search for information about the product category before going into the store. A second option, especially for consumers without much knowledge about the product, is to decide the SKU or brand within the store by paying attention to the commercial stimuli and sell staff's recommendations or, even, to brand perceived quality in order to simplify their purchase process (Hoyer and Brown, 1990). In fact, this second option is quite usual as purchases decided in-store represent around 70% of total purchases (Stilley et al., 2010; Bell et al., 2011).

### Study Data

We have combined three data sources in order to obtain information about computers offered for sale in one representative store in Spain of Europe's largest category killer computer retailer (Retail-Index, 2015): scanner data (to collect weekly data about units sold and prices), survey data (to collect assessment of perceived quality), and observational data (to capture the weeks in which displays or advertising flyers are used for promoting a particular SKU and for collecting from manufacturers' websites the quality attributed to intrinsic characteristics of the computers). Therefore, we created a whole database with several variables that indicate, for each SKU, its price and the number of weekly units sold, the assessment of its perceived quality by consumers depending on (i) its brand and (ii) its technical characteristics, and whether it was stimulated by in-store displays or featured on the advertising flyer each week.

#### Scanner Data

We collected weekly sales data about computers offered for sale for a period of 8 weeks, from mid-February to mid-April 2011, in one representative store in Spain of Europe's largest category killer computer retailer (Retail-Index, 2015). We selected this short period of time in order to (i) avoid holiday periods (such as Christmas or Easter), which could cause any unusual peak or off-peak in this product's sales and (ii) prevent changes in certain issues relating to the analyzed commercial stimuli that could influence their effectiveness (such us the number of promoted SKUs or the placement where they are displayed/featured). The weekly data featured units sold and sales prices of 109 SKUs from all 12 brands offered by the retailer. The retailer provided us the information directly from the scanner data warehouse. This information included SKU stocks that were used to control whether any SKU was out-of-stock. If so, these weekly observations were deleted and not considered because its sales were zero since consumers could not purchase it. By contrast, we maintained observations without sales, if the SKU was offered at the point of sale, for our analysis. Furthermore, some SKUs that were offered for sale in the first weeks did not remain on sale during the whole 8-week study period because they were replaced by other more modern ones and, thus, other SKUs started to be offered by the store during the time period analyzed (e.g., third week, fourth week, etc.). These SKUs are considered in the database for the weeks in which there are offered for sale. Finally, we obtained 599 observations from 109 different SKUs. This information about prices and units sold is completed with other information sources, such as survey and observation, to detail the perceived quality and commercial stimuli.

#### Survey Data

We conducted a consumer survey within the store over the same 8 weeks we collected the information about SKU prices and sales and the use of commercial stimuli. In this survey, we used different items to obtain (i) the assessment of perceived quality attributed to the brand for each of the 12 brands and (ii) the relative importance that different technical characteristics of computers have for the surveyed consumers. In fact, we used a five-point Likert scale for the assessment of a brand's perceived quality and we asked respondents to portion out 100 points among the technical characteristics according to their importance for the purchase. The items used in the questionnaire can be found in the Appendix Section (Appendix 1). First, we collected a pretest with 21 consumers and employees of the store in which we asked them to review the survey for any errors. After fixing a few minor mistakes, we randomly asked 402 customers who were observing and/or comparing computers inside the store, in fact, in the computer aisle or area. We decided to survey consumers who were interested on buying a computer, i.e., customers with a purchase intention, at the moment of making a decision because we believed their assessment could explain more reliably the influence of perceived quality and commercial stimuli on the computer sales. Finally, we obtained 376 valid surveys from consumers, 53 of whom considered themselves experts on computers.

#### Observational Data

We used observational data for two objectives. On one hand, we visited the selected store once a week to capture when a display was used to stimulate purchase of a particular SKU and we reviewed the weekly advertising flyer each week to capture what computer products were featured in it. We also controlled for different issues that could influence the effect of these commercial stimuli, such as the number of items displayed or featured or the place within the store or in the advertising flyer in which the SKU was displayed or featured (Gijsbrechts et al., 2003; Bezawada et al., 2009; Chandon et al., 2009; Luceri et al., 2014). Thus, we preferred to select a shorter period of time in which any of these issues about commercial stimuli changed. For example, the number of displayed computers is the same for each week (except 1 week in which the number is four instead of five) and they were located at the same point within the store so that this issue would not have an influence on their effectiveness. Regarding weekly

#### advertising flyers, they were launched in the same geographic area, with the same number of pages (eight), the same temporal frequency or duration (weekly), the same number of featured computer products (except 2 weeks in which the number is five instead of four) in the same pages and with average discounts without significant differences.

On the other hand, we checked the information about each computer product on the manufacturers' websites in order to obtain the assessment of quality attributed to intrinsic characteristics by its manufacturer. They usually use a five-star system in which they assess the quality of their computers' technical characteristics, on many occasions by following the assessment of the actual manufacturer of the components (e.g., Intel for processor). In the cases in which the manufacturer did not use this five-star system, we used the assessment from the manufacturer of components for the same technical attributes with the same characteristics (speed or capability) and we asked the retailer manager to give us his/her opinion or approval. The assessment used for each technical attribute can be found in the Appendix Section (Appendix 2). **Table 2** provides some descriptive information of the three data sources that we collected and used in this study.

### Operationalization of Variables Measures of Perceived Quality

The customers' perceived product quality comprises both types of attributes, extrinsic and intrinsic, and therefore it was really necessary to measure these two dimensions independently: the perceived quality attributed to the brand and the perceived quality attributed to the intrinsic product characteristics. For these measures, we collected the consumer survey of potential buyers and paid attention to the technical characteristics of the computers.

For the "perceived quality attributed to the brand" (PQB) variable, we turned to assessments of the brands that consumers offered in the survey. Specifically, we directly asked about the level of quality they would attribute to each of the 12 brands in the panel data, on a five-point Likert scale following already validated quality scales, such as those from Keller and Aaker (1992), Grewal et al. (1998), and Erdem et al. (2006). This procedure provided us with a separate score for each of the 12 brands that were offered for sale in the study period. Therefore, we obtained a new enclosed variable—the values are between one and five—which


Study Data

Time Period: mid-February—mid-April 2011 (8 weeks) Place: Salamanca (Spain)


collects the average score of the perceived quality attributed to the brand, according our surveyed consumers. As we searched and asked for the brand quality assessment, all products of a particular brand have the same value for this variable.

For the "perceived quality attributed to the intrinsic product characteristics" (PQIC) variable, we combined information related to the technical attributes and subjective information from the consumer survey. First, for each SKU, we sought the manufacturers' assessment of the four most important intrinsic technical characteristics for purchasing a computer (Mitra and Golder, 2006): processor speed, RAM capacity, hard drive capacity, and graphics card capability. Thus, we obtained enclosed values—between one and five—which collect the quality level of each these four technical characteristics depending on their objective features, such as capacity measured in MegaBites/GigaBites or speed measured in GHz. However, according to our surveyed consumers, these four intrinsic attributes do not have the same importance when they purchase a computer. Therefore, secondly, we weighted the scores obtained for quality level of these technical characteristics depending on their intrinsic features, by considering the average importance that our surveyed consumers granted these technical attributes. Thus, we obtained another enclosed variable—with values also between one and five—which collect the average score of the perceived quality attributed to the intrinsic characteristics for each of the 109 analyzed SKUs.

#### Measures of Commercial Stimuli

We collected the use of both analyzed commercial stimuli as two dummy variables, following previous studies (e.g., Narasimhan et al., 1996; Van Heerde et al., 2004; Inman et al., 2009). They take a value of 1 if the reference appears in the current week in the analyzed stimulus, and 0, otherwise.

For the displays variable, we obtained the observational information about their use by visiting the store during each of the weeks analyzed. During these visits, we explored the whole store and checked whether each SKU was located in a different placement than usual -such as an end-of-aisle or an island- and if it was stimulated by any special presentation or display.

Regarding the advertising flyer variable, we checked all the weekly advertising flyers that the retailer sent during the study period in order to create this dummy variable. This retailer usually launches a different advertising flyer every week by highlighting different SKUs from several product categories that it offers on sale. This advertising flyer is sent to potential consumers' home in a particular commercial area of influence of the store. **Table 3** provides descriptions of the main variables used for the study.

#### Model and Estimation

We proposed a sequential analysis, through multiple linear regressions, in which we considered the effect of each commercial stimulus and its interaction with both types of perceived quality separately in order to explain the weekly product sales. We then presented a general model that jointly compares the direct relationship of the perceived quality attributed to the brand and the perceived quality attributed to the intrinsic product


i, is the SKU for which we collect information; t, is the week in which the information is collected; k, is the product brand

characteristics with sales, as well as their moderating roles in the link between displays or advertising flyers and sales.

Then, we checked whether the necessary assumptions of normality, linearity and homoscedasticity were met. In addition, we found that there was no multicollinearity between the variables based on an analysis of tolerance and VIF (Hair et al., 1998).

Therefore, we defined different models for the sequential analysis. Model 1 is used to contrast the effectiveness of displays and advertising flyers for this product category, i.e., to contrast their direct effects on sales. Furthermore, the product price is used as control variable in order to analyze whether the price level is important in this product category as occurs in FMCG according to most previous studies. Note that we introduce 11 instead of 12 dummy variables for controlling the effect of each brand because a brand must be taken as a reference. Thus:

$$S\_{it} = \alpha + \beta\_1 R R\_{it} + \sum\_{k=1}^{11} \beta\_{2k} R R\_{ki} + \beta\_3 D I\_{it} + \beta\_4 A F\_{it} + \varepsilon\_i \tag{1}$$

Where:

α: is the constant to be estimated βj : are estimation parameters

ε: is the error

Moreover, Model 2 and Model 3 are used to analyze the moderating role of perceived quality on the effectiveness of displays and advertising flyers, respectively. Thus, Model 2 includes interactions of the variable that represents the use of Garrido-Morgado et al. Influence of Customer Quality Perception

displays in the store with the perceived quality attributed to the brand (DIxPQBit) and with the perceived quality attributed to the intrinsic product characteristics (DIxPQICit). Model 3 features interactions between the variable that captures the use of advertising flyers and both types of perceived quality (AFxPQBit) and (AFxPQICit). In both models, the direct effect of the perceived quality attributed to the brand gets omitted, due to its redundancy with the effects of brand dummy variables. Thus, the proposed models are expressed as follows:

$$\begin{split} \mathbf{S}\_{it} &= \alpha + \beta\_1 \mathbf{P} \mathbf{R}\_{it} + \sum\_{k=1}^{11} \beta\_{2k} \mathbf{B} \mathbf{R}\_{it} + \beta\_3 \mathbf{D} \mathbf{I}\_{it} + \beta\_4 \mathbf{P} \mathbf{Q} \mathbf{I} \mathbf{C}\_{it} \\ &+ \beta\_5 \mathbf{D} \mathbf{I} \mathbf{x} \mathbf{P} \mathbf{Q} \mathbf{B}\_{it} + \beta\_6 \mathbf{D} \mathbf{I} \mathbf{x} \mathbf{P} \mathbf{Q} \mathbf{I} \mathbf{C}\_{it} + \varepsilon\_i, \end{split} \tag{2}$$

and

$$\begin{split} \mathbf{S}\_{it} &= \alpha + \beta\_1 \mathbf{P} \mathbf{R}\_{it} + \sum\_{k=1}^{11} \beta\_{2k} \mathbf{B} \mathbf{R}\_{ki} + \beta\_3 \mathbf{A} \mathbf{F}\_{it} + \beta\_4 \mathbf{P} \mathbf{Q} \mathbf{I} \mathbf{C}\_{it} \\ &+ \beta\_5 \mathbf{A} \mathbf{F} \mathbf{x} \mathbf{P} \mathbf{Q} \mathbf{B}\_{it} + \beta\_6 \mathbf{A} \mathbf{F} \mathbf{x} \mathbf{P} \mathbf{Q} \mathbf{I} \mathbf{C}\_{it} + \varepsilon\_i \end{split} \tag{3}$$

Finally, Model 4 is the general model that collects the direct effect of both commercial stimuli as well as their interactions with the perceived quality attributed to the brand and the perceived quality attributed to the intrinsic product characteristics on sales in order to obtain their moderating roles on the effectiveness of displays and advertising flyers to increase computer sales. Therefore, the general model is expressed as follows:

$$\begin{split} \mathbf{S}\_{it} &= \alpha + \beta\_1 P \mathbf{R}\_{it} + \sum\_{k=1}^{11} \beta\_{2k} \mathbf{B} \mathbf{R}\_{ki} + \beta\_3 \mathbf{D} I\_{it} + \beta\_4 A \mathbf{F}\_{it} \\ &+ \beta\_5 P \mathbf{Q} \mathbf{I} \mathbf{C}\_{it} + \beta\_6 \mathbf{D} \mathbf{I} \mathbf{x} P \mathbf{Q} \mathbf{B}\_{it} + \beta\_7 \mathbf{D} \mathbf{I} \mathbf{x} P \mathbf{Q} \mathbf{I} \mathbf{C}\_{it} \\ &+ \beta\_8 A \mathbf{F} \mathbf{x} P \mathbf{Q} \mathbf{B}\_{it} + \beta\_9 A \mathbf{F} \mathbf{x} P \mathbf{Q} \mathbf{I} \mathbf{C}\_{it} + \varepsilon\_i \end{split} \tag{4}$$

#### RESULTS

Results of the estimations are reported in **Table 4**. The results of Model 1 indicate that the use of both commercial stimuli, displays and advertising flyers, for promoting computer products have significant and positive direct effects on their sales (p < 0.01), in line with previous studies about frequently purchased products (Gijsbrechts et al., 2003; Van Heerde et al., 2004; Bezawada et al., 2009; Inman et al., 2009). These results confirm hypotheses H1 and H2. Moreover, we tested these results and the effects of both commercial stimuli are significantly different (p > 0.01), which highlights the need to analyze them separately. Furthermore, Model 1 already indicated that computer products have differences with respect to FMCG because it was observed that the effect of the price variable is not significant. One reason for this is that consumers may focus on the perceived quality attributed to the brand or the quality of the technical attributes in order to reduce the perceived risk linked to this complex product (Neelamegham and Chintagunta, 2004; Sriram et al., 2006). Accordingly, the perceived quality would be more important than prices for these infrequently purchased products.

Model 2 indicates a significant positive moderating effect for both the perceived quality attributed to the brand (p < 0.01) and the perceived quality attributed to the intrinsic product



\*p < 0.1; \*\*p < 0.05; \*\*\*p < 0.01.

<sup>a</sup>Acer is taken as the reference brand.

<sup>b</sup>The effect of Subjective Quality is redundant, because we introduce a constant for each brand.

characteristics (p < 0.10) on the effectiveness of displays to promote computer products. These results indicate that the higher the perceived quality of a product (both types of perceived quality), the greater the displays' effectiveness. It seems these findings are in line with the results of previous studies analyzing FMCG, such as Lemon and Nowlis (2002); however, they use price tiers as proxy of quality, and thus these findings are different and original. Furthermore, we distinguished between perceived quality attributed to the brand and perceived quality attributed to intrinsic attributes and the effects of both types of perceived quality are significantly different (p < 0.01). Thus, we may surmise that the effectiveness of displays on computer sales will be different depending on the two types of perceived quality.

Similarly, Model 3 indicates the moderating effect of both types of perceived quality on advertising flyers' effectiveness on computer sales. In fact, the interactions between advertising flyers and the perceived quality attributed to the brand (p < 0.10) and perceived quality attributed to the intrinsic product

TABLE 5 | Summary of results.


characteristics (p < 0.05) present a significant positive effect. The effect of these interactions are also significantly different (p < 0.01), indicating that each type of perceived quality has a different impact on the effectiveness of each analyzed commercial stimulus. As in the case of the displays, we may conjecture that the effectiveness of advertising flyers on computers sales will be different depending on the two types of perceived quality.

Finally, the results of Model 4 confirm that perceived quality attributed to the brand presents a greater moderating impact on displays (3.998, p < 0.01) than on advertising flyers (2.439, p < 0.10). These effects are significantly different (p < 0.01) and indicate that the perceived quality attributed to the brand enhances the effectiveness of displays more than it does the effectiveness of advertising flyers for sales; therefore, these results are in line with hypothesis H3. In addition, Model 4 also indicates that perceived quality attributed to the intrinsic product characteristics exerts a greater moderating impact on advertising flyers (4.303, p < 0.01) than on displays (2.149, p < 0.10). These effects are also significantly different (p < 0.01) and indicate that the perceived quality attributed to the intrinsic product characteristics enhances the effectiveness of advertising flyers more than it does the effectiveness of displays for sales; therefore these results are also in line with hypothesis H4. We summarize these results in **Table 5**.

### DISCUSSION

Product quality is a key determinant in the purchase of computer products, and in stimulating customer loyalty. Perceived quality has two dimensions, extrinsic quality—linked to the brand and intrinsic quality—related to internal product characteristics. Whereas extrinsic attributes (brand name) are more related to affective loyalty (customers build affect toward the brand on the basis of cumulatively satisfying usage occasions), intrinsic attributes have a more objective nature and are more related to cognitive loyalty.

The main aim of this research was to analyze the importance of both perceived quality attributed to intrinsic characteristics and perceived quality attributed to the brand in explaining the effects of two commercial stimuli, displays inside the store where the customer has less time to process information and, advertising flyers sent to potential customers where the customer has more time to process information. We analyze an infrequently purchased product, computers, which feature substantial technological components and greater perceived risk, because of their complexity and dynamic evolution (Neelamegham and Chintagunta, 2004; Sriram et al., 2006).

According to our results, both stimuli, displays and advertising flyers, help increase sales in this product category. The positive influence on sales is moderated by the two dimensions of the products' perceived quality: one attributed to intrinsic characteristics (the product's own technical characteristics) and one attributed to the brand (assessment of brand quality). The more perceived quality a product has, the greater the impact of the displays and advertising flyers. However, we also emphasize that the perceived quality attributed to the brand improves to a greater extent the effect of displays, which work inside the store as a brand heuristic to help consumers with information processing. By contrast, we find that the perceived quality attributed to the intrinsic product characteristics improves to a greater extent the effect of advertising flyers, which are usually sent to potential consumers' home and favor systematic information processing.

These findings may be considered original and very useful for manufacturers and retailers because previous studies did not analyze the effectiveness of different commercial stimuli (one in-store and one out-of-store stimuli) on infrequently purchased products (such as computer products which feature substantial technological components and greater perceived risk, because of their complexity and dynamic evolution), and because we explain these results from the point of view of consumer behavior, focusing on the two dimensions of quality (perceived quality attributed to the brand and perceived quality attributed to intrinsic attributes). Therefore, our findings entail very important managerial implications for retailers as well as for manufacturers because they show what type of commercial stimuli is more appropriate for technological and infrequently purchased products depending on their extrinsic and intrinsic attributes.

Furthermore, these findings also have theoretical relevance in marketing, retail and consumer behavior research because they prove that commercial stimuli work differently depending on whether they take place in-store (time-constrained condition) or out-of-store (low time-constrained condition), since they trigger different types of purchasing processes (cognitive loyaltyintrinsic perceived quality or affective loyalty-extrinsic perceived quality). Because of this, unlike what was previously believed, commercial stimuli not only work for more hedonic or impulsive product purchases–which are usually acquired through an unplanned purchase process–but they may also be used for products with a higher perceived risk and product purchase involvement. To optimize sales performance, it is necessary to understand how consumers' perceived quality is formed by distinguishing between these two dimensions.

### Managerial Guidelines for Manufacturers and Retailers

According to these results, we recommend that computer manufacturers allocate some resources to convincing retailers to encourage sales of their products through displays and advertising flyers; the results affirm their effectiveness for infrequently purchased products such as computers. In addition, they could devote further efforts to increase the perceived quality attributed to their brand, which may be more profitable in terms of enhancing their image than improving the intrinsic technical product characteristics. In this product category, consumers seek to reduce the high perceived risk associated with purchasing such an infrequently purchased, relatively expensive, complex product by focusing on the brand. Furthermore, many consumers also seek advice from other expert consumers or sellers, who have a great influence as prescribers. Thus, manufacturers should combine a pull strategy to attract potential buyers with a push strategy that encourages the retailer's sellers to recommend their brand instead of other brands.

For retailers, we suggest they pay close attention to their uses of displays and advertising flyers because both commercial stimuli are particularly effective in the computer product category. According to our results, retailers could use displays inside the store for products with higher perceived quality attributed to recognized brands in order to have consumers under time-constrained conditions use heuristics to simplify their decisions and make an unplanned purchase. In contrast, they could prominently feature products with higher intrinsic technical quality on their advertising flyers because consumers are likely to compare the product's attributes carefully and make more reasoned purchases under low time-constrained conditions.

Furthermore, buyers are not extremely price sensitive, perhaps because they assume price cuts when the product is displayed in a special location or featured on an advertising flyer (Gázquez-Abad et al., 2014). An additional reason could be that they focus on the perceived quality attributed to the brand or to the technical attributes for reducing the perceived risk linked to this complex product, rather than gain economic deals. In this line, the perceived quality would be more important than prices for these infrequently purchased products and retailers may optimize their profits by using commercial stimuli without needing to offer great price discounts.

#### Limitations and Further Research

The major limitation of this study is that our data come from only one representative store in Spain belonging to Europe's largest category killer computer retailer. This was due to the need (i) to visit the store weekly in order to collect observational data about the use of displays, as well as (ii) to collect a consumer survey within the store in which the sales data were obtained about their product quality assessment. Related to this limitation, we only

### REFERENCES


were able to use data on the SKUs and brands sold by only one store for a short period of time, although it does pertain to the most important retailer in Europe.

Therefore, given this limitation, further research could expand this study by collecting data from different stores and, even different retailer chains. It would be interesting to compare our results across (i) other category killers, (ii) other retail formats that also sell computers such as hypermarkets, or (iii) other geographic areas. Through these expansions, further research could examine if our findings are moderated by the retail chains' positioning and their characteristics (such as their assortment, their specialization level, their sales staff, etc.). Furthermore, it is possible that results may vary depending on behavioral loyalty, higher/lower price sensitivity, or the purchase involvement and perceived risk in this product category linked to the greater/lesser product knowledge level of consumers in other regions or countries.

Another important limitation is the use of a validated non-specific questionnaire for the assessment of subjective product quality and the importance of the four technical product attributes most analyzed for computer products. Thus, another interesting research extension might devise a specific questionnaire, adapted to technological products, to refine the measure of subjective quality.

In sum, to complement this study, further research might increase the number of SKUs and brands as well as the number of analyzed product attributes, for a longer study period, and examine our findings by considering different variables that could be moderating them, such as retailer positioning, retailer format or differences in behavioral variables depending on areas or countries, in order to gain further insights and improve the managerial guidelines that can be derived from the analysis.

### AUTHOR CONTRIBUTIONS

ÁG, ÓG, and MM have participated in the development of the conceptual framework and methodology as well as the data collection, the realization of the empirical analysis and the results.

### ACKNOWLEDGMENTS

This research was supported by Ministerio de Educación y Ciencia, Grant ECO2011-23381 and ECO2014-53060-R (Spain). The authors also acknowledge the efforts of Cátedra Fundación Areces de Distribución Comercial, which aims to stimulate Spanish research; this article is based on a previous version, which the authors prepared for the Working Paper Collection of this nonprofit organization.

signals on perceptions of automotive brand quality. J. Prod. Innov. Manage. 31, 728–743. doi: 10.1111/jpim.12120


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Garrido-Morgado, González-Benito and Martos-Partal. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## APPENDIX 1. QUESTIONNAIRE ON QUALITY ASSESSMENT OF COMPUTER PRODUCTS

Dear client:

We are conducting a survey for a study by the University of Salamanca in the field of computer products.

If you want to contribute, you need only answer some brief questions according to your opinion. This will only take a few minutes.

In order to ensure maximum confidentiality, your answers will be treated anonymously and in aggregate form. Thank you very much for your willingness to complete this questionnaire.

## QUESTIONS

1. Distribute 100 points into these characteristics according their importance when you buy a computer.


2. Assess the quality of these brands according your opinion.


## APPENDIX 2. ASSESSMENTS OF TECHNICAL ATTRIBUTES


# Consumer Behavior in Shopping Streets: The Importance of the Salesperson's Professional Personal Attention

Natalia Medrano\*, Cristina Olarte-Pascual, Jorge Pelegrín-Borondo and Yolanda Sierra-Murillo

Departamento de Economía y Empresa, Universidad de la Rioja, Logroño, Spain

#### Edited by:

Monica Gomez-Suárez, Universidad Autónoma de Madrid, Spain

#### Reviewed by:

Maria Avello, Universidad Complutense de Madrid, Spain Victor Manuel Molero, Universidad Complutense de Madrid, Spain

\*Correspondence:

Natalia Medrano natalia.medrano@unirioja.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 17 November 2015 Accepted: 25 January 2016 Published: 10 February 2016

#### Citation:

Medrano N, Olarte-Pascual C, Pelegrín-Borondo J and Sierra-Murillo Y (2016) Consumer Behavior in Shopping Streets: The Importance of the Salesperson's Professional Personal Attention. Front. Psychol. 7:125. doi: 10.3389/fpsyg.2016.00125 Since the early 2010s, the emergence of a new consumer has begun. In this context, consumer behavior represents one of the greatest interests of marketing scholars and business managers due to their need to adapt their companies' strategies to the new frontier. In order to advance understanding of this new consumer, this article focuses on analyzing consumer behavior in shopping streets. Thus, the aim of this research is to know what customers value in terms of salesperson–customer interaction quality nowadays. To achieve this, the authors conducted two studies. The results of the first study show that customers cite personal attention as the primary factor motivating their preference for small retailers in shopping streets. However, this motivation is not as relevant one for those who prefer malls. This result provides a point on which to research service quality incorporating personal attention in a second study. Using the SERVQUAL-P scale, the authors elaborate three lenses through which the quality of service from the customer's point of view can be analyzed: normative expectations, predictive expectations, and the importance of each attribute. The most striking result is that the dimensions of expectations (normative and predictive) are the same; these results demonstrate that customers are coherent in making assessments of their expectations, evaluating service quality and satisfaction with similar criteria. However, these dimensions are different from the dimensions of importance. Our main contribution lies in the finding that personal attention, when assessed using the scale of attribute importance, is split into two dimensions: (1) courteous attention and (2) personal relationship. Courteous attention is always welcome, but personal relationships are less valued and are often even rejected. The article concludes with a discussion of the implications of these findings for marketing practices and research.

Keywords: consumer behavior, small retail, mall, shopping street, expectations, personal attention, personal relationship

### INTRODUCTION

Cities in Europe are often characterized by an urban center with commercial streets in which numerous independent stores are located. These stores are often family-owned, small, and specialized, and the employees tend to have an indepth knowledge of the product and a greater focus on customer service. Commercial development, in contrast, has been characterized by the emergence of large malls located on the outskirts of cities. These malls have led to the displacement of consumers toward these urban peripheries, thus hurting the more traditional urban retail trade (Sadahiro, 2000; O'Callaghan and O'Riordan, 2003; Hernández and Jones, 2005). Small retailers located in shopping streets are losing customers every day; these stores eventually close, and over time cities slowly begin to lose their cultural and economic vibrancy. This has led European public authorities to take action to improve the management of their cities' commercial centers and the shops therein (Medway et al., 2000; Paddison, 2003). It could be said that traditional urban small trade of shopping streets has become an endangered species.

At the same time, since the early 2010s the emergence of a new consumer has begun: Consumer 3.0. The influence of sociocultural shifts on this consumer's purchase behavior is highlighted, especially factors that are technological, social, or emotional in nature (Sersland, 2015). Today's consumers want to feel more in control and they want to be seen and valued more than their money. Technology has radically changed the psychology of these new consumers and created a host of new expectations. They do not want to sift through irrelevant information, lengthy explanations, or anything not immediately important. In this context, consumer behavior, which is sometimes guided by self-related motives rather than by rational economic considerations (Cisek et al., 2014), represents one of the greatest interests of business managers due to their need to adapt their companies' strategies to the new frontier. Current research (Grewal et al., 2009) stresses that survival in today's economic climate and competitive retail environment requires more than just low prices and innovative products. Several authors (Badgett et al., 2007; Gentile et al., 2007; Grewal et al., 2009; Tynan and McKechnie, 2009; Verhoef et al., 2009; Rose et al., 2012) emphasize the importance of the shopping experience when choosing among different retailers.

The quality of service and the degree of personal attention are important factors in consumer behavior (Gremler and Gwinner, 2008). Thus, the aim of this research is to know what customers value in terms of salesperson–customer interaction quality nowadays. To achieve this, the authors conducted two studies (**Figure 1**).

Study 1 is based on 220 surveys. We find that customers perceive clear differences between malls and shopping streets. Furthermore, 54% of respondents prefer to do their shopping at malls mainly because of their wide and varied commercial offerings, and 35% prefer going to shopping streets because of their personal attention. The remaining respondents (11%) reported liking both types of retail destination and spread their purchases out between the two on the basis of price, product category, and/or convenience. It is evident from these responses that the kind of personal attention found in shopping streets is valued enough to win the loyalty of a large group of consumers.

In Study 2, we examine in greater detail how personal attention can provide a competitive advantage and explore the fundamental components of personal attention. This study is based on information obtained from a sample of 974 customers of small retailers in shopping streets. We use Mittal and Lassar's (1996) SERVQUAL-P scale, which incorporates aspects of personal attention, and apply it in three different ways. Specifically, we employ three lenses through which to view

the components of personal attention: (1) "what should be" (16 attributes of service quality for customers, four of which are specifically related to personal attention) (i.e., "normative expectations"), (2) "what customers really expect" to happen at the store (i.e., "predictive expectations"), and (3) "how important each attribute is."

In Study 2 analysis A, we demonstrate that customers differentiate between normative and predictive expectations and analyze the gap that occurs between them as well as their dimensions. The most striking result is that the dimensions of expectations (normative and predictive) are the same; however, these dimensions are different from the dimensions of importance. Our main contribution lies in the finding that personal attention, when assessed using the scale of attribute importance, is split into two dimensions: (1) courteous attention and (2) personal relationship. Courteous attention is always welcome, but personal relationships are less valued and are often even rejected (Study 2 analysis B).

The current research focuses on the Spanish city of Logroño. In 1997, it was acknowledged as the first commercial city in Spain. Currently, its commercial area includes two malls (characterized by having one or more large retailers that are a driving force of the city's economy and a city center with several shopping streets that are connected but have no large commercial spaces).

This study was approved by the University of La Rioja Research Ethics Board and according to ICC/ESOMAR International Code on Social Research. Each participant provided informed consent.

### STUDY 1: SHOPPING STREET OR MALL? WHY?

### Literature Review

#### The Customer Experience

Customer orientation is considered a competitive strategy for smaller service enterprises (Polo Peña et al., 2011). Creating a superior customer experience seems to be a central objective in today's retailing environment. A recent IBM report identifies customer experience as a key factor for companies in building loyalty to brands, channels, and services (Badgett et al., 2007). Effective retail management strategies have been linked to the creation of customer experience, which in turn leads to successful performance outcomes (Gentile et al., 2007; Grewal et al., 2009; Tynan and McKechnie, 2009; Rose et al., 2012). Yet, despite practitioners' recognition of the importance of the customer experience, the academic marketing literature on this topic has been limited. Publications on customer experience are mainly found in practitioner-oriented journals or management books.

Previous research on "customer experience" (e.g., Verhoef et al., 2009) recognizes the importance of past customer experiences, store environments, service interfaces, and store brands on future experiences. Some literature on retail experience has focused on store atmospherics and the impact of scents, music, tactile input, and color on customers' affective responses to a retailer (Naylor et al., 2008). Puccinelli et al. (2009) examine store atmospherics and the social environment. Retail environmental factors, such as social features, design, and ambience, can result in enhanced pleasure and arousal (Mehrabian and Russell, 1974; Baker et al., 1992). Other research in this area examines the presence and age of other consumers in the retail or service setting (e.g., Thakor et al., 2008) and the effect of crowds, music, and lighting (Baker et al., 2002). Novak et al. (2000) investigate the impact of website design on the customer's experience. Other research topics within the customer experience domain are personal relationships and service quality expectations.

#### Personal Attention

Mittal and Lassar (1996) define personalization as the social content of interaction between service employees and their customers. In this sense, "personalization" pertains to the service employee's manner of relating to the customer on a human level: cold and impersonal at one end of the spectrum and warm and personal at the other. As such, it includes aspects such as employees' politeness and courtesy, employees' attempts to get to know the customer as a person and to engage in friendly conversation, and the exhibition of personal warmth in employee behavior.

The popularity of relationship marketing stems, in part, from the assumption that building customer relationships yields positive returns in the form of customer satisfaction, loyalty, word of mouth, and purchases (Reynolds and Beatty, 1999). Moreover, interactions between retail employees and customers can have a significant impact on customers' perceptions of the organization (Tsiros and Parasuraman, 2006; Gremler and Gwinner, 2008; Lichtenstein et al., 2010; Otnes et al., 2012; Litz and Pollack, 2015). The rapport between employees and customers represents a particularly salient issue in retail businesses characterized by significant interpersonal interactions (Haas and Kenning, 2014).

Gist (1968) emphasizes that the opportunity to develop personal relationships, and therefore to give personal attention, was one of the factors leading to the emergence of the specialty store in the early nineteenth century<sup>1</sup> . As the specialty store has evolved, this characteristic of personal association between buyer and seller has led to the popularity of the specialty store retailer. Many of today's specialty retailers have become successful by combining this element of personalized service with a merchandise assortment geared toward a particular market segment. It has been argued that the success of Starbucks is due to its ability to create a distinctive customer experience (Michelli, 2007). Gist (1968) concludes in his study that employees from these specialty stores need to be responsive, courteous, and knowledgeable and offer prompt, individualized service as a primary distinguishing characteristic of the shopping experience. In line with this perspective, Puccinelli et al. (2009, p. 24) argue that "the interpersonal nature of the interaction between the customer and employee [. . . ] may be key to customer satisfaction in the retail environment." Ulaga and Eggert (2006) identify service support and personal interaction as core differentiators

<sup>1</sup> Specialty retailing combines the selling of goods and services to the consumer, and consumers expect knowledgeable, helpful staff to assist in the sales procedure (Gist, 1968).

in business relationships. An important insight from Johnson and Selnes (2004) is that firms which position themselves toward offerings with low economies of scale, such as personal services, must build closer relationships to create value.

However, the importance of personal attention is not universal and varies across different service industries and cultures and it should be taken into account in order to differentiate between shopping streets and malls. Several researchers have compared the factors that draw consumers to shopping streets and malls. Reimers and Clulow (2004) believe that malls provide greater spatial convenience than shopping streets. Teller and Reutterer (2008) establish that the commercial mix, value for money, and entertainment element influence the appeal of a shopping street and a mall. Finding one's way around more easily is mentioned as a positive aspect of malls. Store atmospherics (e.g., scent, temperature, air) are a factor in both shopping streets and malls, though they are a more intense factor in malls. The ranking of the retail mix attribute depends on consumers' expectations (Léo and Philippe, 2002). Finally, Reimers (2013) states that people generally perceive malls as more accessible when using their car to go shopping.

Due to the importance of the salesperson–customer interaction<sup>2</sup> in defining the consumer's experience and ultimate satisfaction (Goodwin, 1996; Menon and Dubé, 2000; Stock and Hoyer, 2005; Schau et al., 2007; Gremler and Gwinner, 2008), and in light of our desire to explore the importance of personal attention in shopping streets vs. malls, we propose the following:

H1: Personal attention is the main motivating factor by customers who choose to go to shopping streets but not for those who choose to go to shopping malls.

### Data

**Figure 2** shows our study area. The types of establishments located in the shopping streets of Logroño are mainly specialized small businesses in which the owner and their family serve the customer directly. On the outskirts of the city, two malls characterized by larger shops owned by large companies, a supermarket, and several category killers serve as the primary generators of customer traffic (**Table 1**).

We obtained information from a quota sampling of 220 people. We considered the representativeness of the sample, establishing age and gender quotas. The method used was personal in-home interviews.

### Method

To learn about the importance of personal attention in the shopping experience, we conducted a survey using openresponse questions. We asked respondents whether they preferred going shopping at a mall or a shopping street and why. The aim was to find out, indirectly and without biasing the respondents, whether personal attention is the main motivation for choosing the shopping location (H1). In order to contrast this hypothesis we used a binary logistic regression, applying wall method.

<sup>2</sup>The salesperson–customer interaction is characterized by voluntary, dyadic interpersonal exchanges between the buyer and the seller and is traditionally understood to be the cornerstone of the retail/service customer experience (Meuter et al., 2000).


TABLE 1 | Characteristics of the study area and sample.

\*Source: Data are based on a 2011 survey of retailers by Cámara de Comercio e Industria de La Rioja.

### Results

Of those surveyed, 54.09% reported preferring malls, 34.55% opted for shopping streets, and the rest (11.36%) reported having no preference between the two. Malls were generally preferred by young people (78.08%) and adults (44.18%) and by individuals of both sexes. However, their appeal was greater for men (55.91%) than for women (52.76%). Older people (49.18%) preferred shopping streets, and their appeal was greater for women (38.58%) than for men (29.03%).

When asked why, the respondents gave 352 reasons, which we then grouped into six primary motives. With regard to shopping streets, the motives stated by those surveyed, in order of importance, are as follows:


With regard to malls, the motives stated by those surveyed, in order of importance, are as follows:

1. Commercial offer (49.15%): this is a motive for choosing a mall when shopping for consumer goods. Statements such as



"big shopping," "monthly shopping," "I can find everything," or "there are more things in a mall than in the neighborhood." In addition, there is another key concept: wide offer.


Finally, we encountered buyers who did not have a clear preference, that is, those who shop both in shopping streets and at malls. These buyers reported different priorities:


We analyze the explanatory relationship between the dichotomous dependent variable (preference between shopping streets or malls) and the independent variables (the eight most cited motives). **Table 2** shows the logistic regression results. Goodness of fit were adequate: Nagelkerke's Pseudo R square = 45.6% and percentage correctly classified = 74.9%. Regarding the multicollinearity analysis the major variance inflation factor (VIF) of logit is 1.23 (corresponds to commercial offer).

Results show that people more likely to go to the shopping streets are those who seek personal attention. On the other hand, people more likely to go to the malls are those who look for a wide commercial offer with good price.

#### Discussion

The results of this survey indicate that the main strength of retailers located in shopping streets is personal attention. Thus, we can accept H1. No respondents who preferred malls mentioned this motive when explaining their choice.

The strengths of malls are their wide offer and prices. Spatial convenience appears to be an important factor in both contexts, though with a different meaning. In shopping streets, it means spatial proximity. In malls, it means finding everything the customer needs in one place. In conclusion, the motivations behind choosing one retailer environment over the other are different. In addition, the study enables us to infer certain connotations about the importance of personal attention.

### STUDY 2. SERVICE QUALITY INCORPORATING PERSONAL ATTENTION IN SMALL RETAILERS IN SHOPPING STREETS

The potential for personal attention to serve as a competitive advantage for small retailers located in shopping streets has led us to analyze service quality in greater depth. The objective of this second study is to learn what the most important components of personal attention are for consumers who choose small retailers.

### Literature Review

#### Expectations in Service Quality

Expectations play a significant role in determining customer perceptions and satisfaction. Accordingly, retailers seek to manage customers' service expectations (Mitra and Fay, 2010). The literature on expectations is broad, and many ways of understanding and studying expectations have been found. After a detailed review, we have classified expectations into ten types, which we have then grouped into four main approaches: (1) comparison, (2) ideal amount, (3) levels, and (4) point of assessment (see **Table 3**).

Consumers use expectations in service quality to compare competing offers (Oliver, 1977, 1980; Cadotte et al., 1987; Oliver and Burke, 1999; Andreassen, 2000). But what type of expectations do customers use when it comes to assessing distributors? When trying to answer this question, we find that there is no clear consensus on the topic (Zeithaml et al., 1993; Walker and Baker, 2000).

On the one hand, some research has argued that the service quality a customer receives can be measured using normative expectations and that predictive expectations are more appropriate for measuring customer satisfaction (Boulding et al., 1993; Zeithaml et al., 1993; Dean, 2004; Higgs et al., 2005). On the other hand, several researchers believe that using other types of expectations is appropriate; however, normative and/or predictive expectations are included their work as well (Golder et al., 2012).

Because our framework is related to the identification of potential competitive advantages for shopping streets, we chose the comparison approach for the analysis and classification of expectations (**Table 3**). This approach distinguishes and conceptualizes three types of service quality expectations to compare competitive brands: normative service quality expectations, equitable or deserved quality expectations, and predictive quality expectations.

Normative expectations represent an excellent level of service quality that a person believes a supplier should realistically and feasibly offer for a specific service—that is, what the customer thinks it should be. These expectations are usually related to a particular category of service.

Equitable expectations are defined as the equity level of service the customer feels the seller must supply, taking into account the costs incurred. Under this perspective, equitable expectations are critically determined by a personal assessment of the potential rewards vs. costs.

Predictive expectations represent the calculation a person performs to determine what he or she really expects a supplier to provide in a particular situation.

Of these three types of expectation (normative, equitable, and predictive), we decided to measure only normative and predictive expectations. We leave equitable or deserved expectations to future work, because we would need to take into account the costs the customer incurs, and our empirical study does not analyze actual purchases.

We characterize normative expectations as being more stable over time in the customer's mind (Johnson and Mathews, 1997; Clow et al., 1998). Moreover, predictive expectations are linked to the existence of the next service encounter, while normative expectations do not require any temporal proximity of the service.

#### The Importance of the Attributes of Service Quality (Incorporating Personal Attention)

Another aspect to consider in the assessment of service quality is the importance of each attribute used to measure it. Teas (1993) proposes an assessment model for the quality received, differentiating between the importance of the attributes of service quality and the ideal expectations in service quality. Parasuraman et al. (1991) analyze the importance of the dimensions of service quality.

#### Relationships between Service Quality Expectations and the Importance of the Attributes of Service Quality (Incorporating Personal Attention)

We believe that it is important to know the expectations customers have regarding the various components of personal attention. However, customers seem to obtain simultaneous information about how important each component is. In this regard, we analyze whether normative expectations, predictive expectations, and importance are different concepts in customers' minds. We propose the following hypotheses:

H2: Customers are able to differentiate between normative and predictive expectations of service quality, taking into account personalization.

H3: Customers are able to differentiate between normative expectations and the importance of the attributes of service quality, taking into account personalization.

H4: Customers are able to differentiate between predictive expectations and the importance of the attributes of service quality, taking into account personalization.

### Data

#### Data Collection

A face-to-face survey, applying the SERVQUAL-P scale developed by Mittal and Lassar (1996), which is an adaptation of the SERVQUAL scale developed by Parasuraman et al. (1991), was conducted on a sample of individual customers in a city in northern Spain. Respondents participated voluntarily without any compensation.

#### TABLE 3 | Classification of expectations.

1. The Comparison Approach: This refers to the use of service quality expectations to compare competitive brands. We can differentiate between:

(1a) Normative expectations represent the level of service that a person believes a supplier should provide to offer excellent quality when it comes to a specific service, carrying out a realistic and feasible assessment. For example, if a person chooses a specialty luxury clothing shop to buy a dress for a special occasion, the question we ask, What level of information about the dress should the employees of a luxury clothing boutique provide?

(1b). Equitable or deserved expectations represent the level of service that a customer believes he or she should receive, taking into account the expenses borne—that is, what the customer considers fair. For example, the quality of service a buyer considers fair when purchasing an €800 dress in a boutique dress shop in Logroño's shopping street district. In this example, if a person spends €800, the question for an attribute might be, What would be a fair level of information about the dress to receive?

(1c) Predictive expectations represent the objective calculation that a person carries out regarding what he or she actually expects to receive from a supplier in a specific situation. For example, what level of service a person actually expects employees of a luxury boutique in Logroño to provide when she goes to buy a dress for a special occasion. The question in this sense would be, What level of information about the dress does this person actually hope to receive in this store?

Source: (Miller, 1977; Oliver, 1980; Cadotte et al., 1987; Oliver and Winer, 1987; Bitner, 1990; Boulding et al., 1993; Zeithaml et al., 1993; Clow et al., 1997; Johnson and Mathews, 1997; Clow et al., 1998; Hamer et al., 1999; Kopalle and Lehmann, 2001; Kalamas et al., 2002; Anderson and Salisbury, 2003; Dean, 2004; Higgs et al., 2005; Evans et al., 2008; Mitra and Fay, 2010; Benedicktus, 2011; Lin and Wu, 2011; Yip et al., 2011; Golder et al., 2012; Yang et al., 2013; Hung, 2015; Hung et al., 2015).

2. The Ideal Amount Approach: This refers to what the customer considers an ideal level of service. We can differentiate between:

(2a) Vector expectations refer to attributes for which the ideal amount that the customer requires is infinite; therefore, the customer never reaches his or her maximum utility (e.g., an optician's level of knowledge).

(2b) Ideal point expectations refer to attributes for which the ideal amount that the customer requires is finite (e.g., the temperature of a store). A further differentiation has been made within this ideal point expectation regarding expectations for which the ideal finite amount is feasible (i.e., feasible ideal point expectations) and those that cannot be reached by any supplier (i.e., classic ideal point expectations).

Source: (Miller, 1977; Woodruff et al., 1983; Parasuraman et al., 1985; Teas, 1993; Zeithaml et al., 1993; Parasuraman et al., 1994b; Clow et al., 1997; Higgs et al., 2005; Tsai et al., 2011; Golder et al., 2012).

3. The Levels Approach: This refers to different levels of expectations that set the limits for the interval of tolerance allowed in the assessment of the service. We can differentiate between:

(3a) Desired expectations represent the highest level of performance that a consumer considers can be reached by the suppliers of a product category (e.g., I wish I could pay by mobile phone in the store).

(3b) Adequate expectations reflect the minimum service quality that a consumer believes should be expected from the suppliers of a product category (e.g., I should at least be able to pay with a credit card at the store).

Source: (Miller, 1977; Parasuraman et al., 1991, 1993, 1994a; Johnson and Mathews, 1997; Hamer et al., 1999; Bebko, 2000; Walker and Baker, 2000; Higgs et al., 2005; Nadiri and Hussain, 2005; Yap and Sweeney, 2007; Nadiri, 2011).

4. The Point-of-Assessment Approach: This refers to the point at which the customer creates his or her expectations of service quality. We can differentiate between:


Source: (Oliver, 1980; Bitner, 1990; Bitner et al., 1990; Zeithaml et al., 1993; Mittal and Lassar, 1996; Johnson and Mathews, 1997; Clow et al., 1998; Hamer et al., 1999; Oliver and Burke, 1999; Dawar and Pillutla, 2000; Choi and Mattila, 2008).

For consumers to be able to value predictive expectations, they must have a specific small store in mind. Therefore, we chose a group of stores that represent the most important sectors of the area of study (fashion and accessories, footwear, furniture and decoration, computer stores, gift stores, opticians, and travel agencies).

To measure normative expectations, predictive expectations, and the importance of the attributes of service quality, we applied a sequential process. For each attribute (observable variable), respondents rated on a scale from 0 to 10 the level of service that a store should offer for it to provide excellent service (normative expectations), the service actually expected (predictive expectations), and the importance of each item on the scale (importance). For example, for attribute 9 ("everyone at this retailer is polite and courteous") we obtained the following: what it should be, what respondents actually expect of the service quality, and how important each item is for the customer.

#### Sample Characteristics

Overall, 1088 questionnaires were collected in December 2013, but as some had to be eliminated because they were not complete, 974 usable questionnaires were obtained.

**Table 4** shows that the sample of customers reproduces the structure of the population by sex and age categories.

### Study 2 Analysis A

We want to know how important 16 service quality attributes (incorporating personalized attention) are to the customers and the expectations they have regarding those attributes. Previously, we analyzed whether customers perceive differences between normative expectations, predictive expectations, and importance.

#### Method

To determine whether consumers perceive any differences between normative expectations, predictive expectations, and importance in the attributes of service quality, we performed 2 × 2 contrasts. Then, we obtained the underlying dimensions of the SERVQUAL-P scale for normative expectations, predictive expectations, and importance. To achieve this, we applied exploratory factor analyses, and with the dimensions obtained we applied confirmatory factor analyses to validate and confirm the factors.

#### Results

In the first stage, our results show profound differences between the average values of normative expectations and predictive expectations (**Table 5**). The parametric test (t-test) and the non-parametric test (Wilcoxon) reflect significant statistical differences, with a p-value lower than 0.01 for all variables.

If we compare the normative and predictive expectations with the importance of the attribute, we also observe profound differences in 15 of the 16 variables (p ≤ 0.05).

In the second stage, the results of the exploratory factor analyses conducted on normative expectations and predictive expectations show three dimensions. For importance, we obtain four dimensions (Appendix A in Supplementary Material).

To obtain the solution of the three confirmatory factor analyses, we carried out a series of modifications. We applied the Lagrange multiplier test and calculated the Wald statistic,


which evaluates the effect of freeing (or not) a group of parameters simultaneously (Hair et al., 1999). We considered the convergence of parameters in the factors to respecify the model<sup>3</sup> .

With regard to goodness-of-fit indexes on the three scales, the results were satisfactory (**Table 6**). The composite reliability coefficient shows values >0.7 (Appendix B Supplementary Material). With regard to convergent validity, the indicators converge in the factors assigned (standardized lambda parameters >0.5 and significant). The average variance extracted is ≥0.5 for all factors. Regarding discriminant validity (**Table 7**), the covariance between factors indicates that they differ from each other in each model. The confidence interval for the covariance value does not include the value of 1, and therefore there are no covariance issues among the factors involved.

From **Table 7**, we can identify three factors in the normative expectation (Model 1) and predictive expectation (Model 2) models: "service attitude and trust" (F1), "store appeal" (F2), and "personal relationship" (F3). Our results show that the dimensions of normative and predictive expectations are essentially the same. The only difference is that in F1 (service attitude and trust) of the predictive expectations, there is one additional item: "well-trained and knowledgeable employees." With regard to the importance scale (Model 3), the dimensions obtained were "trust" (F1), "store appeal" (F2), "personal relationship" (F3), and "courteous attention" (F4).

When we performed the confirmatory factor analysis, the fundamental difference between the two expectation scales (normative and predictive) and the importance scale is that in the latter scale, there is a new dimension—"courtesy in the attention" (F4)—in which the variables "polite employees" and "friendly employees" are integrated. We removed these two variables from the predictive and normative expectation scales because the lambda factor was <0.6.

Furthermore, as the exploratory factor analysis (Appendix A in Supplementary Material) shows, they do not have a clear assignment. In addition, we checked whether the model could be improved by introducing a dimension with these two variables. The result was a model with a poorer fit.

For the importance scale, respondents are clearer that there is a dimension relating to polite and friendly behavior (the dimension's attributes covary with one another, but not with attributes of other dimensions).

### Discussion

Our results show that there are significant differences between normative expectations (what it should be) and predictive expectations (what they actually expect the service quality to be). However, the underlying structure, which we obtained from the factorial analysis, is essentially the same. In this sense, we can partially accept H2. With regard to the comparison

<sup>3</sup>As Anderson and Gerbing (1988) note, "you can obtain greater convergence of the model by respecifying one or more problematic indicators from different constructs or by excluding these parameters." In our case, we considered it appropriate to eliminate parameters that contributed little to the factor to which they belonged (lambda < 0.6). However, to carry out any respecification of the model, we considered that the modification must be supported by theory.

#### TABLE 5 | Significant statistical differences.


<sup>a</sup>NE, normative expectations; PE, predictive expectations; Imp, importance of the attribute.

<sup>b</sup>Sig.T, p-value of the t-test; Sig.W, p-value of the Wilcoxon test.

#### TABLE 6 | Goodness-of-fit Indexes of each model.


<sup>a</sup>BBNFI, Bentler–Bonett normed fit index; BBNNFI, Bentler–Bonett non-normed fit index; CFI, comparative fit index; GFI, goodness-of-fit index; AGFI, adjusted goodness-of-fit index; RMSEA, root mean square error of approximation.

between normative and predictive expectations and importance, differences arise both in the observable variables and in the structure of the underlying dimensions. As a result, we can accept H<sup>3</sup> and H4.

#### Study 2 Analysis B

In their SERVQUAL-P scale, Mittal and Lassar (1996) establish four dimensions (reliability, responsiveness, personalization, and tangibles). In the personalization dimension, they include the following aspects: "the store employees show personal warmth in their behavior," "the employees are polite and courteous," "the employees are friendly and pleasant," and "they take their time to know the customer personally." However, our results pertaining to the importance of the attributes show that the personalization dimension is divided into two subdimensions: one related to courteous attention and another to personal relationship. These results encouraged us to study these dimensions in greater detail.

#### Method

We analyzed the average value of the attributes included in the dimensions of the importance scale for service quality. Then, we applied a sequential cluster analysis to examine whether any dimension stands out in any segment.

#### TABLE 7 | Analysis of discriminant validity.


TABLE 8 | Average value of the attributes included in the dimensions of importance.


#### Results

The average values of the attributes included in each dimension (**Table 8**) provide evidence that the most important factors in the average score are trust (8.95) and courteous attention (8.94). These two factors are followed by store appeal (8.22). The least important factor is personal relationship (7.07).

It comes as a surprise that the courteous attention dimension is highly valued while the personal relationship dimension is the least important. Our question now is whether there is any group of consumers for whom personal relationships stand out as the most relevant dimension.

To classify customers into groups, we applied a hierarchical cluster analysis based on the dimensions of the importance scale. We used squared Euclidean distance as a proximity measurement and the Ward method as an algorithm for classifying. A dendrogram enabled us to establish the number of clusters and the centroids to subsequently apply the K-means method. As a result, we obtained four clusters (see **Table 9**), whose validation we carried out through two methods: variance analysis and discriminant analysis. The validation was satisfactory (see Appendix C in Supplementary Material).

With regard to the description of the groups, we assigned a name to each cluster based on the importance the customers gave to the factors of service quality.

#### Group 1: Highly Concerned

These customers are most concerned about high service quality that incorporates personalized attention. This group values service quality in all its dimensions: trust, store appeal, personal relationship, and courteous attention. However, personal relationship is the least valued dimension. This group shows the least differences between the average values of the factors. The gap between the most valued factor (courteous attention = 9.73) and the least valued factor (personal relationship = 8.44) is 1.29 points. This cluster comprises 27.46% of the sample.

#### Group 2: Concerned

This group represents more moderate customers when it comes to the importance placed on service quality (average values of the factors are close to 0 points). The average values of the items of each factor are high. The gap between the most valued factor (courteous attention = 9.14) and the least valued factor (personal relationship = 7.06) is 2.08 points. This group comprises 38.92% of the sample.

#### Group 3: Value Quality Less

This group includes people who are less concerned than Groups 1 and 2 about the service quality of the store. The difference between the most valued factor (trust) and the least valued factor (personal relationship) is 2.23 points. This cluster comprises 28.00% of the sample.

#### Group 4: Reject Personal Relationships

The people in this group have values on the scale that are close to 5 points (on a scale from 0 to 10), and they reject personal relationships (with values lower than 5 points). This group has the largest gap (2.68 points) between the most important dimension (trust) and the least important dimension (personal relationship). This is also the smallest group, representing only 5.62% of the total sample.

Overall, 66.38% of respondents valued courteous attention as the dimension of service quality that is most important, while for the rest it is the second most valued dimension. For 33.62%, the trust dimension—which includes proper service, capable of handling problems, and keeps promises—is the most important. Personal relationship is the least important factor for 100% of the respondents.

### GENERAL DISCUSSION

Consumer behavior represents one of the greatest interests of marketing scholars and business managers due to their need to adapt their companies' strategies to the new frontier. In this context, several authors (Badgett et al., 2007; Gentile et al., 2007; Grewal et al., 2009; Tynan and McKechnie, 2009; Verhoef et al., 2009; Rose et al., 2012) emphasize the importance of the shopping experience when choosing among different retailers. The interactions between retail employees and customers can have a significant impact on customers' perceptions of the organization (Tsiros and Parasuraman, 2006; Gremler and Gwinner, 2008; Lichtenstein et al., 2010; Otnes et al., 2012; Litz and Pollack, 2015). Despite the importance of this interaction, prior research has not answered the question of what customers value in terms of salesperson–customer interaction (Haas and


#### TABLE 9 | Cluster according to importance.

<sup>a</sup>Mf, average value of the factor.

<sup>b</sup>Mi, average value of the factor's items on a scale from 0 to 10.

Kenning, 2014). Thus, our study fills an existent gap on customersalesperson relationship quality in retail.

In study 1, we analyze the primary motivations for customers going to shopping streets or malls since the relationship between a sales associate and a customer is dynamic and not universal and it varies across different services and industries (Kim and Jin, 2001). Our results show that the primary motivations for customers going to shopping streets or malls are different. Personalized attention is the most important factor cited by customers who prefer shopping streets. Therefore, this result is in line with numerous previous findings (Goodwin, 1996; Menon and Dubé, 2000; Stock and Hoyer, 2005; Schau et al., 2007; Gremler and Gwinner, 2008) that suggest the importance of the salesperson–customer interaction in defining the consumer's experience and ultimate satisfaction. However, personal attention is not as relevant for those who prefer malls. This is a source of competitive advantage for shopping streets, and it is the reason we studied consumer behavior in relation to service quality in these types of stores in greater detail.

Expectations also play a significant role in our research question. Retailers seek to manage customers' service expectations (Mitra and Fay, 2010). The literature on expectations is broad, and many ways of understanding and studying expectations have been found. We have classified expectations into ten types, which we have then grouped into four main approaches: (1) comparison, (2) ideal amount, (3) levels, and (4) point of assessment. Several research has argued that the service quality a customer receives can be measured using comparison approach (Boulding et al., 1993; Zeithaml et al., 1993; Dean, 2004; Higgs et al., 2005; Golder et al., 2012). This approach distinguishes and conceptualizes normative service quality expectations and predictive quality expectations. In Study 2 analysis A, we compare normative expectations (what the customer believes the service quality should be) and predictive expectations (what the customer expects the service quality will actually be) of the customers of shopping streets. The results demonstrate deep differences in the mean scores of each attribute of service quality (incorporating personalization). In keeping with numerous previous findings (Boulding et al., 1993; Johnson and Mathews, 1997; Higgs et al., 2005), expectations for normative expectations were generally higher than for the predictive expectations. However, the dimensions from the factor analysis are essentially the same: service attitude and trust (F1), store appeal (F2), and personal relationship (F3). These results are consistent with the findings of Kalamas et al. (2002), but contrary to the conclusions of Higgs et al. (2005). If the underlying dimensions of the factor analysis are the same under normative and predictive expectations, this means that the customers are coherent in their assessment of these expectations. In other words, they evaluate service quality (normative expectations) and satisfaction (predictive expectations) following a similar mindset.

According to Teas (1993), another aspect to consider in the assessment of service quality is the importance of each attribute used to measure it. Its comparison with normative and predictive expectations is crucial in the answer to our research question. Once it has been compared, differences arise both in the items and in the structure of the factors between expectations and importance of the attributes.

We were surprised to find that the personalized attention (in the importance scale) was divided into two subdimensions in our work: courteous attention and personal relationship. In Mittal and Lassar (1996), aspects related to courteous attention and aspects related to personal relationships were integrated in the personalization factor. Likewise, Gagliano and Hathcote (1994) also find a dimension related to personal attention that integrates aspects related to courtesy and the concern to help, though the notion of personal relationships is not included. Haas and Kenning (2014, p. 436) argue that "great service begins with showing courtesy to everyone, customers and coworkers alike." Ulaga and Eggert (2006) identified service support and personal interaction as core differentiators in business relationships.

Our subsequent question was whether there was any group of consumers for whom personal relationship stands out as the most relevant dimension. In Study 2 analysis B, we classified the customers into four groups based on the dimensions of the importance scale.

Courteous attention is the most important factor for two of the groups, while for the other two it is the second most important factor. However, personal relationship is the least valued dimension in all groups.

Our results from Study 2 analysis B are consistent with those from Study 1 and according to Gremler and Gwinner (2008), the results show that the quality of service and the degree of personal attention are important factors in consumer behavior. In Study 1, the personal attention dimension was divided into five central concepts, one of them being personal relationship. Customers who preferred shopping streets value polite and courteous attention as well as close and personalized attention. However, only 4.09% of those surveyed commented that one of their motives is the personal relationship with the retailer. Therefore, there are consumers who like to maintain a personal relationship, but they are a minority.

### MANAGERIAL IMPLICATIONS

This article advances the knowledge about this new consumer behavior and helps business managers, giving them guidelines to adapt their companies' strategies to the new frontier.

Urban shopping streets can be revitalized using personal attention in a professional manner, while stopping short of forging a personal relationship with the customer. With regard to consumer behavior, managers should take care to cultivate what kind of personal attention they offer their customers, recognizing that the personal relationship is the least valued factor in service quality.

Because the motivations for choosing malls vs. shopping streets are different, managers should consider the following recommendations. First, for managers of shops located in shopping streets, it is important to note that for 66.38% of the respondents, the most important aspect of service quality is courteous attention, and for 32.62% of the respondents, trust is the most important aspect. Thus, we recommend the following:


Second, for mall managers, we recommend building on the concept of "big shopping" as an excursion and consolidating their strength in the commercial offer. Commercial communication campaigns should stress the advantages of spatial convenience (everything in the same place) and accessibility (opening hours and parking). Young people show a clear preference for the mall for shopping but not for leisure activities, though in the malls in the study area there are also leisure activities. Our recommendation is to attract young people with appropriate entertainment proposals.

Because customers perceive differences between what should be (normative expectations) and what they actually expect (predictive expectations), management should carry out actions that aim to meet customers' normative expectations. This effort should focus on the attributes that are important to customers (importance scale).

### LIMITATIONS AND FURTHER RESEARCH

Although we believe our results add an important contribution to the literature, it is difficult to determine whether they are unique to retail in the area of study or whether they can be extrapolated beyond this environment. Therefore, we recommend that our study be replicated in other regions, such as the United States, India, or Germany, which have different cultures.

In our work we included all the retailers in the area of study as a whole. However, it would be worthwhile to conduct research that involves comparing various types of stores, as there may be differences for each category.

For Group 4, which rejects personal relationships, it would be worthwhile studying how this preference affects shopping behavior on the Internet to study e-commerce in greater detail.

Our results complement previous studies in that we have found new reasons to go shopping. We identify, among others, solidarity with traditional trade as a reason for opting to purchase in stores located in shopping streets. This result is connected to the world of emotions as a source of competitive advantage for traditional trade.

We agree with Puccinelli et al. (2009, p. 24), who argue that "the interpersonal nature of the interaction between the customer and employee [...] may be key to customer satisfaction in the retail environment." Moreover, retailers located in shopping streets must manage every encounter with the customer as a unique opportunity, in which personal attention should be the main tool for satisfying and building loyalty with the customer. Using this strategy to differentiate themselves from and compete with larger stores and malls, smaller businesses located in shopping streets might be able to effectively reestablish their retail niche and relevance to consumers.

### AUTHOR CONTRIBUTIONS

The four authors have equally participated in literature review, data analysis and writing of the paper.

### ACKNOWLEDGMENTS

This work has been funded by The Ministry of Economy and Competitiveness (Spain), Research Project reference no. ECO2014-59688-R, Programa Estatal de Investigación, Desarrollo e Innovación Orientada a los Retos de la Sociedad, Plan Estatal de Investigación Científica y Técnica y de Innovación 2013–2016.

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fpsyg. 2016.00125

### REFERENCES


in Conceptualization and Measurement of Consumer Satisfaction and Dissatisfaction, ed H. K. Hunt (Bloomington, IN: School of Business, Indiana University), 72–91.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Medrano, Olarte-Pascual, Pelegrín-Borondo and Sierra-Murillo. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Multi-Vendor Loyalty Programs: Influencing Customer Behavioral Loyalty?

#### Teresa Villacé-Molinero, Pedro Reinares-Lara\* and Eva Reinares-Lara

Departamento de Economía de la Empresa, Universidad Rey Juan Carlos, Madrid, Spain

Loyalty programs are a consolidated marketing instrument whose adoption in many sectors has not been associated with appropriate comprehension of either their management elements or their effects. The purpose of this research is to contribute to knowledge about the effect of loyalty programs on repeat purchase behavior. More specifically, it seeks to discover whether joining a program changes the buying behavior of its members, and, if so, to study the profile of those whose behavior changes most. The intention was also to provide new study variables pertaining to multi-vendor loyalty programs, such as where they are joined or purchases in associated outlets as a result of behavioral loyalty. Research was carried out using a sample of 1200 individuals (31,746 purchases) belonging to a multi-vendor loyalty program. The study period was 13 years, 4 months, and split into two phases: before and after the joining the program. Different methodological approaches, such as the use of transactional databases that included pre-program-enrollment data and of the same sampling units throughout the study, were incorporated into the research with the aim of advancing academic knowledge regarding multi-vendor loyalty programs. Moreover, a type of program and market hardly dealt with in the relevant literature was analyzed. The results showed while the loyalty program had managed to reduce the time between purchases, it had not affected purchase volume or average expenditure. They also demonstrated the existence of a differential profile of customers who had changed their buying behavior to a greater extent. Finally, recency was identified as being the decisive variable in behavioral change.

Keywords: customer behavior, customer loyalty, behavioral loyalty, multi-vendor loyalty programs, reward programs

## INTRODUCTION

Competition in today's market is intense, which makes acquiring new customers increasingly complicated and less profitable; hence the importance of securing the loyalty of existing customers. In this regard, Alet i Vilaginés and Nueno (2004) point to a shift in emphasis from gaining product trial to seeking customer loyalty, which, they argue, has become a fundamental strategic component for companies. Likewise, business strategies have gone from focusing on customer satisfaction to focusing on customer loyalty, primarily because of the understanding of the impact of loyalty on profits (Oliver, 1999). Some authors have suggested that customer loyalty is essential to achieving business profitability, as there is a high cost associated with acquiring new customers due to the low return on the initial transactions since, with many customers, profitability increases over the course

#### Edited by:

Ana I. Jiménez-Zarco, Open University of Catalonia, Spain

#### Reviewed by:

Emma Juaneda-Ayensa, University of La Rioja, Spain Emira Josefina Rodriguez, Universidad de Oriente, Venezuela

> \*Correspondence: Pedro Reinares-Lara pedro.reinares@urjc.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 12 December 2015 Accepted: 03 February 2016 Published: 23 February 2016

#### Citation:

Villacé-Molinero T, Reinares-Lara P and Reinares-Lara E (2016) Multi-Vendor Loyalty Programs: Influencing Customer Behavioral Loyalty? Front. Psychol. 7:204. doi: 10.3389/fpsyg.2016.00204 of their relationship with a company (Anderson et al., 1994). Additionally, companies can benefit from other advantages related to having loyal customers: lower service costs, less price sensitivity, higher purchasing levels, and positive word-of-mouth (Sharp and Sharp, 1997).

The various existing conceptualizations of loyalty mainly revolve around two accepted dimensions (Bloemer and De Ruyter, 1998). The first is behavioral loyalty, which refers to commitment as expressed through behavior, that is, actual repeat purchasing. In this dimension of loyalty, importance is not given to the reasons for the repeat patronage, as what matters is the act of repetition itself. Thus, loyalty is measured in terms of the number of store visits, the number of repeat purchases at a single establishment, etc. The second dimension is attitudinal loyalty, which refers to the psychological or affective bond customers show toward a given product or brand in the form of a positive attitude toward repeat purchasing or even by engaging in positive word-of-mouth. Attitudinal loyalty shows the customers' preferences and inclination toward a given provider. Thus, customers are considered loyal when, in addition to engaging in repeat patronage with regard to a given product, brand, or establishment, they show an inclination, conviction, or favorable attitude toward it (Heiens and Pleshko, 1996), that is, when they evince both dimensions of loyalty. Given the importance of this subject, considerable research has been conducted on loyalty with a view to determining how to successfully build a long-term relationship with customers.

However, due to the difficulty of evaluating the psychological component of loyalty, most authors, while not denying the attitudinal dimension, have focused on the analysis of behavioral loyalty (O'Malley, 1998), which is easier to measure objectively.

In keeping with the business goal of achieving behavioral loyalty, recent years have witnessed a boom in loyalty programs based on cumulative reward point cards or obtaining discount vouchers (Zhang and Breugelmans, 2012). In order to encourage repeat purchases, these programs offer customers incentives (points or other exchange units), which can be redeemed for different items of value. The proliferation of loyalty programs reflects a changing market environment that is increasingly characterized by intense competition, more demanding and knowledgeable consumers, and a development toward relationship marketing and customer relationship management in marketing thinking and practice (Liu and Yang, 2009).

Loyalty programs, appropriately managed, are considered to allow structured and effective actions to manage, select, relate, and control customers' buying behavior.

Such programs have been extensively used by companies to motivate customers to increase their purchase volume and frequency. The assumption is that consumers will stay loyal only if suppliers provide an integrated service platform, that assures, that loyal customers are more favored than the rest of the market (Miranda and Kónya, 2008). It is thought, that correctly managed loyalty programs allow for structured, operational actions for managing, selecting, relating and controlling customer buying behavior (Banasiewicz, 2005).

McCall and Voorhees (2010) have argued, that the current lack of understanding of what factors drive a program's success is a major knowledge gap, that affects the optimization of how they are managed. Despite their development, the study of their effect on customer behavior in existing literature has provided contradictory results (Bojei et al., 2013). Dorotic et al. (2012), in their descriptive review approach to this issue, explained this disparity by claiming that the differences in results depended on several factors: (a) the area of application, or rather, the products involved or the markets where the programs were launched, (b) the type of customer segments analyzed, (c) the type of program (mono-sponsor/multi-sponsor), and (d) the methodology used in the research. These authors highlighted the need to continue investigating loyalty programs on the basis of the factors identified which affect the results.

This article will focus on the impact of loyalty programs on repeat purchase behavior (purchase loyalty). This is because, in practice, only repeat purchase behavior is rewarded, and not attitude, as loyalty schemes are often based on classic promotional techniques, with delayed, or immediate rewards (gifts, price reductions, points, etc.) or relationship marketing techniques (access to privileges or services, special status, individualization, etc.), which encourage consumers to purchase more often and remain loyal to the store. We can thus assume, that if a certain number of loyalty cardholders make similar changes in their buying behavior, there will be visible changes in repeat purchase patterns at a store level (market share, penetration, average purchase frequency; Meyer-Waarden and Benavent, 2006).

This research is focused on studying behavioral loyalty in multi-vendor programs. These consist of companies from different sectors joining together under one umbrella brand, sharing the implementation and management costs of the program, with the aim of providing cardholders with a diversified offer of where to obtain and redeem points rapidly. A standalone program itself does not promote the acquisition of new customers as easily as a multi-vendor loyalty program, by which new customers can be won over from other program participants (Rese et al., 2013). According to authors such as Lemon and Wangenheim (2009), this type of program has strategic advantages due to inter-company cooperation favoring crossselling. Nevertheless, others (Sharp and Sharp, 1997; Dorotic et al., 2011) limited the networking effects across competing partners.

Within this context, the proposed general aim was to study whether joining a multi-vendor loyalty program caused a change in buying behavior, although several, more specific objectives were also set. Firstly, the aim was to analyze whether, in those cases where a change in buying behavior did occur after the program was joined, the customers who most modified their behavior had a differentiable profile. If so, companies would be able to carry out relationship marketing activities aimed at this specific group. And secondly, the intention was to deepen the study of multi-vendor loyalty programs in order to examine the possibility of extrapolating the existing knowledge about mono-sponsor loyalty programs toward them. Thus, what were effectively new study variables when related with buying behavior in the field of multi-vendor programs were used. These included: the place of enrolment, purchases in other establishments belonging to the program, the number of other establishments where the customer buys using the program, and even the feasibility of using variables such as recency or accreditation level for the operational management of multi-vendor programs.

The field of application for this study was small businesses; more specifically, medium-priced consumer goods, such as those found at optical shops, were analyzed. This sector application represents an advance in the business management of these programs since, it does not deal with a sector, that has been routinely studied (such as airlines, hotels, and grocery stores).

Finally, for the purposes of making advances in loyalty program research with adequate scientific rigor, a new methodology was applied, consisting of the use of transactional databases incorporating purchases made prior to when the program was joined and sampling units, that remained constant throughout the study (longitudinal analysis).

### CONCEPTUAL FRAMEWORK

### The Influence on Behavioral Loyalty of an Establishment Joining A Multi-Vendor Loyalty Program

Authors such as Lewis (2004), Taylor and Neslin (2005), Kivetz et al. (2006), and Bridson et al. (2008) have stated, that loyalty programs permit the modification of customer buying behavior and, therefore, improve loyalty to the establishment in question. Drèze and Hoch (1998) claimed, that such programs had a positive effect on both the size and the average cost of purchases. Likewise, for the vast majority of researchers, belonging to a program reduced the time between purchases (Lewis, 2004).

Other authors have also confirmed the influence of loyalty programs on behavioral loyalty, but limiting their effects to the short term, therefore, usually associating them with promotional rather than relationship marketing activities (Benavent and Crié, 2000).

However, Dowling and Uncles (1997) and García Gómez et al. (2006) put further limits on the influence of loyalty programs to change customer buying behavior. Sharp and Sharp (1997) even alleged, that such programs neither modified buying behavior nor increased market share. Similarly, Long and Schiffman (2000) maintained that only a small number of customers modified their behavior after joining a program.

Finally, Meyer-Waarden (2008) pointed out, that changing buying behavior in multi-vendor programs was even more complicated than in mono-sponsor ones. In this type of program, in order to obtain their reward, customers do not even need to buy more frequently in outlets, that they did not use before, since, by being multi-vendor, they can continue accumulating points in their usual outlet. Moore and Sekhon (2005) confirmed the lack of influence of multi-vendor programs on improving market share.

This rather contradictory background and the methodological limitations identified in previous research make the formulation of the following hypothesis coherent:

H1: The incorporation of an establishment into a multivendor loyalty program produces a change in the behavioral loyalty of its customers, as measured in terms of buying behavior.

In this research, in accordance with previous literature, buying behavior was measured on the basis of the following variables:


### The Profile of Those Customers Who Most Modified their Buying Behavior after Joining the Program

According to various authors, actions carried out within loyalty programs do not have the same impact on response for all customers (Dorotic et al., 2012). For example, loyalty programs apparently have more effect on the "light" and "moderate buyers" (Lal and Bell, 2003). This, of course, seems logical since the "heavy users" have less capacity to increase their purchasing volume or frequency; nevertheless, they maintain their high usage levels. Moreover, "early adopters" are usually heavy users, given that they are already loyal customers, living close to the shop, holding a joint family card and with a tendency to join up quickly in order to take advantage of the program's benefits (Leenheer et al., 2007; Demoulin and Zidda, 2009). Therefore, since differences in buying behavior due to customer type exist, it can be stated that one of the main advantages of loyalty programs is that they make relationship segmentation possible (Reinares and Ponzoa, 2008; Drèze and Nunes, 2009).

Within this context, the intention was to analyze the buying profile of the customers who most change their behavior, thus allowing for the identification of those customer groups that it was more interesting to incorporate into the program, due to their more favorable response.

With the aim of checking this argument within the realms of multi-vendor programs, a new hypothesis was formulated. Therefore, for each one of the buying behavior variables used to measure behavioral loyalty to the program, the group of customers who most modified their buying behavior after joining the program was selected. This group consisted of 5% of the total. The objective was to compare their behavior with, that of the rest in order to verify whether there existed differences between the two groups. The proposed hypothesis was as follows:

H2: Significant differences exist between those customers who change their buying behavior to the greatest extent after joining a program and the rest.

### The Influence of the Point Accumulation Rate on Behavioral Loyalty

The point accumulation rate is considered to be a very useful variable in detecting high-value customers. It is defined by the number of points that a customer builds up using the program's card for purchases within a given period of time (Ponzoa and Reinares, 2010). Loyalty programs use this variable to classify customers on the basis of their program transactions, or rather, the higher the accumulation rate, the greater the number of purchases. Therefore, the intention was to verify whether, in effect, the accumulation rate had a bearing on the traditional buying behavior variables for multi-vendor programs.

In a multi-vendor program, the accumulation rate depends on the total expenditure within the program, which is to say, within the whole set of establishments pertaining to it. Therefore, it is more easily increased due to the greater number of outlets where points can be accumulated. At the same time, however, it is more difficult to use the accumulation rate to identify those customers with higher transactional volume (since they may be very active in some outlets while being completely inactive in others).

This can be attributed to the "double jeopardy phenomenon": that is to say, the greater the market penetration of the brand, the greater the buying frequency and the category quota bought, and vice-versa, so loyalty programs are more effective for market leaders than for small companies (McGahan and Ghemawat, 1994). Smaller companies not only have fewer customers, they also have a lower purchase frequency (Ehrenberg et al., 1990). Therefore, companies belonging to a loyalty program can obtain very varied results, as previously described, depending on the market penetration of each one's brand.

Thus, with the aim of testing whether this variable could be useful in the taking of operational decisions in the businesses belonging to multi-vendor programs, the following hypothesis was formulated.

H3: The accreditation level (point accumulation rate) of customers in a multi-vendor loyalty program influences behavioral loyalty toward one of the member establishments.

### The Influence on Behavioral Loyalty of the Establishment Where the Program is Joined

Multi-vendor loyalty programs exhibit differentiating elements, with respect to mono-sponsor ones, which it is necessary to analyze. The possibility of accumulating and redeeming points in the different member establishments of the multivendor platform makes a deeper study of this differential aspect necessary. Thus, questions as to whether the establishment in which the program is joined exerts an influence on behavioral loyalty or whether purchasing in other member establishments implies a specific buying behavior arise.

If prior related literature is analyzed, according to Meyer-Waarden (2008), the programs first attract the high-volume shoppers of the establishment in which the multi-vendor program is joined. This is due to the so-called "self-selection" effect by which the "heavy users" show a greater probability of participating in a program than other customer types with a lower buying volume and lesser frequency (Leenheer et al., 2007). Therefore, by extrapolating these results to the multivendor loyalty program environment, one can logically infer that the customers that joined the program through the analyzed retailer (and not through another of the program's member establishments) could well have a buying behavior similar to that of the "heavy users" with regard to the purchase of products in that particular retailer since they were already buying at that particular optical shop prior to its implementing a multi-vendor loyalty program. Thus, these customers would have been "self-selecting" themselves. On the other hand, those customers who joined the multi-vendor program through other affiliated companies (establishments from different, noncompeting sectors: fuels, food, banking, etc.) would behave in a similar way to "light users" in the purchase of optical products (less frequent, lower volume), but would exhibit a greater change in buying behavior.

So, taking this into account, the following hypothesis regarding the establishment where enrollment takes place was proposed:

H4: The establishment where a program is joined has an influence on the change in buying behavior after joining a multi-vendor loyalty program.

### The Effect on Behavioral Loyalty of Buying in Other Multi-Vendor Program Member Establishments

There have been very few studies on the impact, that the associate companies in a multi-vendor loyalty program have on buying behavior, since the majority of studies have been focused on mono-sponsor programs where relationships with "partners" simply do not exist.

Among those who have analyzed multi-vendor programs, Dorotic et al. (2011) compared the response of multi-vendor loyalty program members to individual brand promotions with their response to joint promotions within the program itself. The results demonstrated the low effectiveness of joint promotions once customers had joined the program, in such a way, that the customers responded better to the individual promotions of each brand. Therefore, within the realm of promotions, multi-vendor programs had not displayed any advantage over mono-sponsor ones.

Given the scarcity of bibliographical sources specifically relating to multi-vendor programs, prior literature regarding mono-sponsor programs with associated companies was also analyzed and taken into account, since these are the programs that most resemble multi-vendor ones.

In this area, authors such as Lemon and Wangenheim (2009) studied how the use of and satisfaction with buying the principal product of the program ("core service") influenced the purchase of other products in establishments associated with it.

They specifically verified how, as use of and satisfaction with the core service increased, so did cross-purchasing from other services in the program, and vice versa. However, they confirmed that cross-purchasing in one category did not seem to strengthen cross-buying between other categories; that is to say that the relationship was only confirmed when there existed a main product/core service within the program. Moreover, according to these authors, this relationship was influenced by the kind of service that the partners in the program offered; the more, that this fit with and complemented the core service, the greater was the cross-purchasing effect.

In contrast with these results, other authors have found, that satisfaction with the core service of an enterprise did not exert a positive influence on the cross-purchase of other services, and also, that this itself was negatively influenced by the number of services joined to it in the recent past (Verhoef, 2003). The problem with adapting these results to multi-vendor programs is the lack of a "core service" or central product. Nevertheless, according to Moore and Sekhon (2005), the majority of the purchases of a multi-vendor program customer take place in the two leading member establishments, which act as the driving force of the program.

Thus, it would be interesting to know whether the customers that only shopped at the optical shop (as if the program were mono-sponsor) demonstrated higher behavioral variable values (shopping basket, average purchase price, etc.) than the customers that also bought in other member companies (food, clothing, etc.). Therefore, the following hypotheses were proposed:

H5a: Purchasing in other multi-vendor program member establishments influences behavioral loyalty toward other companies associated with the program (in this case: the optical shop).

H5b: The number of different sectors in the program in which a customer buys using the multi-vendor program influences behavioral loyalty toward other companies (the optical shop) associated with the program.

#### METHOD

As explained in the presentation of the dimensions of loyalty, like other studies (Nilsson and Olsen, 1995), this study focuses on the behavioral dimension. Several authors (Lewis, 2004; Lacey, 2009; Lemon and Wangenheim, 2009) considered that, in order to measure the success of a program, it was necessary to carry out longitudinal research, which observes how buying behavior changes over time, as opposed to cross-sectional studies, which offer a snapshot view of buyer behavior. Similarly, the longitudinal research carried out in this field has normally analyzed either the change in buying behavior after joining the program (Lal and Bell, 2003; Verhoef, 2003; Taylor and Neslin, 2005; Liu, 2007; Tsao et al., 2009), or the change produced by changing from a traditional program to one with specific characteristics (Zhang and Breugelmans, 2012).

With the aim of overcoming the limitations identified in other work, the proposal in this study was to compare consumers' behavior before and after they joined the program.

Of the two multi-vendor loyalty programs existing in the Spanish market, the one that possessed the most ideal formal characteristics for fulfilling the proposed objectives and the greatest typological representativeness of Spanish households was chosen for the transactional data analysis.

The target population consisted of Spanish residents, aged 18 and over, who were loyalty program members. According to data from the company PSM (2013) the study universe consisted of some 22,750,000 individuals.

The study took place in 14 urban areas in Spain and its field of application comprised small retail establishments (more specifically, optical shops). The study period was 13 years 4 months, giving it a greater timespan than previous research. The study period was divided into two parts: (a) "before joining the program" (January 2, 1996–March 30, 2002) and (b) "after joining the program" (April 1, 2002–May 9, 2009). The date the establishment joined the program was April 1, 2002.

The sampling frame comprised those customers of the optical shop chain who, being members of the selected loyalty program, had made at least one purchase in each of the two periods under study. The selection procedure for the sampling units within the program used was simple random sampling.

A sample of 1,200 individuals who had effected a total of 31,746 purchasing actions in the two periods studied was obtained.

Concerning the information about the ethics requirements, this study was made according to criteria of the Ethics Department of the Rey Juan Carlos University and to the ICC/ESOMAR International Code on Social Research. Personal or sensitive data were deleted of the initial database because they did not provide relevant information for the analysis.

The study variables were calculated by cross-referencing the electronic transactional data from the multi-vendor program with that of the retailer studied.

The initial database was modified with a complex conversion process to suit the objectives of the study, and some existing variables were transformed according to the criteria used in the literature; to calculate these variables, algorithms programmed in Visual Basic for Applications were used. Also, specific new variables on the operation of loyalty programs were added based on the recommendations given by industry experts.

Finally, a total of 45 variables grouped into 9 constructs were used for the analysis, related to: (a) purchase behavior, (b) change in buying behavior after joining the program, (c) purchase behavior by category of product purchased, (d) recency, (e) multi-vendor purchase behavior (variables related to purchases made in other establishments), (f) type of consumer, (g) level of accreditation, (h) communication channels used by the multivendor program, and (i) redemption of points.

### RESULTS

In order to verify the indicated hypotheses, a T-test for related samples and, later, a Wilcoxon test to confirm the results obtained were carried out (see **Table 1**).


#### TABLE 1 | Results: T and Wilcoxon tests.

The first variable used to measure the influence of joining the program on behavioral loyalty was volume of basket. The average per customer increased after joining the loyalty program, but, as can be seen in **Table 1**, this increase was not statistically significant since in neither of the two tests was the significance level obtained lower than 0.05.

The second variable used for the same ends was the average purchase price. This value was very slightly lower in the time after joining the program in comparison with that spent in the preceding period. The results confirmed that the difference in averages was significant. This decrease was due to a progressive price reduction that some products at the optical shop had been undergoing and not to the program in itself.

The inter-purchase time was also reduced upon joining the program. The tests used confirmed, that these differences in average were indeed highly significant since the critical levels were almost 0. This variation was very positive for the establishment, as hoped for, since it implied a greater buying frequency. Likewise, and as a result of this change, it was noted that joining the program had increased the average number of visits per year to the member establishment. The tests used to check the difference in averages confirmed, that this difference was also significant.

As happened with the average purchase price, the change in the number of articles and product categories purchased was not as expected since these decreased after the program was joined. This difference in averages was confirmed statistically as can be seen in **Table 1**.

The last measurement variable used for the purposes of this research was the average annual expenditure of the customers. As was the case with the average purchase price, there was a decrease in the second period analyzed. However, here, the two tests used differed. In the T-test the results proved to be significant (p = 0.01), while in the Wilcoxon test they were not (p = 0.06); the significance level was slightly higher than the critical value. It could therefore not be confirmed that joining the program had caused a change in average annual expenditure.

These results partially confirmed Hypothesis 1. Although a change took place in all the variables used, after joining the program, the differences only turned out to be significant and positive for the establishment in the case of reducing interpurchase time and increasing the average number of shop visits.

Next, in order to analyze the profile of those customers who had most changed their buying behavior after joining the program, the sample was split into two groups comprised of those customers who had greatly changed their behavior compared to those who had done so to a much lesser degree. The difference in averages considered was the difference between the 5% of customers that had most increased their buying behavior and the remaining 95% who had done so to a much lesser extent. Afterwards, a comparison of the differences in behavior for the two sub-samples both "before" and "after" was carried out for each of the buying behavior variables.

The results obtained for each one of the buying behavior variables demonstrated, that the change produced was related to several variables from the period prior to when the loyalty program was joined. Thus, the variables, that had most significantly influenced the buying behavior of the customer group analyzed were: recency in the pre-joining period and previous average expenditure on key optical products, lenses and graduated lenses (see **Table 2**). More specifically, recency was significant for all the buying behavior variables. Those customers who exhibited greater recency before joining were identified as those who most changed their volume of basket (202.14 days since, the last purchase), their average purchase price (recency: 417.20 euros), their average inter-purchase time (recency: 375.66 days), the average number of articles bought (recency: 289.74 days), and their average number of store visits (recency: 211.69) and also those who most increased their average annual expenditure (120.87 euros).

In the case of average annual expenditure on core products (lenses and graduated lenses), this greater change was noted for all the buying behavior variables except that of the average number of articles purchased.

It is also interesting to note, that the consumers who showed the greatest change in number of store visits (-0.14) were customers whose enrollment had taken place in program member establishments other than the optical shop in question and who bought in other sectors, such as clothing (0.14). Moreover, purchasing in the fuel sector caused a greater increase in the average purchase price for the customers in the analyzed group (0.016). Therefore, on the basis of these results, Hypothesis 2 (H2) can be accepted.

Having determined the influence of recency on buying behavior change, the influence of the point accumulation rate on behavioral loyalty was then verified. In order to do this, the degree of relationship between the average accumulation rate and each one of the buying behavior variables in the post-joining

#### TABLE 2 | Profile analysis of the greatest buying behavior modifiers (5%).


time period was studied (see **Table 3**). The results of the Pearson correlation coefficient showed a negative relationship between the average accumulation rate and the following variables: volume of basket, average number of articles bought and number of store visits. The relationship was therefore positive for the remaining variables—average purchase price, inter-purchase time, average number of product categories bought and average annual expenditure. The correlation coefficients and p-values confirmed the absence of correlation between the analyzed variables. Therefore, the third hypothesis was rejected.

Finally, the results obtained for the variables particular to multi-vendor programs—the establishment where enrollment occurred and purchases in other program member establishments—were analyzed.

First, the influence of the "enrollment establishment" on buying behavior change after the multi-vendor program was joined was examined. The results of the tests carried out confirmed, that the differences, that existed on the basis of the enrollment establishment were only significant with regard to the variables (see **Table 4**): (a) average number of categories bought on each visit, which was higher for customers who joined at the optical store [T(1, 1095) = −2.79, p = 0.01, and W(1, 1095) = 162.940, p = 0.02]; and (b) average number of visits, which was higher for those who joined at other establishments. Here it should be pointed out that the results obtained with the Ttest [T(1, 1186.9) = 1.71, p = 0.09] were not significant, whereas with the non-parametric contrast of the Wilcoxon test it was confirmed that they in fact were [W(1, 1186.9) = 188, 004, p = 0.05], although with only the minimum critical level. Therefore, Hypothesis 4 was partially rejected.

TABLE 3 | Correlation analysis between average accreditation rate and behavioral loyalty: Pearson's correlation coefficient.


(N = 1200).

With respect to the effect on behavioral loyalty of buying at other multi-vendor program member establishments, it was verified, that the customers who purchased at other program establishments, and not only at the optical shop, had greater basket volumes, a higher average purchase price, a lower interpurchase time, and therefore, a higher average number of visits to the store, and, finally, a greater average annual expenditure. On the other hand, those customers who used the program as if it were mono-sponsor, only using the loyalty program when buying at the optical shop, did so on average with a higher number of articles and categories. However, these differences between the two groups of customers were only found to be significant with respect to the average annual expenditure variable [T(1, 399.84) = −2.0, p = 0.05 y W(1, 399.84) = 103, 437, p = 0.05). Thus, Hypothesis 5a was partially rejected.

#### TABLE 4 | Analysis of the influence on buying behavior change of the establishment where enrollment took place.


\*p < 0.05.

TABLE 5 | Analysis of the influence on behavioral loyalty of the number of other sectors where purchases took place: Kruskal–Wallis H-test.


In order to analyze the influence on behavioral loyalty of the number of other sectors of the program in which a customer purchases using the multi-vendor program, the Kruskal–Wallis H-test was used (see **Table 5**). The results showed that customers who only bought at the optical shop had a higher average number of articles bought on each visit to the store, while those who also bought in another sector of the program had a bigger basket volume and number of visits and, therefore, also a lower inter-purchase time. Moreover, those customers who bought in more than two different sectors within the program exhibited a higher average purchase price and also a higher average annual expenditure. However, the p-values of the Kruskal–Wallis H-test indicated that the differences found on the basis of the number of program sectors with purchases were not significant since they all substantially exceeded the 0.5 level. Thus Hypothesis 5b was also rejected.

### DISCUSSION AND CONCLUSIONS

In accordance with the aim of this research, the outcomes of this study are an improvement over the divergent academic contributions that consider loyalty programs either to act on behavioral loyalty or to be mere promotional tools, with no long-term effect on consumer behavior.

This study shows that, when properly managed, loyalty programs can indeed affect behavioral loyalty, although they are not capable of modifying some of the variables used to measure it. It can thus be concluded that loyalty programs are an ideal tool for managing some of the components comprising customers' behavioral loyalty to commercial establishments.

With respect to the results obtained, as was the case with Lewis (2004), it was confirmed that loyalty programs increased the frequency of visits to the member establishment, thus having a very positive effect on it by: (a) offering greater opportunities for cross-selling; (b) encouraging personal communication and the customization and handling of the offer as opposed to massmedia communication; and (c) diminishing the probability of the customer visiting competing establishments. The company managing the program also benefitted from this increase in the number of visits to one of its member establishments since this, in turn, increased the brand awareness of the program in the minds of its members: acting as a reminder of its possible use in other member establishments and informing them about their accumulated points.

However, joining the program did not have the hoped-for effects on the rest of the buying behavior variables. These results were along the same lines as other previous research that limited the benefits of loyalty programs in terms of loyalty (Dowling and Uncles, 1997; García Gómez et al., 2006) and even in promotional terms (Dorotic et al., 2011).

The existence of differences in the profile of those customers who most changed their buying behavior after joining the program was also demonstrated. Within this group of customers, recency was the determinant variable, as those customers whose time-lapse since their last purchase was greatest were those who most changed their buying behavior. That is to say, these were the customers that reacted most favorably to joining the program. Therefore, integrating itself into the program allowed the joining establishment to awaken those customers described as "sleepers." Nevertheless, it should be pointed out that these customers were not necessarily the most loyal, but rather simply the most sensitive to the activities of the program.

Another variable that defined the profile of those who most changed their buying behavior was that of the greatest average annual expenditure on the primary product of the business before joining the program. The response to joining together with other products, less relevant to the establishment, produced a lesser variation in buying behavior, thus reflecting the importance of analyzing those customers that buy the products making up the core business of the company.

Apart from this, it was verified that those customers who joined the program in the analyzed establishment were not those who most modified their buying behavior after joining, which once again implied that loyalty programs were effective in activating (or re-activating) customers, but not in making them loyal.

It was also confirmed that the category of products sold, in this case less frequent, medium-priced consumer goods, obtained worse results than other types of products more typical of research focused on loyalty programs (flights, food, etc.). For this reason, the results of this work have provided important business decision-taking information for the different stakeholders involved—the program management and the member companies. This information concerns the type of company that should join the program and the need to simultaneously coordinate program activities with those specific to each member brand in order to improve the results.

#### Managerial Implications

An important decision to be taken by companies, with regard to the incorporation of a loyalty program into their strategies, is to choose the type of program to which they adhere. For many companies, the decision to join a particular multi-vendor platform is based on management complexity and the higher costs associated with mono-sponsor programs. However, in view of the results of this work, companies wishing to optimize the benefits offered by these programs should assess their limited effects on behavioral loyalty and consider that their advantages will be conditioned by the type of company that is associated with them, as well as the development of a management style that favors the company's own objectives against those of all the other companies promoting the same multi-vendor program.

It is important to point out that the company managing the program should not only include associate companies from different sectors in the program, given that behavior in one sector does not affect behavior in the rest, but it should also, preferably, make sure that the services of these associated establishments are complementary. This would greatly improve effectiveness. Likewise, it has also been shown that the relative importance of each member company within the multi-vendor loyalty program is a determining factor in behavioral loyalty. The driving-force companies of the program will improve the behavioral loyalty of their customers to a greater extent than those companies of lesser importance within the program itself. Thus, the total accumulation rate in the program will only be an interesting segmentation variable for the companies that carry more weight within the program.

With regard to the benefits related to loyalty that have been shown, the increased frequency of visits to related businesses has very positive effects for the management of the establishment, to the extent that:


The fact that multi-vendor programs represent a very useful marketing tool which permits the management of a heterogeneous client base is also verified. From an operational point of view, the ability to identify the profiles of customers who are more sensitive to program actions subsequent to joining, highlights the need to carry out differentiated relationship activities. However, the process of customer segmentation should be made independently by each of the associated companies and not by the company managing the program. This recommendation, which is derived from the results obtained jointly in the program, does not correspond to the behavior seen in each of the participating establishments as previously explained. In this regard, it is recommended that those responsible for the companies that are associated with a multi-vendor program manage information proactively, adapting it to their needs. This implies a greater commercial training of managers and employees in contact with customers, in addition to building a good relationship with the program promoters, which will help the program to adapt, as well as develop specific targets for each establishment based on their situation and the developments in their respective markets. This has obvious implications for companies that choose to adopt multi-vendor programs in order to reduce their marketing costs. For example, a company associated with these programs does not avoid—if they wish to optimize their benefits—having to realize their own communication actions since the generic communication program integrates all associated brands, making it less effective. Communication activities of the program are designed to benefit the whole program and its objectives do not necessarily match the conditions that each of the companies taking part in the program require. Since communications made by the program are not sufficiently segmented, each associated company must decide what their objectives are and what to communicate. Therefore, if an associated company is participating more actively in the management of the program

to suit their own particular needs, they will obtain better results, thus improving customer loyalty.

### Limitations and Further Research

The relationships between loyalty programs and behavioral outcomes are more complex than has been assumed. This empirical investigation had certain limitations: the effect of the loyalty program was only tested with regard to behavioral loyalty. The integration of attitudinal variables to complement the behavioral approach to loyalty would be a promising area of research.

It would likewise be interesting to continue to progress with the exploratory models used in this study, with a view to more precisely explaining changes in buying behavior after a customer joins a multi-sponsor loyalty program. Therefore, new variables should be incorporated related to purchases at competing establishments, such as share of wallet, a variable that, as seen in the literature review, has been included in numerous studies on behavioral loyalty. In other words, it should be determined whether the increase in the number of store visits is concentrated in the program and, therefore, decreases the number of visits to the competition or if, on the contrary, the frequency of visits increases under the program but remains unchanged with regard to the rest of the stores. This would mean that, although the program would improve purchasing behavior at the associated establishment, high levels of customer loyalty would not be achieved, as customers would continue to divide their purchases between the different competing brands.

REFERENCES


Additionally, future studies should address segmentation in loyalty programs based on customer lifetime value. As proposed by Kumar and Shah (2004), it would be interesting to measure customers' patronage behavior, taking into account the expected future value and not just past data.

Finally, given that the data analyzed correspond to a single multi-sponsor loyalty program (as in other studies conducted at the international level), and even though the program's representativeness is justified as there is only one other program with the same characteristics in Spain, it would be advisable to compare the results with those of other international programs.

### AUTHOR CONTRIBUTIONS

The directors of research have been the professors PR and ER. The three co-authors have participated in all stages of work, including the conception and design of the research, the revision of intellectual content, and drafting the work. The analysis and interpretation of data has been the responsibility of the professor TV.

### ACKNOWLEDGMENTS

This work has been funded by The Ministry of Economy and Competitivity (Spain), Resarch Project with reference: ECO2014- 59688-R, Programa Estatal de Investigación, Desarrollo e Innovación Orientada a los Retos de la Sociedad, Plan Estatal de Investigación Científica y Técnica y de Innovación 2013–2016.


PSM (2013). Loyalty Monitor. Madrid: PSM.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Villacé-Molinero, Reinares-Lara and Reinares-Lara. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Assessing the Growth of Ethical Banking: Some Evidence from Spanish Customers

Fernando E. Callejas-Albiñana<sup>1</sup> \*, Isabel Martínez-Rodríguez<sup>1</sup> , Ana I. Callejas-Albiñana<sup>2</sup> and Irene M. de Vidales-Carrasco<sup>3</sup>

<sup>1</sup> Department of Spanish and International Economy, Econometrics, History, and Economic Institutions, University of Castilla-La Mancha, Ciudad Real, Spain, <sup>2</sup> Department of Psychology, University of Castilla-La Mancha, Ciudad Real, Spain, <sup>3</sup> Department of Political Economy and Public Finance, Economic Statistics, Business, and Economic Policy, University of Castilla-La Mancha, Ciudad Real, Spain

#### Edited by:

Alicia Izquierdo-Yusta, University of Burgos, Spain

#### Reviewed by:

Antonio Calvo, CEU San Pablo University, Spain Javier Iturrioz, Centro de Estudios Universitarios, Mexico

> \*Correspondence: Fernando E. Callejas-Albiñana fernando.callejas@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 08 March 2017 Accepted: 27 April 2017 Published: 24 May 2017

#### Citation:

Callejas-Albiñana FE, Martínez-Rodríguez I, Callejas-Albiñana AI and de Vidales-Carrasco IM (2017) Assessing the Growth of Ethical Banking: Some Evidence from Spanish Customers. Front. Psychol. 8:782. doi: 10.3389/fpsyg.2017.00782 Aristotle, who, having predated Adam Smith by 2000 years, deserves to be recognized as the world's first economist (Solomon, 1995), distinguished between two different senses of what we call economics: oikonomikos, or household trading, which he approved of and considered essential to the working of any even slightly complex society, and chrematisike, or trade for profit, which he considered selfish and utterly devoid of virtue, calling those who engaged in such practices "parasites". Of course, consumers do not purchase and invest for solely economic reasons (Polanyi, 1944). Interest in ethics in economics has been the subject of continuous study. In this regard, the recent financial crisis has had not only economic, but also social, psychological, political, and ethical consequences, which have impacted the financial and banking system. Consumers are no longer drawn only by the economic return but also by ethical factors. Ethical banking is on the rise. This paper aims to explain the reasons for the growth in ethical banking and to answer the following questions: can banking consumers-investors change the characterization of the banking system? Can ethical banking gain ground on traditional banking? And is ethical banking really effective? To this end, it will examine the Spanish case, using econometric causal regression models to identify the reasons why consumers decide to invest in ethical banking and determine its role in the Spanish economy.

Keywords: ethical banking, traditional banking, banking consumers, deposits and loans

## INTRODUCTION

Ethics in finance has long been a subject of study, and today it is being given increasing attention. It seeks to incorporate moral, transparent, fair, and sustainable considerations into the decisions of an ever-growing number of banking consumers. The Goldman Rule that defends the pursuit of profitable opportunities, regardless of the effects on others (Watkins, 2011) appears to be disappearing in general consumption and (albeit to a lesser degree) in the consumption of financial products too.

Formally, ethical banking first emerged in its present-day role as a trader with the founding of the first ethical bank, Triodos Bank, in 1971, following a 1968 study group meeting in Holland.

They sought a way to promote a more rational and conscientious way of using money and, finally in 1971, they set up the Stichting Triodos Foundation (Triodos Bank, 2008–2015:149).

On the supply side these banks take ethical responsibility into account when determining which products – financial or otherwise – to offer consumers, thereby supplying the market with a responsible offer based on trust in those who receive and manage consumers' savings (Green, 1989). This means companies are ethically liable and are not protected by limited liability from the consequences of their actions. The role of bankers is to administer funds based on the trust of those who ask them to administer their money, and they should thus loan this money responsibly.

From this perspective, the severe financial crisis, which began in the summer of 2007 in the U.S. real estate sector and continues, even today, to affect the real economy of many developed countries, had not only economic causes, but also psychological, social, political, and ethical ones (Fernández, 1994:402). The latter include a number of ethical failures (Calvo and Mingorance, 2012) that both led to the crisis and helped to make it as deep, severe, and enduring as it has been. These failures can be divided into three distinct groups (Argandoña, 2012:2). First, there were individual moral failures, as witnessed by the widespread inappropriate, sometimes even criminal, behavior so often on display in the pre-crisis environment of high economic growth, abundant liquidity, low interest rates, and extraordinary opportunities to profit: concealment of information, false advertising, the proliferation of unnecessary transactions to generate higher commissions, inadequate risk assessments, high indebtedness, etc. Second, there were ethical failures related to management or governance, such as the many cases of poor governance and lack of professional competence by directors, senior executives, and analysts at organizations ranging from commercial and investment banks to hedge funds (Ropero and Hurtado, 2015), monolines, rating agencies, supervisory bodies, central banks, and governments. Finally, third, there were social ethics failures, since social conditions were created that encouraged, or at least failed to stop, inappropriate behavior at the personal and organizational levels that notably interfered with the proper functioning of the legal, institutional, and social corrective mechanisms that, in other conditions, would have prevented the moral consequences of such ethically reprehensible behavior.

Clearly, the recent systemic crisis has jeopardized global financial stability by disrupting the efficient allocation of savings and investment – the drivers of economic growth, job creation, progress, and social welfare – and, ultimately, threatening continued public trust in the financial system. Indeed, in recent years, financial systems, including Spain's, have largely distanced themselves from their traditional role of meeting the real economy's needs. The severe financial crisis generated a high degree of uncertainty and volatility in the international markets due, first, to the liquidity and solvency problems that many institutions were experiencing. The situation was subsequently compounded by deep economic recessions and sovereign debt crises resulting from the significant debt governments incurred as a consequence of the implementation of fiscal stimuli and measures to support distressed banks in an environment of high external leverage of the private sector (Argandoña, 2012).

In short, the economic and financial crisis has underscored the need to embrace an investment philosophy based on greater transparency, a greater presence of ethical values, and better and broader risk management. Numerous intellectuals have called for the necessary proliferation of ethical and moral principles in the economic and financial world. However, these principles do not arise spontaneously, but rather in response to the problems generated by the magnitude and frequency of business scandals and the various economic and financial crises we have witnessed in recent years, for which there has never been a politically and socially acceptable solution from a neoliberal perspective (Calvo, 2012:7). The development of an ethical culture in the financial world has resulted in the establishment of new national and international regulatory standards that help to strengthen the regulation, supervision, and risk management of the global financial system, while at the same time eliminating or reducing the harmful effects of ethical conflicts on society as a whole.

### MATERIALS AND METHODS

### Conceptual Framework

The implementation of new national regulatory standards in the financial system, with greater international coordination, has primarily sought to achieve (Aspachs-Bracons et al., 2010:5): greater and better capitalization of banks, more stringent control of liquidity and solvency levels, increased transparency in financial agreements and securitizations, and greater control of risk and the level of leverage assumed by the industry. At the same time, it has revealed the need to establish a new multilateral pillar in order to strengthen the regulation, supervision, and risk management of the international financial system, thereby preventing new systemic crises and mitigating their global impact should they occur.

#### The Current Traditional Banking System

In response to the aforementioned guidelines, the Basel Committee on Banking Supervision developed a comprehensive set of reforms known as Basel III with a view to improving the banking industry's ability to absorb shocks caused by economic or financial stress, whatever the source, improving banks' risk management and governance, and strengthening banks' transparency and disclosures (Rodríguez de Codes Elorriaga, 2010; Bank of International Settlements, 2016). As the new capital and liquidity measures significantly tighten bank regulation, a broad transitional period was established, to last from January 1, 2013, to January 1, 2019, so that the changes could be implemented gradually. These measures were part of a broader reform process, arising from the action plan that the G20 approved at the Washington Summit held on November 14 and 15, 2008, to provide global solutions to the crisis and improve international cooperation.

Subsequently, at the G20's London Summit, held on April 2, 2009 (El País, 2009), in addition to a broad package of measures aimed at restoring growth and employment, strengthening financial institutions, and promoting global trade and investment, the group agreed to establish a new Financial Stability Board (FSB). The FSB would succeed the Financial Stability Forum (FSF) and would consist of the G20 countries, Spain, and the European Commission. Its actions, which it performs in close collaboration with the IMF and the Basel Committee, are mainly aimed at maintaining global financial stability and the openness and transparency of the financial sector. To this end, its members commit to implementing international financial standards and undergoing periodic reviews of their compliance with these obligations (Banco de España, 2016).

Meanwhile, at the European Union level, two new institutions were created (Field and Moreno García, 2010:114). The first was the European Systemic Risk Board (ESRB), an independent body responsible for the macroprudential supervision of the Union's financial system, which aims to identify and prioritize systemic risks and, where applicable, issue recommendations for action and monitor their implementation. The second was the European System of Financial Supervision (ESFS), tasked with microprudential supervision, i.e., the supervision of individual institutions. The ESFS is made up of the national financial supervisors and three new European supervision authorities: the European Banking Authority (EBA), the European Securities and Markets Authority (ESMA), and the European Insurance and Occupational Pensions Authority (EIOPA). These European authorities were set up based on existing European committees in order to harmonize standards, promote the application of European legislation, and, when necessary, help settle disputes between national supervisors.

Finally, in the summer of 2012, the leaders of the European Union undertook to advance on the creation of the European Banking Union (EBU). The EBU first emerged as a step toward financial integration, i.e., toward a single market for financial services and, ultimately, toward the perfection of the construction of the euro by re-establishing a properly functioning monetary policy in the eurozone and restoring confidence in the European banking sector (Abascal Rojo et al., 2014). To achieve these goals of greater and better use of financing sources and higher levels of competition, efficiency, technology, and diversification, the EBU is organized around three main pillars, to which a fourth might be added:


difficulties with minimal costs to taxpayers and the real economy.


However, this financial regulation, at both the national and supranational level, has generally proved insufficient to prevent the numerous ethical conflicts that arise in the world of finance. This is mainly due to the great complexity of financial activities, which facilitates avoidance of the undertaken obligations. Consequently, the various players in these markets must also engage in self-regulation. This can be pursued, first, within the financial institutions themselves, through the establishment of codes of ethics or of conduct, which have enabled significant improvements in institutional transparency, and through a commitment to corporate social responsibility, with the aim of managing the impact these institutions' activities have on their customers, employees, shareholders, local communities, the environment, and society at large. Second, it can be achieved through appropriate financial training and education that equips the recipients of the training with the knowledge, skills, behavior, values, and aptitudes they need to make sensible and informed financial decisions, in addition to leaving them better positioned to deal with the basic financial challenges they will encounter over their lives. Indeed, to paraphrase Melé, "Let us improve regulation and its functioning, but let us also consider improving people's culture and education. This latter may be more expensive, but it is also more enduring" (Melé,, 2009:64).

#### Current Ethical Banking

Recent years have seen a proliferation of new forms of financial business that seek to make economic profitability compatible with respect for human rights and the environment. One such business is ethical banking, which invests only in projects offering value added to society, mainly from an educational, cultural, environmental, and/or social perspective. In this context, ethics is understood as a science aimed at steering human action in a rational direction (Cortina, 1994:17). Many papers have looked at ethical banking and demonstrated the important role it plays as an independent, differentiated financing activity (Lynch, 1991; Cowton and Thompson, 2001; Alsina, 2002; Fergeson, 2004; Barbu and Vintila, 2007; Buttle, 2008; Baranes, 2009; Cowton, 2010).

In the economic and financial arena, ethics must fulfill the important mission of ensuring equal access by agents to all types of goods and services. To achieve this, the financial

system must promote the achievement of monetary and financial stability and effectively contribute to it, in accordance with its purpose of optimizing the use of financial resources, by channeling the savings generated by surplus spending units to borrowers or deficit spending units (Calvo and Martín de Vidales, 2014; Calvo et al., 2014:1). In short, it must facilitate real or productive economic activity and foster overall rather than just individual well-being, which is the first principle of financial ethics (Camacho Laraña, 1996:35).

To date, the concern for ethics in the financial system has led to two main developments: the establishment of ethical banks and the launch of ethical or socially responsible investment (SRI) funds, that is, funds in which certain social values (usually related to environmental quality and quality of life) are given precedence over strictly financial ones (risk-return). The ethical banking model could be included in the market for SRI funds. Ethical banks manage their customers' money, allocating it to investments and projects based on environmental, social, and governance (ESG) criteria with the aim of creating something of social utility for the surrounding community, above and beyond the mere pursuit of profit. This social purpose refers to the social return on the invested capital and to the social responsibility of the investor (Spansif, 2013:6; Benito Hernández, 2014).

Europe has had several experiences with ethical banks, which have positioned themselves at the epicenter of some of the Old Continent's largest economies, especially in innovative sectors with high growth potential. Some have joined the Global Alliance for Banking on Values (GABV), established in 2009 with a view to examining the financial dynamics of capital and the systems for measuring the impact on the overall, environmental, and financial development of sustainable banks. In addition, the GABV established, for the first time, a precise definition of sustainable banking, endorsing the following principles (Global Alliance for Banking on Values [GABV], 2012):


Recognition of the institution as a bank or credit institution by the national authorities is essential (Iturrioz et al., 2005; Valor, 2005; Valor et al., 2007). This dimension is important to distinguish between ethical banks and other financial experiences, such as solidarity programs or foundations that depend on banks but do not operate as real financial institutions. Some traditional banks have foundations that might in and of themselves meet ethical criteria, but are not, strictly speaking, credit institutions, since they depend on the bank's business, which probably has a different kind of social impact. Ethical commitments must thus affect all aspects of the bank and not just part of it or its activities (San-Jose et al., 2011:152–153).

### Why Do Consumers Choose Ethical Banking? An Analysis of the Spanish Case

In Spain, investing according to ethical, sustainability, and/or social responsibility criteria remains considerably less developed than in many northern European countries, despite having become, in recent years, one of the levers for advancing toward a more sustainable economic model able to overcome the effects of the crisis. Indeed, this form of investment remains a niche type of investment dominated by a small number of large institutional investors, which together account for 97% of the assets managed according to ESG criteria (Novaster, 2015:3). Of these, the most active participants are employment pension plans, which are the most important drivers of SRI in Spain, as SRI retail funds have only a very marginal presence, due both to the sparse selection currently offered by the main financial institutions and a lack of knowledge on the part of private investors.

Accordingly, no ethical banks have yet been founded in the Spanish financial system. Therefore, the influence of this type of ethical action can be said to be limited to just three fronts. The first is the launch by some traditional banks of financing lines based on the granting of microcredits, mainly targeted at entrepreneurs, personal and family development projects, microenterprises seeking to meet social needs, and ecologically sustainable or environmentally friendly projects. CaixaBank, BBVA, and the ICO have been quite active in this business segment for many years. The second is the opening in Spain of branches of different ethical banks, specifically, Triodos Bank and Fiare Banca Etica. Finally, the third is the marketing by organizations that are not financial institutions – mainly, service cooperatives – of savings products and ethical financing. In this sphere, the initiatives undertaken by Oikocredit and Coop57 stand out. The most important aspects of the latter four institutions are explained below.

(1) Triodos Bank. Triodos Bank was established in the Netherlands in 1980, following the creation of the Triodos Foundation in 1971, as a way to channel donations and loans to companies excluded from the financial sector. Since then, Triodos has grown and expanded its original approach. Through its extensive European branch network

(it currently also has operations in Belgium, the UK, Spain, and Germany), it finances companies, institutions, and projects that add clear social, environmental, and cultural value with the support of depositors and investors who wish to encourage socially responsible businesses and organizations and a sustainable society. To this end, its business model seeks to strike a balance between people's quality of life, care of the planet, and economic profit, what is known as the triple bottom line approach, or the 'three Ps' (People, Planet, Profitability) (Triodos Bank, 2008–2015).


these workers used part of the compensation they had received in the lay-off to create a fund to promote economic projects aimed at creating quality jobs, especially through the application of cooperative models. With this fund, Coop57 was founded, on June 19, 1995. Although at first it was closely linked to associationist cooperativism, it gradually expanded its social base to include other types of social and solidarity economy organizations. Indeed, in 2005, Coop57 adopted a networked growth strategy throughout Spain based on the interest that it had stirred in other regions of the country, which used the business model and legal and technical structure of the parent company. Currently, in addition to its headquarters in Barcelona, it has regional sections with their own management and participation structures in Aragon (since 2005), Madrid (since 2006), Andalusia (since 2008), and Galicia (since 2009). Its main goal is to collect people's savings in order to channel them toward the financing of social and solidarity economy organizations that promote stable quality employment, foster cooperativism, associationism, and solidarity in general, and promote sustainability and food and energy sovereignty, based on ethical and charitable principles. In short, although Coop57 carries out a financial activity, it does so not in pursuit of an economic goal, but rather a social one: the financing of projects that add value to their communities and society as a whole.

Although the presence of ethical banking operations in Spain remains fairly small, numerous experts have argued that the silent revolution of ethical, social, and charitable finance has begun (Sanchis, 2016) and that it is having a positive impact on growth.

### RESULTS

First of all, it is important to consider that the lack of homogeneous data sources for this sector, statistical gaps, and the absence of a uniform nomenclature for both the activities themselves and the purpose thereof at the analyzed institutions (Coop57, 2007–2015; Fiare Banca Ética, 2008–2015; Triodos Bank, 2008–2015; Oikocredit, 2014–2015a,b) require an initial approximation of the sector in order to measure the impact of this banking model on the Spanish economic system (Supplementary Data Sheet 1). For the empirical research, the whole econometric models were technically validated, and all structural and random hypotheses were supported.

The first aim of the research was to measure the development of ethical banking in the Spanish financial sector compared to traditional banking. To this end, all ethical banking in Spain was represented through the analysis of two variables, which were selected on the basis of relevance and ease of comparison with traditional banking: deposits and loans. In order to ensure conceptual consistency and enable the comparison of ethical banking and traditional banking, the heading "deposits" includes contributions that can be used to fund grants and projects (from own funds to deposits and lending investments), while the

heading "loans" is the sum of allocated amounts (credits, loans, social projects, etc.).

**Figure 1** shows how, following the onset of the financial crisis in 2007 and the subsequent reform of the Spanish banking system, loans declined faster than managed deposits in the traditional banking segment. In contrast, in the ethical banking segment, growth in loans outstripped deposit growth until 2012 (see **Figure 2**). Since then, the trend has reversed, with loans showing only modest linear growth in contrast with soaring deposits.

**Figure 3** shows the 2011–2015 series. As in prior years these indices would yield very high rates, making

TABLE 1 | Econometric model explaining the evolution of loans granted by ethical banks.

Dependent variable: LEB Method: Least squares Date: 09/19/16 Time: 11:55 Sample: 2000 2015 Included observations: 16


it impossible to compare the two ratios, so in order to solve this problem, an index number were calculated. A comparative study of deposits and loans shows that, in traditional banking, the two headings have behaved roughly the same. In contrast, ethical banking has grown faster, with growth in customer deposits being particularly strong.

Finally, a comparative ratio of loans and deposits for both types of banks, reflecting both of the preceding points, since 2007 shows that the banking system in general has become more conservative (see **Figure 4**). In traditional banking, not all existing available funds are allocated to loans. The same has been true in ethical banking since 2012, but in a much more pronounced way.

Both for the comparative analysis and to illustrate the change in behavior that has led consumers to trust ethical banks, several simple but fairly representative econometric models were considered.

First, a model was proposed to analyze the recent evolution of ethical banking, taking consumer behavior in the sector into account (**Table 1**).

In other words, the economic causes that lead consumersinvestors to increase the volume of ethical banking were proposed empirically.

In addition to transparency and the social reasons discussed above, certain economic and financial factors have significantly helped to drive ethical banking business in recent years and could serve as an initial hypothesis.

TABLE 2 | Econometric model of inputs (capital and labor) in the Spanish GDP.

Dependent variable: GDP Method: Least squares Date: 09/22/16 Time: 13:21 Sample: 2000 2015 Included observations: 16



likewise began to decline beginning in 2007, making it less attractive to investors.


FIGURE 6 | Fit of inputs (capital and labor) with the evolution of the Spanish GDP.

In light of the above, the paradigm was obtained with the following econometric model (see **Table 1** and **Figure 5**), where:

LEB: Loans granted by ethical banks (€ thousand, current prices).

DEB: Deposits at ethical banks (€ thousand, current prices).

EUR: EURIBOR.

SI: Stock index.

F06: Dummy variable that corrects a technical problem in the structural change in the model.

$$\begin{aligned} \text{Specification 1:} \quad \widehat{LEB} &= \widehat{\beta}\_0 + (\widehat{\beta}\_1 \* DEB) + (\widehat{\beta}\_2 \* EUR) \\ &+ (\widehat{\beta}\_3 \* SI) + (\widehat{\beta}\_4 \* F0\widehat{e}) \end{aligned}$$

Estimate 1: LEB <sup>d</sup> <sup>=</sup> <sup>430391</sup>.<sup>7</sup> <sup>+</sup> (0.<sup>232091</sup> <sup>∗</sup> DEB) <sup>−</sup> (75139d.<sup>18</sup> <sup>∗</sup> EUR) <sup>−</sup> (115.<sup>0631</sup> <sup>∗</sup> SI) <sup>+</sup> (182388.<sup>5</sup> <sup>∗</sup> <sup>F</sup>06)

In conclusion, ethical banking provides financing based on the level of deposits (direct relationship). At the same time, the model (with a fit of 98.38%) shows that the situation of declining interest rates and a falling stock index justified the increase in these loans, subject to deposit levels.

The next question is whether this consumer choice is favorable or unfavorable to the country's economic system. The following model of the banking sector is revealing in this regard. It shows the importance of ethical banking and traditional banking to GDP, including both the capital input (through deposits at traditional and ethical banks) and the labor input (from the complementary perspective of the unemployment rate or UR). Considering the justification of the proposed specification, as a function of aggregate production (GDP), and the inputs of capital (LEB, LTB) and labor (UR), yields the following econometric model (see **Table 2** and **Figure 6**):

$$\begin{aligned} \text{Specification 2:} \quad \widehat{GDP} &= \begin{array}{c} \widehat{\gamma}\_0 \ + \ \left(\widehat{\gamma}\_1 \* \widehat{LEB}\right) &+ \ \left(\widehat{\gamma}\_2 \* LTB\right) \\ &+ \ \left(\widehat{\gamma}\_3 \* UR\right) &+ \ \left(\widehat{\gamma}\_4 \* F08\right) \end{array} \end{aligned}$$

Estimate 2: GDP <sup>d</sup> <sup>=</sup> <sup>776319</sup>.<sup>6</sup> <sup>+</sup> (0.<sup>360935</sup> <sup>∗</sup> LEB) + (175.4056 ∗ LTB) − (19138.02 ∗ UR) + (65824.71 ∗ F08)

Where:

GDP: Gross Domestic Product (€ million, current prices).


The model conforms to economic theory and the signs are consistent with macroeconomic performance. The levels of loans granted by both ethical and traditional banks (capital input) positively impacted the growth of the Spanish economy. As expected, the unemployment rate (labor input) negatively impacted growth, i.e., the higher the unemployment rate, the lower the economic growth.

### DISCUSSION

Ethical consumption is a growing movement, and consumers are increasingly committed to ethical factors when forming opinions about products and making purchase decisions (Shaw and Shui, 2002; De Pelsmacker et al., 2005). However, several studies have revealed the existence of a possible gap between purchase intentions and actual purchasing behavior (ethical

purchasing gap), showing that ethically conscientious consumers rarely shop ethically (Bray et al., 2011; Goyal and Joshi, 2011; Papaoikonomou et al., 2011; Carrington et al., 2014).

The ethical issues involved in banking practices have become an important aspect that the nations of the world must take into consideration, especially in light of the incessant bank failures and ensuing loss of customers' deposits (Asikhia, 2016). However, the ethical banking sector has grown, not only as a result of the bad banking and financial practices of recent years, but also because of the confluence of the economic crisis, banking reform, declining returns for investors through traditional channels, and, especially, the ethical and social commitment that has recently begun to emerge in many activities.

Based on the preceding empirical analysis, we can offer an initial response to the questions posed at the start of this paper.

#### 1. Can bank consumers-investors change the characterization of the banking system?

By standardizing the coefficients<sup>1</sup> of the first estimated model (**Table 1**), we can see the unit effect of each explanatory variable (deposits at ethical banks, Euribor, and stock index) on the endogenous variable (loans granted by ethical banks), giving:

Although deposits at ethical banks (βˆ EBD<sup>∗</sup> = 0.41055) explain the behavior of ethical banking loans to a greater degree than either the Euribor (βˆ EUR<sup>∗</sup> = −0.4016) or the stock index does alone, they do so in a smaller proportion and with the inverse sign compared to the Euribor and stock index taken together (βˆ SI<sup>∗</sup> = −0.09807). Therefore, when both the Euribor and the stock index increase, deposits at ethical banks will decrease due to the greater profitability of deposits at traditional banks, resulting in a decline in ethical banking loans at high values.

In short, it can be concluded that, if the Euribor continues to fall as it has to date, banking consumers-investors could indeed change the characterization of the banking system by choosing the investments offered by ethical banks through deposits.

#### 2. Can ethical banking gain ground on traditional banking? Is it really effective?

Interpreting the results of the second estimated model (**Table 2**) to determine the extent to which each variable (loans granted by ethical banks, loans granted by traditional banks, and the unemployment rate) contributes to the evolution of GDP, and again standardizing the coefficients, yields:

In this case, the unemployment rate (βˆUR<sup>∗</sup> = −0.794931) explains GDP behavior to a greater extent than loans, whether granted by ethical or traditional banks (βˆ LEE<sup>∗</sup> <sup>=</sup> 0.7791, <sup>β</sup><sup>ˆ</sup> LTB<sup>∗</sup> = 0.69925).

However, when the impact of loans is considered in isolation, loans granted by ethical banks explain GDP behavior to a greater degree, although the difference is not huge (11.42%). This means that an increase in ethical banking loans will be more effective for GDP growth than an increase in traditional banking loans, as long as the Euribor and stock index hold steady or decline (conclusion obtained and discussed from **Table 1**).

In sum, if the Euribor continues to register low values, as it has to date, ethical banking could gain ground on traditional banking, as it is more effective at increasing GDP.

#### 3. Are the principles of ethical banking fulfilled?

**Figures 1**–**5** could cast doubt on whether ethical banking faithfully fulfills some of its basic principles in terms of its "social and environmental bottom lines" and operating "at the service of the real economy".

The foregoing analysis showed that deposits managed by ethical banks are growing at a far higher rate than the loans they grant. At first glance, this finding poses a twofold question regarding the sector's future viability and whether it may not, in the long term, conform more closely to the traditional banking business model.

However, the existing gap between deposits and loans may be justified if ethical banks are allocating the financing collected through deposits to corporate purposes that they share with traditional banks (e.g., hiring staff, the financing of new physical branches of institutions in this business segment, etc.). In that case, it would make realistic economic sense, as a sector that creates jobs and capital input, thereby generating more efficient and sustained economic growth. It must be remembered that the primary labor of this type of banking is the financing, through financial intermediation, of social and solidarity economy projects that promote, among other things, the creation of stable, quality employment, sustainable development, food and energy sovereignty, the social inclusion and employment of vulnerable groups, international cooperation, fair trade, culture, education, and civic engagement.

Therefore, in addition to increasing the economic bottom line, as traditional banking does, by creating jobs and wealth through the acquisition or leasing of office space, it would continue to ensure the sustainability and social purposes that characterize its ideology, resulting in an improvement in people's quality of life and better use of the available environmental resources.

However, any conclusion regarding the consistency of ethical banking will need to await a time when both the Euribor and stock index are rising.

<sup>1</sup>βˆ<sup>∗</sup> (Xj) = β(ˆ Xj) • (SDX<sup>j</sup> SDY); where

<sup>–</sup> SD is the standard deviation of the variable.

<sup>–</sup> X: explanatory or exogenous variable, and

<sup>–</sup> Y: variable to be explained or endogenous variable.

### AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication.

### ACKNOWLEDGMENTS

This work was funded by the Spanish Ministry of Economy and Competitiveness through Research Project ECO2014- 59688-R, as part of the National Program for Research, Development, and Innovation Oriented toward Societal

#### REFERENCES


Cortina, A. (1994). Ética de la Empresa. Madrid: Trotta.

Challenges, under the 2013–2016 National Research, Technology, and Innovation Plan and besides was funded by financial assistance to Research Group of Commercial Observatory of Innovation in Distribution by University of Castilla-La Mancha.

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fpsyg. 2017.00782/full#supplementary-material



**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Callejas-Albiñana, Martínez-Rodríguez, Callejas-Albiñana and de Vidales-Carrasco. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Investor Behavior and Flow-through Capability in the US Stock Market

#### Carlos Cano<sup>1</sup> , Francisco Jareño<sup>1</sup> \* and Marta Tolentino<sup>2</sup>

<sup>1</sup> Department of Economic Analysis and Finance, University of Castilla-La Mancha, Albacete, Spain, <sup>2</sup> Department of Economic Analysis and Finance, University of Castilla-La Mancha, Ciudad Real, Spain

This paper analyzes investor behavior depending on the flow-through capability (FTC) in the US stock market, because investors seek protection from inflation rate changes, and the FTC (a firm's ability to transmit inflation shocks to the prices of its products and services) is a key factor in investment decisions. Our estimates of the FTC of firms listed on the US stock exchange at the sector level are significantly different among industries, and we demonstrate a direct relationship between changes in stock prices (at the sector level) and FTC. These results would be relevant because they have important implications on investor behavior.

Keywords: flow-through capability, inflation rate, stock return, sectoral analysis, investor behavior

### INTRODUCTION

#### Edited by:

Alicia Izquierdo-Yusta, University of Burgos, Spain

#### Reviewed by:

Jorge Pelegrín-Borondo, Universidad de La Rioja, Spain Inés González, Universidad Pública de Navarra, Spain

> \*Correspondence: Francisco Jareño francisco.jareno@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 09 March 2016 Accepted: 21 April 2016 Published: 09 May 2016

#### Citation:

Cano C, Jareño F and Tolentino M (2016) Investor Behavior and Flow-through Capability in the US Stock Market. Front. Psychol. 7:668. doi: 10.3389/fpsyg.2016.00668 Campbell (2006), Sekscinska (2015), and González et al. (2016), among others, point out that investor behavior is concerned about managing the economic and financial risk. To that end, a measure of the firm's ability to transmit inflation shocks to the prices of its products and services could be really relevant for investors and portfolio managers. Thus, the main objective of this research is to estimate the capability of American companies to transmit inflation shocks to the prices of the products that they sell and/or the services that they provide (Asikoglu and Ercan, 1992; Jareño and Navarro, 2010) and to identify any significant differences among sectors, because the investor behavior may be quite different depending on this flow-through capability.

Flow-through capability (FTC) is defined as a firm's ability to transmit inflation shocks to the prices of its products and services. The concept of "FTC" was introduced by Estep and Hanson (1980), but Asikoglu and Johson (1986, 1990) and Asikoglu and Ercan (1992) were the pioneers in analyzing the negative relationship between inflation and stock returns in the United States at the sector level. The authors conclude that the negative effect of an increase in the inflation rate on the stock price of a company is inversely related to the company's capability to transmit inflation shocks to its prices.

Empirical evidence (Jareño, 2005; Jareño and Navarro, 2010) suggests that FTC has an effect on stock prices and that there are significant differences in FTC at the sector level. Namely, industries in which the flow-through (FT) coefficients are greater exhibit stock prices that are less sensitive to inflation shocks. Therefore, increments in FT capability are linked to increments in stock prices.

Jareño and Navarro (2010), studying the Spanish stock market, obtain evidence of a strong negative relationship between the sensitivity of stock returns to changes in nominal interest rates and the ability to absorb inflation. Specifically, absorption ability can explain ∼50% of the differences in stocks' sector durations in the face of changes to nominal interest rates.

Other studies, such as Ertek (2009), are based on the concept of FT capability and conclude that the higher the percentage of inflation shocks that translate into growth in a company's profit rate, the greater the company's stock price (and vice versa). In addition, Ertek (2009) incorporates FT capability into a quantitative model for stock selection to create quality portfolios with inflation hedge in the European sphere. These portfolios are built by selecting individual companies based on a series of characteristics believed to affect returns provided by different assets, considered to be linked to inflation hedge.

Given that related literature is scarce, we highlight this study's contributions. First, we use quarterly data, as opposed to the semiannual data used in previous studies. Second, we propose an alternative method of measuring the ability to absorb inflation, using a proxy variable that is different from the production level. Third, we confirm, for the sample analyzed, a positive relationship between changes in stock prices and companies' FT capability at the sector level, in agreement with Asikoglu and Ercan (1992). Therefore, we find that investor behavior may be quite different according to the FTC of the sector of activity that each company belongs to.

The remainder of the study is structured as follows. Section Estimation of the FTC describes not only the data but also the methodology that was proposed and employed. Section The Relationship between FT Coefficients and Stock Prices shows how we estimated companies' ability to absorb inflation, as analyzed by sector in the US stock market. The primary results are gathered and interpreted in Section Overall Results. Finally, Section Discussion highlights this study's primary conclusions.

#### MATERIALS AND METHODS

This study assumes that the investor behavior may be different depending on the FTC, because investors want to protect from interest and inflation rate changes (González et al., 2016), so the FTC is a relevant factor in investment decisions. Moreover, according to Asikoglu and Ercan (1992) and Jareño (2005), among others, companies characterized by higher FT capability would show higher stock prices. As a result, in sectors with higher FT coefficients, stock prices would be less sensitive to inflation shocks.

Moreover, Kusev and van Schaik (2011), and Sekscinska (2015), among others, assume that financial and economic decisions, considered by previous investors as good or bad ones, may affect risk preferences, so prior experiences could determine their subsequent decisions. Blackburn et al. (2014) also find that investor decisions may depend on past returns. Therefore, the analysis of the FT capability would be really crucial for investors because it enables them to better manage interest and inflation risk.

Thus, this research is based on the methodology proposed by Jareño (2005) and Jareño and Navarro (2010) as applied to the United States, in order to estimate this relevant FT capability of each company depending on the sector belongs to. We used quarterly data, which is a marked improvement compared to previous studies that primarily used semiannual data. The study period is 2000 through 2009, and 40 observations were obtained.

In Jareño (2005), which the starting point of this research, the FT coefficients of the Spanish case are estimated, taking as a reference semiannual data from 150 companies traded in the stock market during the sample period, classified by activity sector (according to the classification established by the Madrid stock market).

This study begins by proposing the following theoretical model:

$$
\Delta V\_t = p\_{t+1} \cdot q\_{t+1} - p\_t \cdot q\_t \tag{1}
$$

where V<sup>t</sup> is company sales during period t, p<sup>t</sup> <sup>+</sup> <sup>1</sup> is the mean price of products sold by the company during period t+1 and q<sup>t</sup> <sup>+</sup> <sup>1</sup>is the mean production (in physical units) during period t+1.

However, data related to goods sold and/or services provided are not available, which is why we used a proxy variable, namely, the number of employees. According to Jareño and Navarro (2010), this variable allows a good approximation if we assume constant productivity. Because the number of years in the sample is not very high, we assume that hypothesis. One important contribution of this research is to propose an alternate proxy variable from that used by Jareño (2005) and Jareño and Navarro (2010); with that alternative proxy variable, it is possible to make a second estimation of the FT capability of North American companies at the sector level.

Substituting production in Equation (1) with its equation according to productivity and the number of employees, we obtain the following formula:

$$
\Delta V\_t = \Delta p\_t \left(\omega\_{t+1} \mathfrak{p} dd\_{t+1}\right) + p\_t \mathfrak{p} dd\_t \Delta \omega\_t + p\_t \omega\_{t+1} q\_t \Delta \mathfrak{p} dd\_t \quad (2)
$$

where ω<sup>t</sup> <sup>+</sup> <sup>1</sup> is the mean number of employees in t+1 and pdd<sup>t</sup> <sup>+</sup> <sup>1</sup> is the productivity for each employee. Assuming constant productivity, we would obtain the following equation:

$$\frac{\Delta V\_t}{V\_t} = \frac{\Delta p\_t}{p\_t} \left[ \frac{\alpha\_{t+1} - \alpha\_t}{\alpha\_t} + 1 \right] + \frac{\Delta \alpha\_t}{\alpha\_t} = \frac{\Delta p\_t}{p\_t} \frac{\Delta \alpha\_t}{\alpha\_t} + \frac{\Delta p\_t}{p\_t} \frac{\Delta \alpha\_t}{\alpha\_t} \tag{3}$$

The relative increase in company sales is equal to the sum of two terms and their cross product. If we assume that the latter is negligible, we obtain the following simplified equation:

$$\frac{\Delta V\_t}{V\_t} \approx \frac{\Delta p\_t}{p\_t} + \frac{\Delta \phi\_t}{\alpha\_t} \tag{4}$$

The annual growth of the sales variable is calculated as the quarterly payment with respect to the same payment for the previous year, thus avoiding seasonal problems.

$$\frac{\Delta V\_t^{t-4}}{V\_{t-4}} = \frac{V\_t - V\_{t-4}}{V\_{t-4}}\tag{5}$$

where V<sup>t</sup> is the business net turnover during the quarterly period "t".

Conversely, there is no data available for variablep<sup>t</sup> , and thus we assume that the increase in prices for the company's goods and services sold/provided have a linear relation to the economy's previous inflation period:

$$
\Delta p\_t = f\left(\triangle IPC\_t, \triangle IPC\_{t-1}, \dots\right) = \alpha\_0 \triangle IPC\_t + \alpha\_1 \triangle IPC\_{t-1} + \dots \tag{6}
$$

where IPC<sup>t</sup> is the US consumer price index in period t, and α0, α<sup>1</sup> are the FT coefficients (such as the Spanish coefficients estimated by Jareño, 2005).

Considering Equation (6) in (4), the general theoretical model proposed by Jareño (2005) and Jareño and Navarro (2010) is as follows:

$$\frac{\Delta V\_t^{t-4}}{V\_{t-4}} = \beta\_0 + \beta\_1 \frac{\Delta \phi\_t^{t-4}}{\alpha\_{t-4}} + \beta\_2 \pi\_t^{t-4} + \varepsilon\_t \tag{7}$$

where V<sup>t</sup> is the "sales" variable for quarterly period t, ω<sup>t</sup> refers to the mean number of employees in t, 1V t− t 4 /Vt−<sup>4</sup> is the growth rate of the "sales" variable, ∆ωt<sup>−</sup> t 4 /ωt−<sup>4</sup> represents the growth rate of the proxy variable of production, π t−2 t is the year-on-year inflation rate, β0, β1, and β<sup>2</sup> show the parameters of the model and ε<sup>t</sup> is the error term, which follows a normal distribution with zero mean.

According to Equation (7), to estimate FT capability, two exogenous variables are included, assuming variations in sales of the company can be caused by fluctuations in the inflation rate and the company's production level:


$$\frac{\Delta\phi\_t^{t-4}}{\alpha\_{t-4}} = \frac{\omega\_t - \alpha\_{t-4}}{\alpha\_{t-4}}\tag{8}$$

where ω<sup>t</sup> is the mean number of employees during quarterly period t. That notwithstanding, this paper presents a proposal to approximate the production level.

Conversely, Leibowitz et al. (1989) assume that the growth of a company's return rate can be approximately measured using the following equation, as assumed by Jareño and Navarro (2010):

$$
\log \approx \lg\_0 + \chi r + \lambda \,\pi \tag{9}
$$

where g is the growth rate of profits—equivalent to g in the simplified equation of the dividend discount model by Gordon (1959) and Gordon and Shapiro (1956), r is the real interest rate, π represents the expected inflation rate, g<sup>0</sup> is a constant that represents the long term growth rate, γ denotes the sensitivity of the growth rate of future profits to changes in the real interest rate and λ is the FT coefficient.

Following the logic of Jareño and Navarro (2010), the FT capability is related to a company's ability to translate inflation shocks into increased product/service prices, namely, 1p<sup>t</sup> [as shown in Equation (6)]. Specifically, we assume that:

$$\frac{\Delta p\_t}{p\_t} = f\left(\pi\_t, \pi\_{t-1}, \dots\right) = \alpha\_0 + \alpha\_1 \pi\_t + \alpha\_2 \pi\_{t-1} + \dots + \mu\_t \tag{10}$$

where α<sup>i</sup> measures the company's ability to transmit such shocks (previous ones and current ones) to the prices of its products, that is, it represents the essence of the FT coefficient concept, the estimation of which is one of this study's primary objectives.

This way, combining Equations (4) and (10) results in the following relationship, wheree<sup>t</sup> is the error term:

$$\frac{\Delta V\_t}{V\_t} = \alpha\_0 + \alpha\_1 \pi\_t + \alpha\_2 \pi\_{t-1} + \dots + \delta \frac{\Delta \alpha\_t}{\alpha\_l} + e\_l \tag{11}$$

Thus, to relate Equation (11) with the equation provided by Leibowitz (9) and the FT coefficient, we assume that the growth of the return rate of company (g) depends on both the relative changes in its turnover and on other variables that we denote as θ (omitted variables such as technological changes and other macroeconomic factors):

$$g = f\left(\frac{\triangle V\_t}{V\_t}, \overline{\theta}\right) = f(\alpha\_0 + \alpha\_1 \pi\_t + \alpha\_2 \pi\_{t-1} + \dots)$$

$$+ \delta \frac{\triangle \omega\_t}{\alpha\_l} + e\_t, \overline{\theta} \rangle \tag{12}$$

In addition, if we assume that f is linear with respect to π, we can write:

$$\mathbf{g} = m \left( \beta\_0 + \beta\_1 \frac{\triangle \alpha\_l}{\alpha\_l}, \overline{\theta} \right) + \theta\_1 \pi\_t + \theta\_2 \pi\_{t-1} + u\_t \tag{13}$$

Assuming that Et[ut] = 0 and taking into account Equations (13) and (9), we arrive at the following:

$$
\log + \gamma r \approx m \left( \beta\_0 + \beta\_1 \frac{\triangle \alpha\_t}{\alpha\_t}, \overline{\theta} \right) \tag{14}
$$

This way, the growth of a company's return rate over the long term and the real interest rate are related to its economic situation and indirectly to increases in the labor force and other residual variables, such as technological changes and other macroeconomic factors (θ).

The remaining terms in Equation (13), φ1π<sup>t</sup> + φ2πt−1, are related to FT capability (πλ). Specifically, Jareño and Navarro (2010) assume that 8 is monotonically related to λ, thus demonstrating a negative relationship between FT capability and the duration of stocks equivalent to a negative relationship between parameters 8<sup>i</sup> and the sensitivity of stocks to changes in the nominal interest rate.

This relationship is really important, because this connection would point out the relevance of this study for investor behavior (Blackburn et al., 2014; González et al., 2016). Thus, if this research could find significant differences in the FTC of companies by sector, then investors would seek protection from interest and inflation rate changes by taking into account this FT capability as a key factor in their investment decisions.

#### ESTIMATION OF THE FTC

#### Data and Sector Classification

To estimate the ability to absorb inflation (or FT ability) in the context of the United States, we use as a starting point Equations (7) and (11). Following this logic, we use individual data from 500 companies traded on the S&P 500. In addition, we classify the companies first by subsectors and then by activity sectors.

In addition, the study period is 2000–2009 (inclusive) and the frequency of data is quarterly, leading to 40 observations. The data were extracted from public reports of balance sheets and profit-and-loss statements, specifically, the quarterly sales of companies in the S&P 500, the number of employees in each quarter and the alternative proxy variable of production for each company. Some examples that support the selection of this proxy variable are Everaert and De Simone (2007), He (2010), Marple (1986), Rodrigues and Brady (1992), and Staunton (1986). Finally, we used data on the United States' inflation rate.

The United States inflation rate was obtained through data supplied by the Harmonized Consumer Price Index (for 2005), published by Eurostat. This way, we first calculated the interannual inflation rate with a monthly frequency to eliminate seasonal problems. Then, we converted the data to a quarterly frequency so that it coincided with the rest of the variables: i.e., turnover of the companies listed in the S&P 500 index and two proxy variables corresponding to production level (number of employees and operating costs).

The data referring to each company in the S&P 500's turnover during 2000–2009 were extracted from the Thomson Reuters database. However, the data for the proxy variables of the production level were available by sector from the North American Industry Classification System (NAICS).

Thus, the analyzed companies were grouped by NAICS sector, as shown in the Supplementary Material Table 1. We used 17 sectors with the following names and corresponding NAICS codes: Agriculture, Forestry, Fishing and Hunting (11), Mining (21), Utilities (22), Construction (23), Manufacturing (31–33), Wholesale Trade (42), Retail Trade (44–45), Transportation and Warehousing (48-49), Information (51), Finance and Insurance (52), Real Estate Rental and Leasing (53), Professional, Scientific and Technical Services (54), Administrative and Support and Waste Management and Remediation Services (56), Educational Services (61), Health Care and Social Assistance (62), Art, Entertainment and Recreation (71), and Accommodation and Food Services (72). Given its residual character, we ruled out the Business Administration (55) and Other Services (81) sectors. Similarly, to be more precise in our proposed classifications, we used the Bloomberg website, where one can enter each listed company's acronym to obtain additional information.

With respect to the sectors presented in the Supplementary Material Table 1, the category of Utilities includes activities that consist of the supply of gas and electricity generated using energy sources such as nuclear, solar, wind, hydraulic, geothermal, biomass, etc. The Manufacturing sector includes the largest number of companies and a diverse array of activities, including production machinery, textiles, food, drinks, tobacco, chemicals, wood, etc.

After performing the NAICS sector classification, we see that the quarterly data referring to operating costs, although they fit with the classification, are further aggregated. Therefore, to make homogeneous estimations that incorporate the independence of the estimation alternative used (e.g., number of employees or operating costs), we modified the sectors proposed in the NAICS classification.

Consequently, Agriculture, Forestry, Fishing and Hunting, and Mining were combined into a single sector. The same was done with Finance and Insurance, and Real Estate Rental and Leasing, Professional, Scientific and Technical Services, and Administrative and Support, and Waste Management and Remediation Services, Educational Services and Health Care and Social Assistance and finally, Arts, Entertainment, and Recreation and Accommodation and Food Services.

Following that modification, we had 12 activity sectors to consider (see **Table 1**).

Once we defined our sectors, we followed the same methodology that we used for the inflation rate and applied yearto-year calculations to the rest of the variables to avoid seasonality problems.

In addition, we conducted different tests to analyze the seasonality of the series: Augmented Dickey-Fuller (ADF), Phillips-Perron (PP), and Kwiatkowski-Phillips-Schmidt-Shin (KPSS). The results of the tests are not shown, but they are available upon request.

The "turnover" variable exhibits a unit root in levels, although the variable in the first differences was stationary at standard levels of significance.

#### Proxy Variable of Production Level

As previously stated, because the S&P 500-listed companies' production volume is a variable that cannot be observed directly, we were forced to seek a proxy variable. One important contribution of this study is the use of the proxy variable of production level, not only the number of employees (Jareño, 2005; Jareño and Navarro, 2010) but also some alternative variable such as operating costs. However, for comparison purposes, we performed the estimation of the FT coefficients considering the two variables selected as proxies for production level.

TABLE 1 | Adapted sectoral NAICS classification used in this study.


Source: Own elaboration based on http://www.naics.com/search.htm.

#### A. Operating Costs: Companies of the S&P 500

The data referring to operating costs were obtained on a monthly basis from the website of the Bureau of Labor Statistics in the form of index numbers for the year 2002. These costs are representative of production value and were extracted from public reports published annually by the actual companies studied. In this way, we extracted the information corresponding to the 12 sectors included in **Table 1** for 2000–2009.

To homogenize the data, we obtained the desired frequency by determining quarterly means and then doing year-to-year calculations. The series of operating costs is stationary in its first differences according to the tests noted above.

#### B. Number of Employees: S&P 500 Companies

The other alternative that we considered as a proxy variable for the production level of S&P 500 companies, following Jareño (2005) and Jareño and Navarro (2010), is number of employees. The reason that we selected this variable is that it has a direct and positive relationship with production volume. Accordingly, we were able to compare the results obtained using both variables and evaluate them for consistency.

Thus, data related to the number of employees was extracted from the website of the Bureau of Labor Statistics, on a monthly basis, for the 17 activity sectors included in the Supplementary Material Table 1 (in thousands of employees and for the sample period). However, we were required to adapt this information by adding the data in the 12 abovedescribed sectors and by obtaining quarterly and annual frequencies.

The stationarity and unit-root tests confirm the stationarity of the time series in first differences.

Finally, the Supplementary Material Table 2 collects the primary descriptive statistics of the variables used in this study. We observe that the mean and median are positive for all of the sectors using the variables "turnover" and "operating costs" and the variable "inflation," which indicates growth for both variables during the studied period. On the contrary, the mean and median of the variable "number of employees" shows a negative sign for more than half of the sectors analyzed, which indicates a reduction. With respect to standard deviation, the variable "turnover" is recurrently more volatile than the proxy variables for production level ("operating costs" and "number of employees") and "rate of inflation," fundamentally S6 and S1. The asymmetry exhibited by the explanatory variables is clearly negative, whereas the dependent variable ("turnover") does not show a clear sign. All of the variables exhibit excessive kurtosis.

After studying the data, we estimated the FT coefficients according to the proposed sector classifications. The idea was to analyze companies' ability to transmit inflation shocks to their prices, noting that those inflation shocks are a function of the sector of economic activity in which the companies were classified during 2000–2009; that analysis is this study's primary objective. In addition, we used two alternative estimates: the first using operating costs and the second using number of employees.

### US Sectoral Results of FT Coefficients Using an Alternative Estimation of US FTC

Starting with Equations (7) and (11) derived from previous studies, we created a system of 12 equations (one for each activity sector analyzed) using the following format:

$$d\left(\frac{\text{CNeg}\_{ti} - \text{CNeg}\_{t-4i}}{\text{CNeg}\_{t-4i}}\right) = \beta\_0 + \beta\_1 \cdot d\left(\frac{\text{CteOp}\_{ti} - \text{CteOp}\_{t-4i}}{\text{CteOp}\_{t-4i}}\right) + \varepsilon$$

$$\beta\_2 \cdot d\left(\frac{\text{TInf}\_t - \text{TInf}\_{t-4}}{\text{TInf}\_{t-4}}\right) + \varepsilon\_t \qquad \text{(15)}$$

where CNegti refers to the turnover for each sector i, CteOpti reflects the operating costs of the different sectors i and Tinf<sup>t</sup> the American inflation rate for 2000–2009. In addition, β<sup>0</sup> represents the independent term, β<sup>1</sup> is the coefficient that measures the variation in turnover for each activity sector as a result of unit variations in operating costs, β<sup>2</sup> is the FT coefficient—that is, it measures the capability of companies in the sector to transmit to prices (and, therefore, to turnover) inflation shocks in the economy—and ε<sup>t</sup> alludes to the error term.

This study used the Seemingly Unrelated Regression (SUR) method to obtain the FT coefficients. This method avoided problems related to heteroscedasticity and a possible contemporary correlation between the different equations' error terms (i.e., autocorrelation). The results can be seen in **Table 2**.

**Table 2** shows the estimated coefficients β0, β1, and β2, which represent the FT coefficients. As seen, in four sectors (S7, S9, S11, and S12), the results are significantly different from zero and the sign of the FT coefficient is positive: "Finance and Real Estate," "Manufacturing," "Transportation and Warehousing," and "Utilities." For the rest of the sectors, the results are not significantly different from zero.

### US Sectoral Results of FT Coefficients Using the Jareño and Navarro (2010) Methodology

Next, we estimated the FT coefficients following the method proposed by Jareño and Navarro (2010) and incorporating as the proxy variable for production level the aggregated number of employees by sector. In this way, we were able to compare the results obtained using both procedures to analyze their robustness.

Some previous studies have applied the Jareño and Navarro (2010) methodology. Thus, the robustness of this proposal has been tested, at international level, mainly in Peiró (2016), and, at Spanish level, in Ballester et al. (2011), Jareño and Tolentino (2012), and Díaz and Jareño (2013), among others. Moreover, this and other previous research has relied on this FT methodology in order to include a better explanation of each evidence found.

Therefore, the model used for the estimation is shown in the following equation:

$$d\left(\frac{\text{CNeg}\_{ti} - \text{CNeg}\_{t-4i}}{\text{CNeg}\_{t-4i}}\right) = \beta\_0 + \beta\_1 \cdot d\left(\frac{\text{NEm}\_{ti} - \text{NEm}\_{t-4i}}{\text{NEm}\_{t-4i}}\right) + \varepsilon$$

$$\beta\_2 \cdot d\left(\frac{\text{TInf}\_t - \text{TInf}\_{t-4}}{\text{TInf}\_{t-4}}\right) + \varepsilon\_t \quad \text{(16)}$$

Cano et al. Investor Behavior and Flow-through Capability



This table shows the results of the model proposed by Jareño and Navarro (2010) to estimate the FT capability of companies, but applying the alternative proxy variable representing the production level. The sample extended from 2000–2009 and the regression was estimated using SUR methodology:

$$\mathcal{O}\left(\frac{\text{C\u0eq}\_{\text{U}} - \text{C\u0eq}\_{\text{U}-4i}}{\text{C\u0eq}\_{\text{U}-4i}}\right) = \beta\_{\text{U}} + \beta\_{\text{I}} \cdot \mathcal{O}\left(\frac{\text{C\u0eq}\_{\text{U}} - \text{C\u0eq}\_{\text{U}-4i}}{\text{C\u0eq}\_{\text{U}-4i}}\right) + \beta\_{\text{Z}} \cdot \mathcal{O}\left(\frac{\text{T\u0eq}\_{\text{I}} - \text{T\u0eq}\_{\text{I}-4}}{\text{T\u0eq}\_{\text{I}-4}}\right) + \upsilon\_{\text{I}}$$

where Cnegti refers to the turnover for each sector i, CteOpti reflects the operating costs of the different sectors i, TInf<sup>t</sup> reflects the American inflation rate and ε<sup>t</sup> reflects the error term. <sup>a</sup>p < 0.15; <sup>b</sup>p < 0.05 (t-statistics in parenthesis).

where Nempti reflects the number of employees in each activity sector i during 2000–2009. In addition, β<sup>0</sup> represents the independent term, β<sup>1</sup> is the coefficient that measures the variation in turnover in each activity sector resulting from a unit variation in their number of employees, β<sup>2</sup> is the FT coefficient, and ε<sup>t</sup> is the error term.

Using the SUR methodology, we estimated a system consisting of 12 equations (one per sector). The results can be observed in **Table 3** as shown below.

According to this alternative estimation, we see that in addition to the four sectors that present a significantly positive FT ability in the previous estimation, S4 "Retail Trade" shows results that are significantly different from zero even though in this case, the sign of the FT coefficient is negative.

In general, the results are very consistent with the use of one or the other proxy variable for production level.

Adjustment of the model is very high in the sectors that exhibit a significant FT ability, reaching almost 70% in S9 with the use of operating costs to approximate production level.

#### Interpretation of Sectoral Results

Following the logic of Asikoglu and Ercan (1992) and Jareño and Navarro (2010), there are clear differences between the FT coefficients obtained for the industries analyzed. This demonstrates that not all industries have the same ability to maintain growth in their turnover and/or profits during an inflationary period, given that multiple differential factors can affect certain industries: market power, competition level, competitive strategies, economic context, etc.

In general, it is true for both estimation formulas that the significant FT coefficients have a positive sign. This leads to the conclusion that in the face of increased inflation, turnover also exhibits a positive variation because companies are capable of transmitting—to a lesser or greater degree—inflation shock to the price of their products or services. That is, the sign of the FT coefficients indicates a positive relationship between both variables for the sectors for which data are significant (S7 Finance and Real Estate, S9 Manufacturing, S11 Transportation and Warehousing and S12 Utilities).

On the other hand, S4 Retail Trade is only significant when using the second alternative (using the number of employees as the proxy variable for production level) and holds a negative sign. This confirms that inflation shocks have a negative effect on that sector's profits.

#### S7 Financial and Real Estate sector

The FT coefficients estimated using both alternative procedures are 4.2943 (using operating costs as a proxy variable for production level) and 3.6408 (using number of employees). The coefficients obtained are similar. Therefore, both estimation procedures could be substituted for each other, which in principle, verifies our selection of operating costs as an alternative variable to number of employees.

From an economic perspective, we must highlight that the sectors related to real estate and finance suffered greatly from the 2007 financial crisis. However, until then, the financial sector, and to a greater degree the real estate sector, was marked by a continuous increase in the price of its services provided and products (i.e., homes sold/rented). This could explain the extremely high value of its FT coefficient.

However, price increases ended with the burst of the real estate bubble, which presumably decreased those sectors' ability to transmit inflation shocks to their prices. Therefore, it would be interesting to perform estimations of the FT capability



This table gathers the results of the model proposed by Jareño and Navarro (2010) to estimate the FT capability of companies, using the number of employees as proxy variable for production level. The sample extended from 2000–2009 and the regression was estimated using the SUR methodology:

$$\mathcal{O}\left(\frac{\text{CN\#}g\_{\text{U}} - \text{CN\#}g\_{t-4i}}{\text{CN\#}g\_{t-4i}}\right) = \\_\circ + \beta\_1 \cdot \mathcal{O}\left(\frac{\text{N\#}\eta\varphi\_{\text{U}} - \text{N\#}\eta\varphi\_{t-4i}}{\text{N\#}\eta\varphi\_{t-4i}}\right) + \beta\_2 \cdot \mathcal{O}\left(\frac{\text{T\#}t\_1 - \text{T\#}t\_{t-4}}{\text{T\#}t\_{t-4}}\right) + \upsilon\_{\text{U}} $$

where Nempti reflects the number of employees of each activity sector i, Cnegti reflects the turnover for each sector i, TInf<sup>t</sup> reflects the American inflation rate and ε<sup>t</sup> reflects the error term. <sup>a</sup>p < 0.15; <sup>b</sup>p < 0.10; <sup>c</sup>p < 0.05 (t-statistics in parenthesis).

differentiated by sub-periods as a function of the economic cycle, but the scarcity of data prevented us from doing so.

#### S9 Manufacturing Sector

The FT coefficients estimated using both procedures are 3.5957 (using operating costs) and 4.4087 (using the number of employees). Again, the value of the coefficients is similar and significant in both cases, confirming the consistency of the results obtained.

This sector includes manufacturing related to machinery for production, textiles, food, beverages, tobacco, chemicals, wood, etc. Therefore, in the majority of cases, the companies that operate in this sector have great market power. For example, large American tobacco companies enjoy an oligopolistic position. A similar phenomenon occurs, to a lesser degree, with the production of certain production machinery, because the United States is not only at the vanguard of new technology but also makes significant investments into research, development and innovation. This could explain the elevated values of the FT coefficients, which are equivalent to those of the financial and real estate sector.

#### S11 Transportation and Warehousing Sector

The FT coefficients estimated are 2.8149 and 2.5387, respectively, for the use of operating costs and the number of employees as proxy variables for production level. It is seen again that the values obtained are practically the same.

In addition, the coefficients are lower than those of the sectors analyzed previously: the transportation and warehousing sector has no companies that enjoy a clearly dominant position in the market because it is a sector with an elevated level of competition (even foreign). However, the United States is the biggest world power, which makes the country brand a marketing tool with great international reach and therefore, the demand for American products is overwhelming. Thus, orders of American products must be transported to their destinations and, until their departure, they are stored in the place of origin. This could explain the values obtained for the coefficients.

#### S12 Utilities

Utilities constitute the last of the four sectors for which the FT coefficients are significant using both estimation procedures. The values obtained are 4.3839 and 4.5480, respectively, for the use of each proxy variable for production level. The values taken by the coefficients are similar, as is the level of significance.

This sector includes activities consistent with the supply of gas and electricity generated through energy sources such as nuclear, solar, wind, hydraulic, geothermal, biomass, etc., some of which are subject to government regulation. Therefore, these businesses are regulated. Consequently, if the economy experiences an inflation shock, then US authorities can exert their control over these companies, transmitting inflation to prices. This justifies the extremely elevated FT coefficients resulting from this estimation.

#### S4 Retail Trade sector

As mentioned previously, the coefficients relative to this sector are only significant when they are estimated using the proxy variable "number of employees" (not when using operating costs).

In addition, their sign is negative (–1.0549), indicating that in the face of an increase in inflation, the turnover of the companies included in this sector decreases. There are several possible economic explanations for this result:


The remaining unmentioned sectors (S1, S2, S3, S5, S6, S8, and S10) exhibit FT coefficients that are not significantly different from zero. However, there are subtleties worth mentioning with respect to those cases for which the FT coefficients take on a different sign using one estimation procedure from that when using the other estimation procedure.


Until the middle or final months of 2007, the number of homes grew exponentially, driven by the real estate bubble that formed during those years. Many investors bought and sold homes with the single goal of obtaining profits, without stopping to think whether the sale value represented the home's real value. This created a difficult future for the real estate sector and by extension, for the construction sector, which did later arrive. Construction has been one of the most affected sectors from the beginning of the global economic crisis, forcing the closing of many construction businesses that had been encouraged by artificial optimism.


We did the same for those sectors that can be broken down into two activity sub-sectors, such as S1 Leisure and Accommodation, S6 Forest and Mining Exploitation, S7 Finances and Real Estate, and finally, S10 Professional and Business Services. In this way, we propose a new system with 10 equations, estimated using SUR methodology through Equation (16), although it was only applied to the sub-sectors to which we have alluded because we can differentiate an FT coefficient for each of them. The results are shown in **Table 4**.

Before carrying out the disaggregated estimate of the FT coefficients, sector S2 Health Care and Educational Services showed a negative sign in the case of alternative 1 (operating costs) and a positive sign in the case of alternative 2 (number of employees). After breaking down the calculations for the different sub-sectors, one can see that the signs for both continue to be positive through the second alternative and are not significant.

Similarly, the FT coefficients of S7 Finance and the Real Estate Sector, which were positive and significant in the previous analysis, maintain their sign and level of significance when broken down into two sub-sectors.

Sub-sectors with significant coefficients resulting from the separate analysis include Accommodation Services (included in


This table shows the results of the model proposed by Jareño and Navarro (2010) to estimate the FT capability of companies, applying the number of employees as the proxy variable for production level. The sample extended from 2000–2009 and the regression was estimated using the SUR methodology:

$$\mathcal{O}\left(\frac{\text{CN\#}g\_{\text{\%}}-\text{CN\#}g\_{t-4\%}}{\text{CN\#}g\_{t-4\%}}\right) - \beta\_{\text{\%}} + \beta\_{\text{\%}}\mathcal{O}\left(\frac{\text{N\#}\eta\text{\mu}\_{\text{\%}}-\text{N\#}\eta\text{\mu}\_{t-4\%}}{\text{N\#}\eta\text{\mu}\_{\text{\%}-4\%}}\right) + \beta\_{\text{\%}}\mathcal{O}\left(\frac{\text{Th}\eta t\_{t}-\text{Th}\eta t\_{t-4}}{\text{Th}\eta t\_{t-4}}\right) + \varepsilon\_{\text{\%}}$$

where Nempti reflects the number of employees in each activity sector i, Cnegti reflects the turnover for each sector i, TInf<sup>t</sup> reflects the American inflation rate and ε<sup>t</sup> reflects the error term.

<sup>a</sup>p < 0.10; <sup>b</sup>p < 0.05 (t-statistics in parenthesis).

S1 of our NAICS adapted classification) and Mining (within S6). The FT coefficient for accommodation services (0.8274) is relatively low, given that it is a sector with a great deal of competition. However, we must remember that it is a positive and significant coefficient, which suggests that businesses in the sector are capable of partially transmitting inflation shocks to service prices. Conversely, Mining shows a very elevated value (11.812), which could be due to the sector's need to use natural resources that are increasingly scarce and expensive. This would explain the elevated ability to absorb inflation exhibited by the companies in this sector, given that an increase in the price of those resources would not disincentivize their purchase by clients with great purchasing power, given that they could be considered "luxury" goods (gold, silver, etc.).

### THE RELATIONSHIP BETWEEN FT COEFFICIENTS AND STOCK PRICES

This section addresses the study of the relationship between estimated sectoral FT coefficients and variations in stock prices corresponding S&P 500 companies during 2000–2009.

In that sense, we have data for the FT coefficients by sector, notwithstanding the fact that because we made estimations using two alternatives, we have sectoral data derived from each of them. In this case, we consider a simple cross-section regression that offers important information.

Conversely, for the case of the stock prices of S&P 500 companies, we have quarterly data for 2000–2009 (40 observations), extracted from the Thomson Reuters database. This means that we aggregated stock trading data from companies in the 12 sectors listed in the NAICS classification, as adapted for this study. We then calculated the variation in stock prices during the sample period to analyze their relationship with the previously estimated FT coefficients:

$$
\Delta \text{Cost}\_i = \frac{\text{Cost}\_{4T \ 2009i} - \text{Cost}\_{1T \ 2000i}}{\text{Cost}\_{1T \ 2000i}} \tag{17}
$$

where Cot<sup>i</sup> refers to the quarterly prices of stocks of companies included in sector i and T represents the quarter in question.

Following the work of Asikoglu and Ercan (1992), we estimated, using ordinary least squares (OLS) adjusted by White (to avoid heteroscedasticity problems), Equation (18), which relates the variation in quarterly stock prices to the estimated FT coefficients (both at the sector level). Our goal was to analyze the value and sign of the coefficient that relates to both magnitudes and statistical significance. The equation is as follows:

$$
\Delta \text{Cost}\_i = \delta\_0 + \delta\_1 \cdot \text{CFT}\_i \tag{18}
$$

where CFT<sup>i</sup> gathers the estimated FT coefficients for each sector i, δ<sup>0</sup> is the independent term, and δ<sup>1</sup> is the coefficient that relates the variations in stock prices with the FT coefficients.

As indicated previously, this equation was estimated twice (**Table 5** and **Figure 1**): once with the FT coefficients obtained through the first alternative (operating costs) and second, with the coefficients resulting from the second alternative (number TABLE 5 | Estimation of the relationship between sectoral FT coefficients and variations in stock prices.


This table gathers the results of the model proposed by Asikoglu and Ercan (1992) to study the relationship between the FT ability of companies classified at the sector level and changes in their stock prices. The sample extends from 2000–2009 and the regression was estimated by ordinary least squares (OLS) adjusted by White (to avoid heteroscedasticity issues):

$$
\Delta \mathbf{C} \mathbf{x} t\_i = \boldsymbol{\delta}\_{\mathbb{O}} + \boldsymbol{\delta}\_1 \cdot \mathbf{C} \boldsymbol{\delta} \boldsymbol{\tau} t\_i
$$

where CFT<sup>i</sup> represents the FT coefficients estimated for each sector i, δ<sup>0</sup> represents the independent term and δ<sup>1</sup> represents the coefficient that relates variations in stock price with FT coefficients.

<sup>a</sup>p < 0.05 (t-statistics in parenthesis).

of employees). Thus, we were able to compare the results of both alternatives and to verify (or not) the positive relationship between stock prices and the FT coefficients, as defended by Asikoglu and Ercan (1992) and Jareño and Navarro (2010).

As shown above, the δ<sup>0</sup> coefficient is not significant if we use operating costs as the proxy for production level. Therefore, although it has a positive sign, as argued by Asikoglu and Ercan (1992) and Jareño and Navarro (2010), it does not have statistical significance. However, when estimating the model proposed with FT coefficients resulting from using the number of employees as the proxy variable for production level, the δ<sup>1</sup> coefficient (=0.1040) is similar in magnitude to the previous estimation, but in this case it is significant and has a positive sign. In this way, we see results in line with the existing literature.

According to Asikoglu and Ercan (1992), Jareño and Navarro (2010) and our own results, there is a positive relationship between changes in stock prices and FT capability, which varies as a function of the sector under consideration. As a result, investors are willing to pay a higher price for stock in those companies that are capable of transmitting larger portions of inflation shocks to their product prices, which would mean, in the end, a growth in profits/dividends. Accordingly, the empirical evidence presented suggests that FT ability has an effect on stock prices.

Similarly, Asikoglu and Ercan (1992) argue for a negative relationship between inflation and stock prices at the sector level in the US and therefore, the same is true for the relationship between inflation and FT capability. Following this logic, increasing the FT capability is associated with higher stock prices and in industries in which such coefficients are higher, stock prices are less sensitive to inflation shocks. This last conclusion is demonstrated for our case in **Table 6**, which corroborates the primary idea of the FT model proposed by Asikoglu and Ercan (1992): when inflation increases, pressure on stock prices through the discount rate is counteracted to a certain point by increments in the growth expected of the nominal rate of return of equity securities. This compensatory effect is positively related to the FT coefficient of the industry in question.

TABLE 6 | Relationship between stock price variation and FT coefficients at the sector level.


CFT collects the estimated FT coefficients for each sector.

Comparing the column "Var. Stock Price 2000/09," which indicates the variation in stock price during 2000–2009 with the FT coefficients (last two columns), we see that in general, in those sectors in which these coefficients are higher, variation is lower. This fact can be verified for S9 Manufacturing and S7 Finance and Real Estate. However, it does not hold true for S6 Forest and Mining Exploitation because it has the highest FT coefficient using both estimation alternatives, but the variation in the stock prices is not the lowest at the sector level, although it does show one of the highest stock prices.

In contrast, empirical evidence presented by Asikoglu and Ercan (1992) suggests that in addition to the effect of FT ability on stock prices, the coefficients that measure it in different sectors differ significantly. Therefore, we analyzed whether different sectors' different capacities to absorb inflation (analyzed in a pair-wise fashion) exhibit statistical significance.

To that end, we performed the Wald test for both estimation alternatives (see **Table 7**). That test is a symmetrical matrix that evaluates, in a pair-wise fashion, whether the estimated FT coefficients are significantly different. There is no reason to estimate the principal diagonal because it compares of each sector to itself (with an equality probability of 100%), and the upper diagonal exactly coincides with the lower one.

**Table 7** shows the results of the test performed, which offers the value of the Chi-squared distribution for each pair of FT coefficients and for the level of significance. Accordingly, if two coefficients are significantly different, then the ability of companies in one sector to absorb inflation is significantly different then the ability exhibited by companies in the other sector.

As expected, notably different results are obtained depending on which sectors we compared. We must highlight that in all cases in which the pairs of FT coefficients are significantly different (for both of the estimations been proposed here), there is involvement by one of the sectors for which the estimated coefficients are shown to be significant in **Tables 3, 4**.

The remaining sectors, analyzed in a pair-wise fashion, offer results that are not significant, which is why we can assume that the capability of companies in one sector to absorb inflation is significantly different from that of companies in an alternate sector that is being compared.

The fact that we have found cases in which the differences are significant and others in which they are not corroborates the idea that sectors have different abilities to transmit inflation shocks to the price of their products and services. Therefore, our results are in line with the studies carried out by Asikoglu and Ercan (1992), Jareño (2005), and Jareño and Navarro (2010).

#### OVERALL RESULTS

In short, estimates about the capability of American companies to transmit inflation shocks to the prices of the products that they sell and/or the services that they provide (i.e., estimates about the FT capability) seem to be quite different among sectors. These results are in line with previous literature, such as Asikoglu and Ercan (1992), Jareño (2005), Ertek (2009), and Jareño and Navarro (2010), and, finally, Ang et al. (2011), at company level.

Second, there may be a positive relationship between changes in stock prices and FT capability, and also, this relationship varies among sectors. Previous studies (Asikoglu and Ercan, 1992;


TABLE 7 | Wald test to evaluate whether sectoral FT coefficients are pair-wise significantly different (two alternatives).

Ertek, 2009; Díaz and Jareño, 2013; Taylor, 2013) find similar evidence.

Finally, as suggested by Asikoglu and Ercan (1992) and Jareño and Navarro (2010), the FT coefficients by sector are statistically and significantly different, and our study corroborate it. As a result, this research may confirm that investors should consider the FT capability in decision-making (Kusev and van Schaik, 2011).

### DISCUSSION

This research is based on the impression that the investor behavior may be different depending on the company's capability to transmit inflation shocks to the prices of its products and services (i.e., the FTC), since investors seek protection from interest and inflation rate changes, among other sorts of risk. Additionally, higher FT capability is associated with higher stock prices, and in industries in which FT coefficients are higher, stock prices are less sensitive to inflation shocks. Therefore, the FT capability is a key factor in investment decisions.

To that end, the primary objective of this study is to evaluate the ability of American companies (listed in the S&P 500 index) to transmit inflation shocks to the prices of their products and/or services, grouping companies into 12 sectors based on their activity, according to the adapted North American NAICS classification that we propose. The study period is inclusive of 2000–2009 and the data frequency is quarterly, which also represents an improvement compared to previous studies with semiannual frequencies.

First, we made two alternate estimations of companies' power to transfer inflation shocks to their prices as a function of two proxy variables for production level (a variable that is not directly observable): operating costs and number of employees. The estimated FT coefficients differ considerably as a function of each sector, in line with previous studies (Asikoglu and Ercan, 1992, and Jareño and Navarro, 2010). This result can be explained as a function of the peculiarities that affect companies in each sector differently, such as level of competition, competitiveness strategies, market share, economic context, etc. Moreover, this result would explain differences in decision-making by investors depending on the FT capability.

This research also demonstrates the existence of a positive relationship between the variation exhibited by stock prices in the companies of a single sector and its corresponding FT coefficient. This positive relationship exists because investors

<sup>1</sup> (a) p <0, 15;(b) p <0, 10;(c) p <0, 05.

are willing to pay a higher price for stocks when a larger part of the inflation rate provided is transmitted to them in the form of growth in earnings/dividends obtained. Thus, increments in FT coefficients are associated with higher stock prices.

Similarly, the empirical evidence indicates that in those industries that have a greater ability to transmit to prices, in the majority of cases, stock prices are less sensitive to inflation shocks. This is not due to the presence of a negative relationship between inflation and stock price levels, although it is true that in those sectors in which FT capability is relatively high, inflation shocks are transmitted, practically in their entirety, to the price of products sold and services provided. Therefore, investors trust stock prices given that their valuation can remain intact.

In general, we see that the results obtained are in agreement with those of the existing literature, because the FT capability by sector considerably differ. However, we propose that we extend this work in the future by widening our sample (with respect to the number of observations and companies analyzed) so that we can further break down our proposed sectors by activity and outline the results by sub-periods in greater detail, differentiating between boom times and economic crises. In addition, other lines of research that consider different proxy variables, such as the size of the sector analyzed in each case, are also open. Finally, for the companies analyzed in the present study, it would be interesting to investigate whether there is an inverse relationship between FT capability and the sensitivity of returns against changes in interest rates.

To conclude, investor behavior should consider the FT capability of the sector that each company belongs to before

#### REFERENCES


making an investment decision. As suggested by previous literature, our results support the state-dependent nature of the investor behavior in the inflation analysis. Similarly, this study may find a herding behavior of investors, because in some scenarios, investors disregard their own information and exhibit herding behavior, which is often extremely optimistic or pessimistic and may lead to an unreasonable reaction to movements in inflation rates when the FTC is higher or lower than the expected one. Finally, we confirm the null hypothesis that investor behavior may depend on different factors that affect the decision-making. Therefore, aspects such as the sector that traded stock belongs to and the business cycle definitely impact on investment behavior. Consequently, the FTC would be a key factor in investment decisions.

#### AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication.

#### ACKNOWLEDGMENTS

This research is funded by "Fundación Ramón Areces (Convocatoria Pública año 2013), Riesgo de interés e inflación: estudio del mercado bursátil norteamericano" grant.

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fpsyg. 2016.00668

variable approach. Empir. Econ. 33, 449–468. doi: 10.1007/s00181-006- 0109-y


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Cano, Jareño and Tolentino. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Interest and Inflation Risk: Investor Behavior

#### María de la O González <sup>1</sup> , Francisco Jareño<sup>1</sup> \* and Frank S. Skinner <sup>2</sup>

<sup>1</sup> Department of Economic Analysis and Finance, University of Castilla-La Mancha, Albacete, Spain, <sup>2</sup> Department of Economics and Finance, Brunel University, Uxbridge, UK

We examine investor behavior under interest and inflation risk in different scenarios. To that end, we analyze the relation between stock returns and unexpected changes in nominal and real interest rates and inflation for the US stock market. This relation is examined in detail by breaking the results down from the US stock market level to sector, sub-sector, and to individual industries as the ability of different industries to absorb unexpected changes in interest rates and inflation can vary by industry and by contraction and expansion sub-periods. While most significant relations are conventionally negative, some are consistently positive. This suggests some relevant implications on investor behavior. Thus, investments in industries with this positive relation can form a safe haven from unexpected changes in real and nominal interest rates. Gold has an insignificant beta during recessionary conditions hinting that Gold can be a safe haven during recessions. However, Gold also has a consistent negative relation to unexpected changes in inflation thereby damaging the claim that Gold is a hedge against inflation.

#### *Edited by:*

Alicia Izquierdo-Yusta, University of Burgos, Spain

#### *Reviewed by:*

Jorge Pelegrín-Borondo, Universidad de La Rioja, Spain Inés González, Universidad Pública de Navarra, Spain

#### *\*Correspondence:*

Francisco Jareño francisco.jareno@uclm.es

#### *Specialty section:*

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

*Received:* 29 January 2016 *Accepted:* 04 March 2016 *Published:* 18 March 2016

#### *Citation:*

González MO, Jareño F and Skinner FS (2016) Interest and Inflation Risk: Investor Behavior. Front. Psychol. 7:390. doi: 10.3389/fpsyg.2016.00390 Keywords: unexpected inflation, interest rates, stock return, business cycle, investor behavior

## INTRODUCTION

A lot of previous financial research assumes that investors are rational agents, so they try to optimize wealth and minimize risk (Campbell, 2006). Thus, the study of two relevant sources of risk such as interest and inflation rate movements is very interesting for deepening on the analysis of investor behavior as well as for portfolio managers. Furthermore, the recent financial crisis confirms that investor behavior changes over time (Ferrando et al., 2015), so this analysis is really challenging to achieve a better understanding of investor behavior. Moreover, according to Blackburn et al. (2014), investor behavior may depend on different factors that affect the investment or trading decision. Therefore, aspects such as the sector that traded stock belongs to and the business cycle—among others—apparently impact on investment behavior.

The US stock market is a world reference market so unexpected changes in US nominal interest rates can affect stock markets worldwide. Moreover, being the most active equity market with the longest series of detailed quality data, the US stock market is a natural laboratory to study the relationships between unanticipated inflation and its co-dependents, unanticipated changes in real, and nominal interest rates, in detail by sector and by varying economic conditions. It is important to examine these relations by sector because there is no reason to expect that individual sector returns are always inversely related to unanticipated changes in inflation and real and nominal interest rates. For instance, according to the flow through model of Estep and Hanson (1980), the impact of inflation on stock prices can be neutral if the firm can pass on inflationary price increases to consumers. If so, then an investment in stocks can serve as a safe haven for investors as stock prices rise with inflation. Additionally, the impact of unanticipated real and nominal interest rate changes can vary by sector depending upon the characteristic leverage and competitive structure of the sector. Moreover, it is also important to examine these relations by time period as conventionally inverse relations can turn positive as economic conditions change. For instance, it could be that an investment in cyclical industries such as the Industrial sector can have a positive relation with unanticipated inflation during boom economics conditions that turns negative during recessions.

**Figure 1** presents the evolution of the US stock market index (S&P500) and the 10-year Treasury bond yield from September 1989 to February 2014. On the one hand, the US stock market exhibits an increasing trend during most of the period, only interrupted by the dot-com bubble burst in 2000 and the global financial crisis at the end of 2007 (Bartram and Bodnar, 2009). On the other hand, the 10-year Treasury bond yield shows a decreasing tendency. So at first glance, we observe clear evidence of the inverse association between US stock market returns and changes in the nominal interest rate. However, we raise the question of whether this inverse relation is consistent by subperiod and whether this inverse association is maintained when we break down unexpected changes in the nominal interest rate into unexpected changes in the real interest and inflation rates, especially when we examine these relations by sector, industry, and by economic condition. Thus, the crucial aim of this paper is to analyze the details of the relation between returns on US stocks and unexpected changes in nominal and real interest rates and inflation, because the investor behavior may be quite different depending on the sector, industry, and the state of the economy.

The literature examines the sensitivity of stock returns to unexpected changes in nominal interest rates finding a negative and significant relationship between stock returns and unanticipated changes in nominal interest rates. See O'Neal (1998), Fraser et al. (2002), Hevert et al. (1998a,b), Tessaromatis (2003), Jareño (2006, 2008), Ferrer et al. (2010), Korkeamäki (2011), Ferrando et al. (2015), and Campos et al. (2016) as examples. Some have examined these relations for the overall stock market (Elyasiani and Mansur, 1998; Oertmann et al., 2000; Shamsuddin, 2014) while others have mainly studied these relations for financial companies (Flannery and James, 1984; Fraser et al., 2002; Staikouras, 2003, 2006; Au Yong and Faff, 2008; Drehmann et al., 2010; Ballester et al., 2011; Memmel, 2011; Bessler and Kurmann, 2012; Abdymomunov and Gerlach, 2014) or for Utilities (Sweeney and Warga, 1986). Others have deepened the analysis by decomposing unexpected changes in nominal interest rates into unexpected changes in real interest and unexpected inflation rates (Tessaromatis, 2003; Jareño, 2006, 2008; Jareño et al., 2016).

This paper is one of the few to estimate the stock return response to unexpected shocks in the nominal interest rate and its components, unexpected changes in the inflation rate, and the residual that we interpret as unexpected changes in the real interest rate. Moreover, this study tries to approximate investor behavior analysing sector stock response to changes in sources of risk in different scenarios. To accomplish this task, we use an extension of the Stone (1974) two-factor model proposed in Jareño (2006) and, partly, in Jareño (2008), and Jareño and Navarro (2010). Using this approach, we make two contributions. First, we analyze these relations at the sector, sub-sector and industry level. Thus, we estimate not only the relation between stock returns and unexpected nominal interest rate changes but also the relations between stock returns and unexpected changes in the real interest and inflation rates by sector, sub-sector, and individual industries. Second, we examine a long time period, from September 1989 to February 2014. This period encompasses a wide variety of economic conditions, including one of the longest expansion periods for

the US economy, one of the most severe credit contractions in living memory and several recessions. This sample variation in economic conditions allows us to explore the stability of these relations overall, and by sector, sub-sector, and industry. This detailed investigation into the stability of these relations allows us to search for special industries whose response to unexpected changes in nominal and real interest rates, and unanticipated inflation, is consistent, either positive or negative, thereby providing valuable information for investors and policy makers who have to consider these important sources of systematic risk.

In general, we find that investor behavior seems to be quite different over time (according to the business cycle) and by sector. Specifically, some financial (as well as non-financial) sectors have insignificant relations and we even find some contrary results when examining the relations by sector, subsector, and industry. Some industries have a consistent significant positive relation between stock returns and unexpected changes in real and nominal interest rates. Interestingly, Gold, among others, has a negative relation to unanticipated inflation in the overall sample and in the contraction and expansion sub-periods suggesting that it is exposed to inflation risk.

The rest of the paper is structured as follows. Section Materials and Methods present the main methodology used in this research. Section Data describes the data and variables included in our empirical analysis. Section Empirical Results comments on the results of our research, and finally, Section Discussion makes concluding remarks.

### MATERIALS AND METHODS

In this section, we explain how we measure unexpected changes in the nominal rate of interest. Then we explain how we measure the expected rate of inflation that is used as an input to decompose the unexpected change in nominal interest rates into unexpected changes in inflation and unexpected changes in the real rate of interest. Finally, we describe how we classify the state of the economy into expansion and non-expansion (contraction) states.

### Unexpected Changes in Nominal Interest Rates

Sweeney and Warga (1986), Kane and Unal (1988), Bartram (2002), Oertmann et al. (2000), and Olugbode et al. (2014) amongst others use changes in long-term interest rates as a proxy for unexpected changes in nominal interest rates because long term interest rates incorporate the expectations of economic agents and because long term interest rates are important as they determine the cost of corporate borrowing. Thus, long term interest rates strongly influence the investment decisions of firms and therefore affect the value of companies. Alternative proxies for unexpected changes in nominal interest rates such as forecast error of an empirical ARIMA process for long term interest rates or survey data on the US federal funds rate (Benink and Wolff, 2000) have their own advantages and disadvantages (Froot, 1989) so no one proxy dominates. Therefore, we follow common practice and use the first difference of the long-term interest rate as a proxy for unexpected changes in the nominal interest rate.

The returns on Treasury securities for different maturities are usually used as risk-free interest rate proxies because Treasury securities are commonly assumed to have no default risk. Of the long term maturities, 10 years tends to be the most liquid and accurately estimated as the Fed has continuously auctioned 10 year Treasury notes throughout our sample period so there is always a recently issued 10-year note that the Fed can use to accurately estimate 10 year treasury yields. Therefore, we use changes in the 10-year US Treasury bond yields, as reported by the Federal Reserve Bank of New York, as our approximation for unexpected changes in the nominal interest rate. We repeat our empirical results using 3-month, 1-, 3-, and 5- year US Treasury bond yield changes and find the results are very similar. These are available from the corresponding author upon request.

### Expected Inflation Rates

Although, previous studies have applied a variety of methodologies to estimate expected inflation rates, a lot of related, and crucial papers (Schwert, 1981; Pearce and Roley, 1988; Fraser et al., 2002; Mestel and Gurgul, 2003; Jareño, 2008), use simple time series ARIMA models to estimate the expected inflation component. These studies assume that the current total inflation rate (πt) can be broken down into the sum of its expected (π e t ) and unexpected (π u t ) components. Thus, the expected component is estimated using ARIMA models thereby assuming that this component depends upon its own past series. Then the forecast errors from the ARIMA model form our estimate of unanticipated changes in inflation. We also use ARIMA models because authors, such as Joyce and Read (2002) and Browne and Doran (2005), observe similar results using ARIMA and other alternative and more sophisticated procedures. These models, in contrast to structural models, do not need additional information for doing forecasts, because they use lagged inflation values. We have repeated this procedure until the end of sample, with one-step-ahead forecast, obtaining the expected component of inflation rate.

We use the Akaike information criterion (AIC) to choose the ARMA (1, 0) process to predict the month-to-month annualized inflation rate. Therefore, we suppose short-sighted expectations as in Leiser and Drori (2005). Unit root tests confirm that inflation rate is a I(1) series, so this result is consistent with short-sightedness expectations. That is,

$$E\_t\left(\pi\_{t,\ t+1}\right) = \rho \pi\_{t-1,\ t}$$

In other words, expectations are formed in part [ρ according to the ARMA(1, 0) process], as of time t for the expected rate of annualized inflation π over the next month t+1 based on the most recent monthly annualized inflation rate that evolved from t−1 to t.

### Unexpected Changes in the Real Rate of Interest

As mentioned above, we use changes in the 10-year US Treasury bond yield as our approximation for unexpected changes in the nominal interest rate. To obtain unexpected changes in the real rate of interest we assume the Fisher approximation and subtract the expected rate of inflation Et(πt,<sup>t</sup> <sup>+</sup> <sup>1</sup>) as estimated above from the nominal rate of interest i<sup>t</sup> .

$$r\_t \approx i\_t - E\_t \left(\pi\_{t, t+1}\right)$$

Then, changes in the above relation form our approximation for unexpected changes in in the real rate of interest.

### The Stone (1974) Two-Factor Model

The literature focuses mainly on the Stone (1974) two-factor model to measure the interest rate sensitivity of stock returns (Lynge and Zumwalt, 1980; Sweeney and Warga, 1986; O'Neal, 1998; Bartram, 2002; Fraser et al., 2002; Soto et al., 2005; Staikouras, 2005; Jareño, 2006, 2008; Ferrer et al., 2010). We use an extension of the Stone (1974) model that decomposes unexpected changes in the nominal interest rate into unexpected changes in real interest and inflation components in the nature of Tessaromatis (2003), Cornell (2000), Jareño (2006, 2008). However, all of these studies do not examine any sector other than the financial or the utility sector. Thus, we propose an analysis at the sector, sub-sector and industry level using an extension of the Stone (1974) model.

Typically, studies of interest rate sensitivity of stock returns start from the Capital Asset Pricing Model CAPM augmented by unexpected changes in nominal interest rates (Stone, 1974) to better explain the stochastic process that generates security returns. Therefore, adjusting Arango et al.'s (2002) model of stock returns by sector, sub-sector and industry we have,

$$r\_{jt} = \alpha\_j + \beta\_j \cdot r\_{mt} + \gamma\_j \cdot \Delta i\_t^u + \varepsilon\_{jt}$$

where rjt is the stock (sector, sub-sector or industry) j return in month t, β<sup>j</sup> shows the stock sensitivity to market movements, rmt is the return on the market portfolio, 1i u t represents unexpected changes in nominal interest rates, and, finally, εjt is the error term.

We extend the Stone (1974) model by applying the Fisher approximation to break down nominal interest rates i<sup>t</sup> into real interest r<sup>t</sup> and expected inflation Et(πt,<sup>t</sup> <sup>+</sup> <sup>1</sup>) components. Taking the first difference in interest rates as unexpected changes in nominal interest rates at time t, we then have unexpected changes in the nominal interest rate 1i u t as a linear combination of unexpected changes in the real rate 1r<sup>t</sup> and unexpected changes in the anticipated inflation rate 1Et(πt,<sup>t</sup> <sup>+</sup> <sup>1</sup>). Thus, the second model estimated in this paper is the following:

$$r\_{jt} = \alpha\_j + \beta\_j \cdot r\_{mt} + \beta\_{jr} \cdot \Delta r\_t + \beta\_{j\pi} \cdot \Delta E\_t^{\text{ORT}} \left(\pi\_{t, t+1}\right) + \varepsilon\_{jt}$$

where rjt is the stock (sector, sub-sector or industry) j return in month t, β<sup>j</sup> shows the stock sensitivity to market movements, rmt is the return on the market portfolio, 1r<sup>t</sup> represents unexpected changes in real interest rates, 1E ORT t πt,t+<sup>1</sup> shows shocks in the expected inflation rate (hereafter, unexpected changes in the inflation rate that we later explain is orthogonalized), and finally, εjt is the error term.

To avoid possible high collinearity between the explanatory variables, the financial economics literature uses some orthogonalization procedure. In **Table 1A** we observe a high, significant correlation between unexpected changes in real interest and unexpected changes in the inflation rate (about −83%). We also find two other significant correlations that we do not need to orthogonalize as they do not simultaneously occur in our model; the first is between changes in real and nominal interest rates (about 44%) and the second is between unexpected changes in inflation and nominal interest rates (about 15%). So, as in Lynge and Zumwalt (1980), Flannery and James (1984), Sweeney (1998), and Fraser et al. (2002), we orthogonalize the relation between unexpected changes in the real interest rate and unexpected changes in the inflation rate by regressing changes in the unexpected inflation rate on a constant and changes in the unexpected real interest rate using ordinary least squares regression. The residual from this regression forms our proxy for the orthogonalized unexpected change in the inflation rate. Thus, the effect of each factor is isolated and the movement that remains is captured by the residuals.

We choose this orthogonalization method because this is in line with the aim of this research, which is to estimate the response of stock (sector, sub-sector and industry) returns to unanticipated changes in nominal interest rates and its' decomposition, unexpected changes in real, and unexpected changes in inflation rates. Therefore, we analyze direct and indirect effects of interest rate shocks and obtain clear economic intuition. We find similar results to those obtained without orthogonalizing and also very similar results when we

TABLE 1 | Correlation matrix between explanatory variables included in the model.


rmt is the return on the market portfolio; 1i u t represents (unexpected) changes in nominal interest rates (10-year US Treasury bond yield changes); 1r<sup>t</sup> changes in real interest rates and 1E ORT t πt,t+<sup>1</sup> shows unexpected changes in the orthogonalized inflation rate. \*, \*\*, \*\*\* indicate statistical significance at the 10, 5, and 1% levels, respectively.

interchange the dependent and independent variables. Thus, our results seem to be robust, since this orthogonalization process evidently only eliminates the correlation between variables.

The final correlations between explanatory variables included in our model are reported in **Table 1B**. Notice that the correlation between unexpected changes in the real interest rate and unexpected changes in the inflation rate is zero.

TABLE 2 | US business cycle expansions and contractions.


Source: NBER (The National Bureau of Economic Research).

NBER'S classification is only available until June 2009, so we extend it by classifying all months with growth above the previous peak as expansion and all other months as contraction according to the US GDP after seasonal adjustment (Díaz and Jareño, 2013). For more information, please see the guide to the National Income & Products Accounts of the United States (NIPA) www.bea.gov/national/pdf/nipaguidepdf.

#### State of the Economy

Like Veronesi (1999), Knif et al. (2008), and Díaz and Jareño (2009, 2013), we assume that the stock market response to unanticipated changes in nominal and real interest and inflation rates depends on the business cycle. Therefore, we need to classify the state of the economy. We follow the National Bureau of Economic Research (NBER's) classification, but this is only available until June 2009. Therefore, we extend this classification by examining the evolution of the annual growth of the US GDP after seasonal adjustment (as in Díaz and Jareño, 2013) in order to identify expansion and non-expansion (contraction) months. Specifically, a contraction begins with a recession as defined as two or more quarters of negative seasonally adjusted growth. A contraction continues throughout the recovery period and converts to an expansion only when seasonally adjusted GDP rises above the peak of GDP just prior to the recession.

**Table 2** and **Figure 2** show the business cycle timing. This classification follows NBER announcements for the most part and divides the state of the economy in expansion and contraction months. During the 292 month period, from November 1989 to February 2014, the US Economy was in an expansion during 237 months and in contraction during 55 months. So, there were three contraction and four expansion periods.

### DATA

Our data set includes monthly indices for the US sector, subsector, and industries from November 1989 to February 2014, 292 monthly observations in all. The US sector index is based on the "Global Industry Classification Standard" GICS as developed by Morgan Stanley Capital International and Standard &Poor's.

This classification aims to enhance the investment research and asset management process for financial professionals worldwide. Also, GICS is the result of numerous discussions with asset owners, portfolio managers and investment analysts. Finally, this classification is designed to respond to the global financial community's need for an accurate, complete, and standard industry definition. The sub-sector and individual industry indices are refinements of the GICS compiled by and obtained from Bloomberg. We also use the monthly S&P500 market index from Bloomberg and the monthly 10-year US Treasury yields from the Federal Reserve. Finally, we use the monthly expected inflation rates as explained in Section Expected Inflation Rates.

The Supplementary Material Table A reports the sector, sub-sector, and industry classifications according to the GICS combined with the Bloomberg refinements. In this paper we analyze 10 sectors, subdivided into 33 sub-sectors, and further refined into 82 industries. The largest US industry sectors by market capitalization (as of April 29, 2010), are Information Technology (19.02%), and Financials (16.58%). There are five other noteworthy sectors with weights around 10%, Consumer Discretionary, Consumer Staples, Energy, Health Care, and Industrials.

**Table 3** reports the monthly returns for the S&P500 Index and the US sector indices. The mean and median returns for all sectors and the market are positive and fairly large; the mean monthly return is 58 basis points or 7.2% on an annual basis. Changes in the 10-year US bond yield, our proxy for the unexpected changes in the nominal interest rate, are negative as are unexpected changes in real interest and inflation rates, reflecting the decreasing trend of long-term interest rates as shown earlier in **Figure 1**. The most volatile sector is Information Technology, followed by Financials. Also, sector and market stock return volatilities are higher than nominal and real interest and inflation rate volatilities. Except for the real rate of interest, all the variables exhibit negative skewness and all variables have excess kurtosis, especially for unexpected changes in the inflation rate. The Jarque-Bera test rejects the null hypothesis of Normal distribution in all cases at the 5% significance level except for unexpected changes in nominal interest rates.

We examine the stationary of the variables in the second part of **Table 3** using the augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) unit root tests and the Kwiatkowski-Phillips-Schmidt-Shin (KPSS) stationary test. Similar to Jareño (2008), Czaja et al. (2009), and Ballester et al. (2011) these tests corroborate that the series of sector and stock market portfolio returns, real interest rate, and expected inflation rates, are stationary.

In order to obtain continuously compounded returns for industry sectors, sub-sectors, and industries rjt, we compute the log relatives using the closing index of the last day of the current month Pjt relative to the closing index of last day of the previous month Pjt <sup>−</sup> <sup>1</sup>. That is,

$$r\_{\vec{\jmath}t} = \log\left(\frac{P\_{\vec{\jmath}t}}{P\_{\vec{\jmath}t} - 1}\right).$$

To avoid income smoothing, we use index values net of dividends. We use the S&P500 index as a suitable representative of the US stock market and compute the log relative return in an analogous way as in (3) to obtain market log-returns.

TABLE 3 | Descriptive statistics of sector and market returns, 10-year US Treasury bond yield changes (nominal interest rates), and real interest and expected inflation rate changes.


This table presents the main descriptive statistics of monthly sector and market portfolio returns and changes in 10-year US Treasury bond yields over the period from November 1989 to February 2014. They include mean, median, maximum (Max.), and minimum (Min.) values, standard deviation (Std. Dev.), and Skewness (Skew.) and Kurtosis measures. JB denotes the statistic of the Jarque-Bera test for normality. The results of the augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) unit root tests and the Kwiatkowsky-Phillips-Schmidt-Shin (KPSS) stationarity test are also reported in the last three columns. As usual, <sup>∗</sup> , ∗∗ , ∗∗∗ indicate statistical significance at the 10, 5, and 1% levels, respectively.

### EMPIRICAL RESULTS

We estimate two models, (1) examines the relation between stock returns and unanticipated changes in nominal interest rates and (2) estimates the relation between stock returns and unanticipated changes in real interest and inflation rates. Both models are applied separately by sector, sub-sector, and industry and are estimated throughout the sample period and during expansion and contraction economic sub-periods from September 1989 to February 2014. We estimate models (1) and (2) separately using the "seemingly unrelated regression" SUR technique (Zellner, 1962) for each of the sector, sub-sector, and industry samples, six SUR regressions in all, thereby taking into account possible contemporaneous correlation in the error terms across sectors, sub sectors, and industries as well as heteroskedasticity.

### Results at the Sector Level

We regress models (1) and (2) at the sector level and we report the results in **Table 4**. **Table 4A** reports the results for the entire sample period and **Tables 4B,C** report the results for the contraction and expansion sub-periods, respectively. The adjusted R squares of both models are very similar where for model 1, the adjusted R square ranges between about 65% for

TABLE 4 | Coefficients of sector stock returns to variations in nominal interest rates (Model 1) and real interest and expected inflation rates (Model 2).


Model 1: rjt = α<sup>j</sup> + β<sup>j</sup> · rmt + γ<sup>j</sup> · 1i u <sup>t</sup> + εjt, Model 2: rjt = α<sup>j</sup> + βjm · rmt + βjr · 1r<sup>t</sup> + βj<sup>π</sup> · 1E ORT t πt,t+<sup>1</sup> + εjt.

rjt represents stock returns at time t for each sector j, rmt is the return on the market portfolio, 1i u t represents changes in nominal interest rates, 1r<sup>t</sup> represents changes in real interest rates, 1E ORT t πt,t+<sup>1</sup> shows movements in expected inflation rates (orthogonalized), and finally, ε<sup>t</sup> is the error term. The sample extends from Nov. 1989 to Feb. 2014 and the following regression has been estimated using SUR methodology. t-statistics in parentheses \*p < 0.10, \*\*p < 0.05, \*\*\*p < 0.01.

Information Technology, and about 24% for Utilities. All sectors exhibit a positive and significant market beta for both models overall and in the contraction and expansion sub-periods. While the betas are different in the contraction and expansion subperiods, there is no discernible pattern to these differences. The beta coefficients are nearly the same by sector for the two models. For the overall period, the beta coefficients vary between the least risky Utilities 0.47 to the most risky Information Technology sector 1.38.

Looking at model 1 for the overall sample period, the results confirms a noteworthy relationship between sector stock returns and unexpected changes in nominal interest rates as 6 of the 10 sectors have a statistically significant coefficient. Interestingly, the sign of this relationship is not always negative. Consumer Staples, Health Care, and Utilities are conventionally negative but Energy and Materials are marginally positive and Information Technology is significantly positive. Clearly, the positive coefficient for Information Technology is not due to mere chance. Moreover, the relation between stock returns and unexpected changes in nominal rates for the Information Technology sector remains significantly positive for the contraction and expansion sub-periods. This suggests that investors who seek protection from unanticipated interest rate changes can view an investment in a portfolio of Information Technology stocks as a natural hedge against interest rate risk.

Meanwhile, the conventionally negative relation between stock returns and unexpected changes in nominal interest rates for Consumer Staples, Health Care, and Utilities remain negative for the recession and expansion sub-periods but only the Consumer Staples coefficient remains highly significant in both sub-periods. Clearly, an investment in the Consumer Staples sector is subject to a significant amount of interest rate risk. Finally, there are two sectors without any significant relation between stock returns and unexpected changes in nominal interest rates for the entire sample but show significant coefficients, with opposite signs, for the contraction and expansion sub-periods. Specifically, Consumer Discretionary, and Industrials have the conventional inverse relation during contraction which turns positive during expansion suggesting that firms in these industries can pass on additional financing costs when economic conditions are robust.

When decomposing unexpected changes in the nominal rate of interest into unexpected changes in the real rate of interest and unexpected changes in the inflation rate (model 2), we discover comparable results for unanticipated changes in the real rate of interest but in this case, there are just four rather than six sectors that are statistically significant. Consumer Staples and Utilities have a significant inverse relation between stock returns and unexpected changes in the real rate of interest whereas Energy and Materials have a significant positive relation. However, none of these relations remains consistently significant and of the same sign for the contraction and expansion sub-periods with the exception of Energy. Even then the positive coefficient in the expansion sub-period is only marginally significant.

Similarly, the signs of the relation between stock returns and unanticipated inflation are not always negative. Specifically, we find significant negative coefficients for Consumer Staples, Health Care, and Utilities and one positive relation for Information Technology. However, only Consumer Staples has a consistent inverse relation for both economic sub-periods suggesting that unexpected changes in inflation are an important source of risk for investments in the Consumer Staples sector. Interestingly, stock returns in the Industrials sector are directly related to unanticipated inflation in expansion sub-period but are inversely related to unanticipated inflation in contraction subperiod suggesting that firms in this sector can pass on unexpected inflationary costs during robust economic conditions but are less able to do so during harder economic times.

In summary, we find that when there are significant relations between stock returns and unanticipated changes in nominal interest rates and their components, unanticipated changes in the real rate of interest and inflation, these relations are most commonly negative. The Consumer Staples industry sector shows this tendency most strongly as the relation between stock returns and unanticipated changes in the nominal interest rate as well as unanticipated changes in the inflation rate are significantly negative overall and in the contraction and expansion subperiods. Even the relation between stock returns and unexpected changes in the real rate is negative but significantly so only for the contraction sub-period. Meanwhile we observe the contrary positive relation more rarely. The clearest example is the Information Technology sector. Specifically, while all the significant relations between stock returns in the Information Technology sector and unanticipated changes in nominal interest rate, real rate and inflation rate are always positive, they are consistently and significantly positive overall and in the in the contraction and expansion sub-periods only for unexpected changes in the nominal rate of interest. The next step is to see if we can discover more instances of these significant relationships as we further refine our analysis by examining more refined sub sector portfolios.

#### Results at the Sub-Sector Level

In the second step of our analysis, we estimate model 1 and 2 at the sub-sector level as defined in Supplementary Material Table A. **Table 5** shows the number and percentage of subsectors that have a significant response of stock returns to unanticipated changes in each factor (nominal interest, real interest and inflation rate) and the average significant coefficient and the average positive and negative coefficients for each factor. **Table 5A** shows this information for the entire sample period while **Tables 5B,C** report this information for the contraction and expansion sub-periods, respectively.

For both models, we find a positive and significant market beta for the total sample and for the expansion and contraction sub-periods for all sub-sectors with just one exception. There are <100% sub-sectors with a statistically significant positive market beta during the expansion sub-period because the beta for Construction Materials, while positive, is statistically insignificant. The average beta is close to the theoretical beta of 1, being a little higher in the contraction sub-period and a little lower in the expansion sub-period. For the overall period, betas range between about 0.4 for Electric Utilities and 1.5 for Semiconductors and Semiconductor Equipment. For the sake of

TABLE 5 | Coefficients of sub-sector stock returns to variations in nominal interest rates (model 1) and real interest and expected inflation rates (model 2): % of significant exposure.



Average Ad. R<sup>2</sup> = 45.21%

#### MODEL 2


Average Ad. R <sup>2</sup> = 45.18%

Total number of sub-sectors = 33

#### (B) CONTRACTION SUB-PERIOD MODEL 1


#### MODEL 2


(Continued)

#### TABLE 5 | Continued


Average Ad. R <sup>2</sup> = 58.32%

Total number of sub-sectors = 33

#### (C) EXPANSION SUB-PERIOD

MODEL 1


Average Ad. R <sup>2</sup> = 39.08%

#### MODEL 2


Total number of sub-sectors = 33

Model 1: rjt = α<sup>j</sup> + β<sup>j</sup> · rmt + γ<sup>j</sup> · 1i u <sup>t</sup> + εjt; Model 2: rjt = α<sup>j</sup> + βjm · rmt + βjr · 1r<sup>t</sup> + βj<sup>π</sup> · 1E ORT t πt,t+<sup>1</sup> + εjt.

rjt represents stock returns at time t for each industry j, rmt is the return on the market portfolio, 1i u t represents changes in nominal interest rates, 1r<sup>t</sup> represents changes in real interest rates, 1E ORT t πt,t+<sup>1</sup> shows movements in expected inflation rates (orthogonalized) and, finally, ε<sup>t</sup> is the error term. The sample extends from Nov. 1989 to Feb. 2014 and the following regression has been estimated using SUR methodology. t-statistics in parentheses \* p < 0.10, \*\* p < 0.05, \*\*\* p < 0.01.

brevity, we do not report the coefficients for each of the 33 subsectors. They are available from the corresponding author upon request.

The average significant sub-sector coefficients, along with the average of the significant positive and negative coefficients are shown in column 3 of **Table 5**. The average relation between stock returns and unexpected changes in the nominal interest rate (model 1) and unexpected changes in the real interest and inflation rates (model 2) are negative for the overall period and for the contraction and expansion sub-periods with just one exception. Specifically, in **Table 5C** the average coefficient for unexpected changes in the real rate of interest is a positive 0.483 for the expansion sub-period. Moreover, when a coefficient is significant, it is most often negative, again except for the expansion sub-period for unexpected changes in the real rate of interest. Specifically, **Table 5C**, column 2 shows that five of the seven sub-sectors have a significant positive relation between stock returns and unexpected changes in the real rate of interest.

Clearly, the overall results are consistent with most of the prior literature as the relations between stock returns and unexpected changes in the rate of inflation are most often negative. Specifically, column 2 shows that around 42, 33, and 27% of the sub-sectors for the total sample, contraction, and expansion sub-periods, respectively, have stock returns that are significantly and negatively related to unexpected changes in the inflation rate.

Nevertheless, there are some exceptions to the conventionally inverse relations. For instance, **Table 5A**, column 2 reports that there are six sub-sectors that have a significant positive relation between stock returns and unexpected changes in the nominal rate of interest in the overall period. In addition, we find three contrary positive relations for unexpected changes in the real rate of interest and six contrary positive relations for unexpected changes in inflation rate for the overall sample period. Breaking down the results by sub-period, we observe that with a smaller sample size, there are fewer statistically significant coefficients. During the contraction sub-period, there are more instances of inverse relations and during the expansion sub-period, there are proportionally more instances of positive relations suggesting that on average companies find it easier to pass on unexpected costs during expansions.

In summary, we find that on average, the relation between stock returns and unanticipated changes in the nominal rates of interest (model 1) and unanticipated changes in the real rate of interest and the inflation rate (model 2) are negative. This result is consistent with the literature. However, as we saw at the more aggregate sector level, we continue to find contrary positive relations at the sub-sector level. This motivates us to examine individual industries to see if we can find exceptional industries where investments in these industries can form a natural hedge against sources interest rate and inflation risk.

#### Results at an Industry Level

As a last step, we regress models 1 and 2 at the industry level. We again examine the relations for the total sample, contraction, and expansion sub-periods and obtain some remarkable results. **Tables 6A1,A2** shows the results for model 1 and 2, respectively, for the overall period and **Tables 6B1,B2,C1,C2** show the results for model 1 and 2 for the contraction and expansion subperiods, respectively. All panels present the information in the same way. For instance, **Table 6A1**, columns 2–5 show by sector the number of industries, the proportion that have a significant response to each factor and the number industries that have a positive and a negative response to each factor, respectively. Column 6 reports the average significant coefficient for the sector and the range of coefficient values by sector while columns 7 and 8 reports the size of the average positive and negative coefficients.

For both models in all six panels from A1 to C2, all industries exhibit positive and significant market betas for the overall sample and for the expansion and contraction sub-periods with just one interesting exception. While all of the industry betas during the contraction sub-period in the Materials sub-sector are positive for both models, **Tables 6B1,B2** show that only 11 of 12 industries have significant betas. The exceptional industry is Gold, long rumored to be an industry that can provide a safe haven during recessions.

Model 1 (in **Table 6A1**) reports that at the industry level, there are more instances of contrary positive relations between stock returns and unanticipated changes in nominal interest rates. In fact, the average industry weighted significant coefficient is positive for six of the nine sectors and only three have significant negative average coefficients. The Utilities industry sub-sector is not segmented into industries by Bloomberg so we can conduct our individual industry analysis only within the remaining nine industry sub-sectors. The sectors with the highest number of industries with significant coefficients are Consumer Staples with an average significant coefficient of −2.8 and Information Technology with an average significant coefficient of 3.7. Meanwhile in the Industrials sector, only 1 of 12 industries, namely Building Products, has a significant relation to unexpected changes in nominal interest rates with a coefficient of −4.15. Industries in the Energy sector exhibit the highest average significant response to unexpected changes in the nominal rate of interest (7.256) whereas industries within the Materials and Financials sectors are the least sensitive to unexpected changes in the nominal rate of interest at 0.399 and 0.578, respectively. The sectors with the most heterogeneous industries are Financials, Health Care and Materials as the range of significant coefficients is very large. In contrast, the most homogeneous industry is Consumer Staples where six of nine industries have a significant negative relation between stock returns and unanticipated changes in nominal interest rates.

In **Table 6B1** we observe that the results of the contraction sub-period is mostly similar to the total sample but with a few peculiarities. First, the stock returns of more industries are inversely related to unanticipated changes in nominal interest rates to the point where Consumer Discretionary and Financials sectors now join the previous three industry sectors that have an average negative significant coefficient. Second, three more industries in the Energy sector and the Industrials sector now have a significant relation between stock returns and unanticipated changes in nominal interest rates. However, the three additional industries for the Industrials sector are all negatively and the three industries for the Energy sector are all positively related to unanticipated changes in the nominal rate of interest. Meanwhile, it is remarkable that two fewer industries in the Information Technology sector exhibit a significant relation between stock returns and nominal interest rate changes in the contraction sub-period and for the remaining significant Information Technology industries, the coefficients become more positive. Third, in general, we observe that stock returns are more responsive to unexpected changes in nominal interest

TABLE 6 | Coefficients of industry stock returns to variations in nominal interest rates (model 1) and real interest and expected inflation rates (model 2): Significant industry sensitivity.



#### (A2) MODEL 2 TOTAL SAMPLE (FROM NOV. 1989 TO FEB. 2014)

Average Ad. R <sup>2</sup> = 40.81%




Average Ad. R <sup>2</sup> = 40.82%




Average Ad. R<sup>2</sup> = 54.95%

#### (B2) MODEL 2 CONTRACTION SUB-PERIOD

Model 2 Industries with signific. 10% Average Coeff. *rmt* Nr. Ind SignifCoeff. Posit. Coeff. Negat. Coeff. Signif. Coeff. (range) Posit. Coeff. Negat. Coeff. Industries of S1 consum. discretionary 16 16/16 16 0 1.310 (0.437, 2.124) 1.310 na Industries of S2 consumer staples 9 9/9 9 0 0.684 (0.447, 1.389) 0.684 na Industries of S3 energy 7 7/7 7 0 0.965 (0.600, 1.287) 0.965 na Industries of S4 financials 11 11/11 11 0 1.602 (0.644, 2.536) 1.602 na Industries of S5 health care 5 5/5 5 0 0.664 (0.410, 0.855) 0.664 na Industries of S6 industrials 12 12/12 12 0 1.306 (0.906, 2.134) 1.306 na Industries of S7 inform. technology 9 9/9 9 0 1.414 (0.962, 1.816) 1.414 na Industries of S8 materials 12 11/12 11 0 1.402 (0.719, 2.136) 1.402 na Industries of S9 telecommunications 1 1/1 1 0 0.562 (0.562, 0.562) 0.562 na Industries of S10 utilities 0 na na na na na na Total number of industries 82 81 81 0




Average Ad. R

<sup>2</sup> = 54.88%

#### (C1) MODEL 1 EXPANSION SUB-PERIOD







Model 1: rjt = α<sup>j</sup> + β<sup>j</sup> · rmt + γ<sup>j</sup> · 1i u <sup>t</sup> + εjt; Model 2: rjt = α<sup>j</sup> + βjm · rmt + βjr · 1r<sup>t</sup> + βj<sup>π</sup> · 1E ORT t πt,t+<sup>1</sup> + εjt.

rjt represents stock returns at time t for each industry j, rmt is the return on the market portfolio, 1i u t represents changes in nominal interest rates, 1r<sup>t</sup> represents changes in real interest rates, 1E ORT t πt,t+<sup>1</sup> shows movements in expected inflation rates (orthogonalized) and, finally, ε<sup>t</sup> is the error term. The sample extends from Nov. 1989 to Feb. 2014 and the following regression has been estimated using SUR methodology. t-statistics in parentheses \* p < 0.10, \*\* p < 0.05, \*\*\* p < 0.01.

rates, irrespective of the sign, during the contraction sub-period. Moreover, it is noteworthy that Health Care and Materials are again the sectors with the most heterogeneous response to unexpected changes in the nominal rate of interest and the range of significant values are even larger during the contraction sub-period.

In the expansion sub-period reported in **Table 6C1**, we find that compared to the overall sample, there are four fewer industries that have a significant relation between stock returns and unexpected changes in nominal interest rate. In fact, the Energy and Telecommunication sectors do not have even one industry that has a significant relation between stock returns and unexpected change in nominal interest rates. The range of significant coefficients is typically smaller as well. In terms of absolute values of the coefficients, stock returns of industries in the Financial sectors have the largest average response (nearly 5) to unexpected changes in the nominal rate of interest whereas industries in the Materials sector have the lowest average response (nearly 1.6) to unexpected changes in the nominal rate of interest.

There is an interest phenomenon contained within these results. Stock returns for the Diversified Metals and Mining industry (within the Materials sector) have a positive and significant relation between stock returns and unexpected changes in nominal interest rates for the overall, contraction and expansion sub-periods. This suggests that an investment in these industries can form a natural safe haven against unexpected changes in the nominal interest rate.

Meanwhile, Model 2 **Tables 6A2,B2,C2** provides the following interesting observations. First, the stock returns of most industries have no significant relation with unexpected changes in the real rate of interest. For instance, in the overall period, only 16 of 82 industries have a significant coefficient and independent of the sample period, the stock returns of all industries in the Health Care sector does not have a significant relation to unexpected changes in the real rate of interest. There are a few more industries with a significant relation between stock returns and unexpected changes in the real rate of interest in the contraction sub-period and a few less in the expansion sub-period, 21 and 11, respectively. Clearly, the stock returns of most industries do not respond to unexpected changes in the real rate of interest.

However, within these general results we find three industries, one each in the Energy, Industrials and Materials sectors, have a consistently significant, and positive, relation between stock returns and unexpected changes in the real rate of interest. Specifically, we find that stock returns in the Integrated Oil and Gas, Commercial Services, and Supplies and Diversified Metals and Mining industries have a consistently positive relation with unexpected changes in the real rate of interest for the overall, contraction, and expansion sub-periods. This suggests that investors can find that an investment in these industries can provide some insulation from unexpected changes in the real rate of interest.

We find that the stock returns of many industries respond to unexpected changes in the inflation rate. Overall, 23 of 82 industries respond significantly to unexpected changes in inflation, while during the contraction sub-period the number of significant relations rises to 30 and during expansion the number of significant relations falls only slightly to 21. The stock returns for industries in the Energy sector exhibit the highest average response to unexpected changes in the inflation rate for the total sample (7.19) and contraction sub-period (12.58) whereas firms in the Financial sectors have the highest average response in the expansion sub-period (4.65). In contrast, industries in the Consumer Discretionary sector have the lowest average response to unexpected changes in the inflation rate for the total sample (−0.29) while industries in the Materials sector have the lowest response in the contraction and the expansion sub-periods, −2.46 and 0.37, respectively.

On average, the majority of sectors, most notably Consumer Discretionary, Consumer Staples, Financials, Industrials, and Materials, have an industry weighted negative significant relation between stock returns and unexpected inflation. While overall, and in some of the sub-periods, we can find industries with a significant positive relation between stock returns and unexpected inflation, we are unable to find an industry that has a consistently positive relation with unexpected inflation. However, we do find that stock returns in the Household Durables, Pharmaceuticals, and Gold industries have a negative relation to unanticipated inflation in the overall sample and in the contraction and expansion sub-periods suggesting that stocks in these industries are exposed to significant inflation risk.

### Overall Results

As mentioned previously, according to most of literature, the response of stock returns to changes in nominal and real interest rates is usually negative. Our results generally agree with these previous findings. Also, like Booth and Officer (1985), Bae (1990), Jareño (2008), Ferrando et al. (2015), and Jareño et al. (2016), we find that some financial (as well as non-financial) sectors have insignificant relations. However, we also find some contrary results when examining the relations by sector, subsector, and industry. We find that three industries, specifically Integrated Oil and Gas, Commercial Services and Supplies, and Diversified Metals and Mining have a consistent significant positive relation between stock returns and unexpected changes in real interest rates while one industry, Diversified Metals and Mining, has a significant consistently positive relation between stock returns and unexpected changes in nominal interest rates. These positive relations suggest that long investments in portfolios of stocks in these particular industries can form a safe haven from unanticipated changes in nominal and real interest rates. Moreover, we find that Gold has an insignificant beta during recessionary conditions hinting that investments in the Gold industry can indeed be a safe haven during recessions. Interestingly, we find that three industries, specifically Household Durables, Pharmaceuticals, and Gold have a negative relation to unanticipated inflation in the overall sample and in the contraction and expansion sub-periods suggesting that these three industries are particularly exposed to inflation risk. It is remarkable that stock returns are inversely related to unexpected inflation for the Gold industry, thereby damaging the image of Gold as a hedge against inflation. Therefore, investor behavior seems to be quite different over time (according to the business cycle) and by sector.

### DISCUSSION

Many studies have analyzed the sensitivity of stock returns to changes in nominal interest rates (Sweeney and Warga, 1986; Hevert et al., 1998a,b; O'Neal, 1998; Oertmann et al., 2000; Fraser et al., 2002; Tessaromatis, 2003; Jareño, 2006, 2008; Ferrer et al., 2010), finding a negative and significant relationship between stock returns and unexpected changes in nominal interest rates. We too examine this relationship but at the sector, sub-sector, and industry levels for both contraction and expansion sub-periods as well as for the overall sample period. In general, we find significant and negative relationship between stock returns and unexpected changes in nominal interest rates. Nevertheless, we observe important exceptions where some of these relations are insignificant and other relations that are consistently positive, even at the level of an individual industry in the case of the Diversified Metals and Mining industry.

At the sector level, we find that the most vulnerable sector to fluctuations in 10-year government bond yields are Utilities, so regulated and seriously indebted sectors seem to be the most interest rate sensitive, particularly in the expansion sub-period. Also, we note that Consumer Discretionary and Industrials have the conventional inverse relation between stock returns and unanticipated changes in the nominal rate of interest during the contraction sub-period that turns positive during the expansion sub-period so that for the overall period, there is no significant relation. This suggests that firms in these industries can pass on additional financing costs when economic conditions are robust.

In order to deepen in our analysis, we decompose unexpected changes in the nominal interest rate into unexpected changes in the real interest and inflation rates. In general, the stock returns by sector, sub-sector and industry are inversely related to unexpected changes in the real interest rate movements, and unexpected changes in the inflation rate overall and more so in the contraction than in expansion sub-period. However, it is unusual to find industries with a consistent negative relation between stock returns and unanticipated changes in the real interest rate and the inflation rate. There are three exceptions however. Evidently, inflation is an important source of risk for investments in Household Durables, Pharmaceuticals and Gold industries as they have a negative relation to unanticipated inflation in the overall sample and in the contraction and expansion sub-periods.

It is remarkable that stock returns are inversely related to unexpected inflation for the Gold industry, thereby damaging the image of Gold as a hedge against inflation. Another interesting result is that the stock returns in the Gold industry are not significantly related to the market return during contraction economic sub-periods thereby bolstering Gold's reputation as a safe haven during recessionary conditions.

Interestingly, we find that investments in three industries, specifically Integrated Oil and Gas, Commercial Services and Supplies, and Diversified Metals and Mining can provide a safe

#### REFERENCES


haven against unexpected changes in the real rate of interest. Specifically, we find that the stock returns in these industries have a consistently positive relation with unexpected changes in the real rate of interest for the overall, contraction and expansion sub-periods. This suggests that investments in these industries will tend to increase if real rates of interest unexpectedly rise, thereby offsetting extra costs associated with a rise in the real rate of interest.

Our empirical results support the state-dependent nature of the investor behavior in the interest rate sensitivity analysis. Also, this study may find a herding behavior of investors in some scenarios, because in certain times of market stress, investors disregard their own information, and exhibit herding behavior, which is often extremely optimistic or pessimistic and may lead to an unreasonable reaction to movements in interest rates. Finally, we confirm the null hypothesis that investor behavior may depend on different factors that affect the investment or trading decision. Therefore, aspects such as the sector that traded stock belongs to and the business cycle definitely impact on investment behavior.

#### AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication.

#### ACKNOWLEDGMENTS

This research is funded by "Fundación Ramón Areces (Convocatoria Pública año 2013), Riesgo de interés e inflación: estudio del mercado bursátil norteamericano" grant.

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fpsyg. 2016.00390


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 González, Jareño and Skinner. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Attitudes vs. Purchase Behaviors as Experienced Dissonance: The Roles of Knowledge and Consumer Orientations in Organic Market

#### María Hidalgo-Baz <sup>1</sup> \*, Mercedes Martos-Partal <sup>2</sup> and Óscar González-Benito<sup>2</sup>

<sup>1</sup> Department of Business Economics and Administration, University of Salamanca, Salamanca, Spain, <sup>2</sup> IME Instituto Multidisciplinar de Empresa (Multidisciplinary Institute for Enterprise), University of Salamanca, Salamanca, Spain

This research focuses on the incongruity between positive attitudinal responses but a lack of purchase behavior in organic markets. According to cognitive dissonance theory, consumer orientations toward the benefits attributed to organic products (environmental protection, health, hedonic) relieve the dissonance that results from this attitude–behavior incongruity. Knowledge also functions as a transmitter, from positive attitudes to purchase behaviors, thereby reducing the incongruity. Using quota sampling in a survey study, this paper tests the hypotheses from linear regression models. The results show that orientations and knowledge improve the congruity between attitudes and purchase behaviors toward organic products. Moreover, interaction effects arise between the environmental protection orientation and knowledge and between the hedonic orientation and knowledge. Increasing knowledge mitigates the difference between attitudes and purchase behaviors, especially for consumers with environmental protection or hedonic orientations. These findings have several important implications for research and practice.

Keywords: organic products, environmental protection orientation, health orientation, hedonic orientation, organic market knowledge, cognitive dissonance theory

## INTRODUCTION

Health concerns and environmental protection are increasingly important societal issues (Bhattacharya and Sen, 2004; Moisander, 2007; Groening et al., 2009; Pagiaslis and Krystallis-Krontalis, 2014; Nielsen, 2015), leading to the developments of green and organic markets. Whereas green markets focus on social and environmental responsibilities (Akehurst et al., 2012), organic markets address broader consumer concerns for health, environmental protection, and food safety by relying on agricultural systems that are free of human-made chemicals (Pino et al., 2012). Although related, these markets differ, in that the organic market entails not just environmental issues but also health and food safety concerns. With its focus on health and environmental protection (Zanoli and Naspetti, 2002; Baker et al., 2004; Essoussi and Zahaf, 2008; Bauer et al., 2013), the organic market also has experienced huge growth (Kareklas et al., 2014), transforming from a niche to a central product trend in the food industry (Van Doorn and Verhoef, 2015). Global sales of organic food and beverages reached US\$72 billion in 2013 (Willer and Lernoud, 2015), a nearly 5-fold increase over 1999 sales (US\$15.2 billion), and then expanded even further to

#### Edited by:

Maria Pilar Martinez-Ruiz, University of Castilla–La Mancha, Spain

#### Reviewed by:

Carlos Flavian, University of Zaragoza, Spain Salvador Ruiz De Maya, University of Murcia, Spain

#### \*Correspondence:

María Hidalgo-Baz marhibaz@usal.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 25 May 2016 Accepted: 08 February 2017 Published: 24 February 2017

#### Citation:

Hidalgo-Baz M, Martos-Partal M and González-Benito Ó (2017) Attitudes vs. Purchase Behaviors as Experienced Dissonance: The Roles of Knowledge and Consumer Orientations in Organic Market. Front. Psychol. 8:248. doi: 10.3389/fpsyg.2017.00248 US\$80 billion in 2014 (IFOAM – Organics International, 2016). In Spain, expenditures on organic products reached 1018 million euros in 2013, equivalent to a 5.5% increase in sales compared with 2011 (Prodescon, 2014).

Parallel with this growing social interest and rapid sales expansions, academic research into the organic market has increased as well (Kim and Chung, 2011). Hughner et al. (2007) classify consumer responses to organic products, according to their consideration in previous literature. For example, some studies address factors that facilitate or encourage these responses, whereas others focus on inhibiting factors. Among the former, most studies concentrate on motivations or consumer orientations related to environmental protection and health (Borin et al., 2011; Kim and Chung, 2011; Bauer et al., 2013), as well as prosocial or altruistic values, together with hedonic or selfbenefiting values (Cornelissen et al., 2008; Urien and Kilbourne, 2011; Yang et al., 2015). Among the latter, negative factors, we find studies of price, consumer confidence (Bhaskaran et al., 2006; Terrachoice Environmental Marketing, 2009; Gleim et al., 2013), and ineffective marketing (Krystallis et al., 2006; Hughner et al., 2007; Aertsens et al., 2009; Ngobo, 2011). Such factors all can influence consumers' attitudes toward and purchases of organic products (Tsakiridou et al., 2008; Ngobo, 2011; Akehurst et al., 2012; Bickart and Ruth, 2012; Tucker et al., 2012; Atkinson and Rosenthal, 2014).

However, attitudes toward organic products appear to differ from purchase behavior in this market. That is, sustainable consumption and healthy eating remain top priorities among modern consumers (Pagiaslis and Krystallis-Krontalis, 2014; Nielsen, 2015), yet an attitude–behavior gap or values–action gap arises, such that consumer express environmental concerns, but those concerns do not translate into purchase behaviors (Akehurst et al., 2012). Therefore, understanding the behaviors of organic consumers, the antecedents of organic consumption, and the incongruity between attitudes and behavior is critical. According to Akehurst et al. (2012), who study the difference between green purchase intentions and green purchase behaviors, this gap is less evident for consumers with high ecological consciousness.

The current research in turn investigates the incongruity between consumer responses, in the form of attitudes, and their behaviors, in the form of purchases, of organic products. We focus on the organic food market instead of the green market. Using cognitive dissonance theory (Festinger, 1957), we study the incongruity by analyzing the role of consumer orientation in relation to consumer behavior (Kim and Chung, 2011; Akehurst et al., 2012; Izagirre-Olaizola et al., 2013; Van Doorn and Verhoef, 2015). Specifically, we consider the potential effects of environmental protection, health, and hedonic (i.e., taste and satiety) orientations, together with consumer knowledge about organic products.

In the next section, we review previous literature and develop our hypotheses. After detailing our methodology and analysis, we present the study results. Finally, we discuss some conclusions, implications, and limitations of this study, as well as ideas for further research.

### THEORETICAL FRAMEWORK

### Effect of Knowledge on Attitude–Behavior Incongruity in Organic Markets

Even when consumers state very positive attitudes toward organic or green products, they frequently exhibit incongruous behaviors and fail to purchase these products. That is, a positive attitude does not translate into a purchase (D'Souza et al., 2007; Pickett-Baker and Ozaki, 2008; Florenthal and Arling, 2011; Moraes et al., 2012; Gleim et al., 2013). Thus, organic market is characterized by an attitude-behavior incongruity. Ongoing studies seek to explain this incongruity, using a variety of factors. For example, organic food consumption might be barred by high prices, lack of consumer confidence (Bhaskaran et al., 2006; Terrachoice Environmental Marketing, 2009; Gleim et al., 2013), or lack of consumer knowledge about these products. A lack of knowledge makes it difficult for consumers to differentiate organic from conventional products (Chryssochoidis, 2000; Gleim et al., 2013; Gfk, 2014). Thus, increased knowledge might be a key factor that would encourage organic purchase behavior. Aertsens et al. (2009) also note that providing more information or increasing awareness of organic products can help lower consumers' uncertainty about the unique attributes of organic offerings, as well as mitigate their lack of confidence about certification methods. Such reduced uncertainty then might improve purchase likelihood (Thøgersen, 2007).

Overall, if more knowledge about organic products influences consumers' decisions and increases their willingness to pay (Barnes et al., 2009), it should have a positive effect on organic product purchases, while also attenuating other consumption barriers, such as a lack of confidence and high prices (Aertsens et al., 2009; Barnes et al., 2009). Knowledge thus might function as a transmitter, from attitudes to purchase behaviors. Therefore,

H1. More knowledge about organic products leads to greater congruity between consumers' attitudes and purchase behaviors.

### Cognitive Dissonance

According to cognitive dissonance theory, each person maintains a cognitive view of him- or herself, past behaviors, beliefs, attitudes, and environments (Oshikawa, 1968). Elements of this view might become dissonant if they are inconsistent or contradict each other. In an organic market setting for example, consumers express positive attitudes toward organic products but do not buy them, so they might experience dissonance between their own attitudes and behaviors. Therefore, we define the incongruity characterizing organic market as a dissonance arised from contradictory responses that consumers state. Cognitive dissonance theory suggests that such inconsistencies generate a disturbing, unpleasant sensation for the consumer, who then tries to avoid or prevent the inconsistency (Festinger, 1957).

Dickerson et al. (1992) argue that this sensation of dissonance can result from hypocrisy, due to a discrepancy between actual behaviors and norms for what people should do to benefit the environment, according to their own beliefs, concerns, or orientations. Nonetheless, when faced with an incongruity between their attitudes and purchase behaviors in the organic market, consumers likely seek to modify the dissonant elements (Oshikawa, 1968), in accordance with their concerns or orientations. In fact, Cornelissen et al. (2008) point out that previous behavior of a consumer is used as a heuristic basis for later decisions. Therefore, consumer may change their purchase behavior toward organic products, rather than their attitudes. That way, consumers may show different orientations in their current behavior, based on their perception about their previous behavior. In this respect, Becker et al. (1977) show a health orientation leads people to engage in healthy behaviors; and Schlegelmilch et al. (1996) show an environmental orientation prompts them to make green decisions.

More broadly, orientations related to the attributes and benefits of organic products should lead consumers to relax or correct the cognitive dissonance they experience, due to the difference between their attitudes and purchase behaviors, by increasing their purchase responses. In organic markets, the benefits associated with the products are mainly environmental protection and health (Essoussi and Zahaf, 2008; Kareklas et al., 2014). Relative to conventional products, organic products generally are perceived as offering more nutritional value and being produced in a more natural way, without chemicals or harmful pesticides (Ott, 1990; Wilkins and Hillers, 1994; Wandel and Bugge, 1997; Squires et al., 2001; Pino et al., 2012). In this sense, organic products also are assumed to be more environmentally friendly (Wilkins and Hillers, 1994; Hughner et al., 2007).

Consumer orientations related to these organic benefits (i.e., environmental protection and health) therefore should have a positive impact on the purchases of organic products, thereby reducing the difference between attitudes and purchase behaviors. Akehurst et al. (2012) concur that the gap between purchase intentions and purchases is less evident for green products when consumers' environmental consciousness is high. Thus, we propose:

H2a. Consumers' higher environmental orientation leads to greater congruity between their attitudes and purchase behaviors toward organic products.

H2b. Consumers' higher health orientation leads to greater congruity between their attitudes and purchase behaviors toward organic products.

We also consider hedonic orientations, because previous research indicates that consumers perceive organic products as tastier and offering better visual appearances and scent (McEachern and McClean, 2002; Cervellon and Carey, 2014). For example, McEachern and McClean (2002) link perceptions of better flavor to the increased safety associated with organic food and cite these notions as the primary reasons consumers buy organic products. Cervellon and Carey (2014) also note that consumers consider the hedonic attributes of organic food, such as their visual appearance, scent, and texture, more positively in their post-purchase assessments. Therefore, consumers with a more hedonic orientation might be more consistent in their attitudes and purchase behaviors toward organic products, such that they may experience less dissonance. In further support of this prediction, Van Doorn and Verhoef (2015) show that consumers oriented toward product quality and taste are less concerned about prices. Therefore, the negative effect of the price premium on the purchase of organic food may be weaker for consumers concerned about the quality and taste of the food. This preference and orientation can help overcome the barriers of organic consumption and facilitate the translation of positive attitudes into purchases. Therefore,

H2c. Consumers' higher hedonic orientation leads to greater congruity between their attitudes and purchase behaviors toward organic products.

## Moderating Effect of Knowledge

Previous literature has analyzed environmental concerns and knowledge about organic products as factors that might explain organic or green purchase behavior (Kollmuss and Agyeman, 2002; Mostafa, 2007; Pagiaslis and Krystallis-Krontalis, 2014). For example, Pagiaslis and Krystallis-Krontalis (2014) propose a mediation relationship, following a sequence of orientation– knowledge and belief–behavior, such that consumers who are more oriented toward environmental protection are also more informed and have more positive beliefs about green products. Therefore, knowledge and beliefs may be necessary for purchases of these products to take place.

We propose that knowledge might be a moderator, as well as a mediator, in these relationships. Beyond organic literature, knowledge is considered as moderator in the relationship attitude-behavior. Specifically, Berger et al. (1994) take into account that moderating variable in the relationship between attitude and ecological behavior by studying different kinds of heating systems. Those authors note knowledge increases the attitude strength and, consequently, the effect of the attitude on behavior will be greater. In fact, subjective knowledge will be an important indicator in high involvement, high risk, search product categories. Information is collected over time for those cases. So, attitude would be stronger as subjective knowledge increases and, accordingly, its effect on behavior.

We have predicted that the dissonance experienced as a result of incongruity between attitudes and behaviors might be lower among consumers with more knowledge, such that knowledge might overcome some of the barriers to the consumption of organic products. In this sense, it could facilitate the transformation from positive attitudes to purchases of organic food. Nonetheless, we propose knowledge as moderator in the relationship between orientations and attitude-behavior congruence as well. We expect that more informed consumers, who know the attributes and benefits of organic products, respond in ways that are more consistent if their orientations also are linked to these attributes and benefits. Consumers with more knowledge about organic food will buy even more if their orientations also are aligned with the benefits attributed to organic products, such that the difference between their attitudes toward organic food and their purchase behavior will be smaller. This reasoning implies interaction effects between knowledge and orientations, such that knowledge accentuates the positive effect of an environmental orientation, health orientation, or hedonic orientation on the congruence between attitudes and purchase behaviors toward organic food. In other words, knowledge positively moderates the relationship between consumers' orientations and the congruence between their attitudes and behavioral responses.

H3a. Consumers' environmental orientation exerts a stronger effect on the congruence between their attitudes and purchase behaviors when consumers have more knowledge of organic products.

H3b. Consumers' health orientation exerts a stronger effect on the congruence between their attitudes and purchase behaviors when consumers have more knowledge of organic products.

H3c. Consumers' hedonic orientation exerts a stronger effect on the congruence between their attitudes and purchase behaviors when consumers have more knowledge of organic products.

### METHODS

To test the hypotheses and obtain pertinent empirical evidence, we conducted a survey in an urban area in Castilla y Leon, Spain. A pretest prior to the main data collection ensured the comprehensibility of the items within the survey, as well as the appropriateness of the data collection procedure. Respondents reported to be responsible for or actively involved in purchasing food for their households. We ensured that the data came from a wide range of ages and both genders, using a quota sampling method. Finally, after studying the existence of outliers, we use data from 305 (out of 311 obtained initially) consumers that have been collected between April and June 2013.

The information requested in the survey refers to the responses of consumers to organic food, the benefits sought, consumer orientations and their values, and socio-demographic characteristics. A seven-point Likert scale (0 = "strongly disagree" and 6 = "strongly agree") applied to all the items except for sociodemographic characteristics.

### Responses to the Organic Product

We consider three consumer responses: cognitive, affective, and conative. A cognitive response implies awareness of the existence of the object, retained information, and knowledge about an object. An affective response refers to the emotions a person feels relative to objects or events, such as a preference or dislike of a product or service. Finally, conative response implies the form of the reaction, such as a purchase (Lavidge and Steiner, 1961; Lambin, 1993; Peter and Olson, 2005). All consumer responses are built as means and scale reliability is assessed from Pearson Correlation and Cronbach's Alpha, as **Table 1** shows. Those coefficients confirm the suitability of our measures about consumer responses. Additionally, we check construct validity using Confirmatory Factor Analysis (CFA) that shows a good model fit (χ 2 statistic: 10.608 (0.101); CFI: 0.994; RMSEA: 0.050).

The knowledge variable measures cognitive responses. We use items adapted from scale of Pagiaslis and Krystallis-Krontalis (2014) about subjective knowledge. So, our variable is constructed as the average of two items that refer to consumer knowledge about the attributes of organic products and the differentiation from their counterpart, conventional products.

Attitude toward organic foods constitutes the measure of affective responses; it is constructed as the average of two items pertaining to overall assessments of organic products and consumer preferences. This measure is congruent with the attitude used by Ajzen and Madden (1986). Those authors define specifically the attitude as an overall assessment related to different consumer beliefs about a certain object or product.

We also calculate the average of two items to measure the conative response, as reflected in the purchase variable. Likewise, our measure is congruent with the measure proposed by Ajzen and Madden (1986) dealing with the likelihood of certain behavior. In this case, the two items pertain to the consumer's purchase and intention to consume organic products.

The "purchase deducting attitude" (PDA) variable is the difference resulting when we subtract the attitude variable from the purchase variable. Thus, it reflects the level of congruence between purchase behaviors and attitudes. In our sample, all respondents indicated more positive attitudes than purchase behaviors, such that the PDA values range from −6 to 0. Consumers with a positive attitude who buy organic products frequently (i.e., exhibit congruent responses) take values near 0. But if consumers indicate more incongruent responses, the PDA values are more negative and farther away from 0.

### Consumer Orientations

We used six items adapted from the "Self-perception" scale of Cornelissen et al. (2008) to measure consumer orientations. Hence, we consider orientations as self-perception of the consumer about their past behavior related to the different benefits of organic product. Specifically, we extracted three orientations: environmental orientation, health orientation, and hedonic orientation. Every orientation is calculated as the average of two items about self-perception of consumer related to either proenvironmental behavior or health behavior or hedonic behavior, depending on the orientation. Therefore, those measures are assessed similarly to consumer responses, scale reliability is measure with Pearson Correlation and Cronbach's Alpha. The statistical results for this assessment are in **Table 2**. Although those Cronbach's Alphas are lower than the common rule of 0.7 some update research point out that "there is no universal minimally acceptable reliability value. An acceptable reliability value depends on the type of application, and furthermore, the focus should be on the population reliability value and not on the sample reliability value" (Bonett and Wright, 2015, p. 4). Similarly to consumer responses, we check construct validity using CFA and the model fit is good (χ 2 statistic: 7.367 (0.288); CFI: 0.994; RMSEA: 0.027).

### Consumer Characteristics

The survey also gathered demographic profiles, reflecting the respondents' gender, age, household size, and presence of

#### TABLE 1 | Consumer responses to organic food.


#### TABLE 2 | Consumer orientations to organic food.


M, mean; SD, standard deviation; PC, pearson correlation/ \*\*p < 0.01.

children (younger than 6 years or 7–12 years old) in the house. Gender was a dichotomous variable (0 = male, 1 = female). Women represent 86.5% of the sample. Age is a count variable; 45.85 years is the average (SD = 12.06). Open questions assess both household size and the number of children. Household size is a count variable, with the following distribution: one-member households (19.6%), two members (27.7%), three members (28.3%), four members (21.2%), and five or more members (3.1%). Finally, 10.6% of respondents have children younger than 6 years, and 11.6% have children between 7 and 12 years of age in their households.

#### RESULTS

We use linear regression models to test the hypotheses. The results and their interpretations reflect widely accepted significance values (p < 0.05 and p < 0.01), though in some cases, we consider values of p < 0.10 as well. The following specification provides the test for our first hypothesis and second block of hypothesis:

$$\text{PDA} = \alpha + \Sigma \beta \text{ (CV)} + \gamma \text{Knowledge} + \Sigma \text{ or (O)} + \xi,\quad \text{(1)}$$

where PDA is the "purchase deducting attitude" variable, α is a constant we use to estimate the model, CV is the vector of the control variables in the study (gender, age, household size, number of children), β denotes the vector of parameters to estimate the effect of the control variables, γ is a parameter to estimate the effect of knowledge, O represents the vector of variables related to consumer orientations (environmental, health, and hedonic), σ denotes a vector of parameters we use to estimate the effect of the variables related to those consumer orientations, and ξ refers to the error term of the model.

The estimation results of our tests of H1 and H2a–c are in the first column of the **Table 3**. The effect of knowledge on PDA is positive and significant at a confidence level greater than 95%, in support of H1. Higher knowledge leads to greater congruity in consumer responses toward organic products; consumers who are more familiar with organic products buy more of them. In turn, the difference between their attitudes and purchase behaviors shrinks. According to the coefficients in **Table 3**, an environmental orientation has a positive effect on PDA too (95% confidence level). For the health orientation and hedonic orientation, there is no significative effect. Nevertheless, they are congruent with our hypothesis. Although no signiticative, their effects are positive. In conclusion, these results offer only support for H2a. Consumers oriented toward the environment, which also is the main benefit gained from organic food, have more consistent responses to organic products.

To test H3a–c, we use the following specification:

$$\begin{split} \text{PDA} &= \alpha + \Sigma \,\beta \text{ (VC)} + \text{y Knowledge} + \Sigma \,\sigma \,\text{(O)}\\ &+ \Sigma \,\lambda \text{ (Knowledge} \times \text{O)} + \xi, \end{split} \tag{2}$$



<sup>+</sup>p < 0.10, \*p < 0.05, \*\*p < 0.01.

where (Knowledge × O) denotes the vector of variables reflecting the interactions between knowledge and consumer orientation, and λ is the vector of parameters we use to estimate the effect of these interactions.

The estimation results for H3 are in the second column of the **Table 3**. Firstly, we check changes for coefficient of determination (R 2 ) between a regression model about direct effects only (first column of **Table 3**) and a regression model including interaction effects (second column of **Table 3**). Comparing their R<sup>2</sup> , it proves that interaction effects get to improve the goodness of model fit for the two presented regressions on the table. As **Table 3** shows, both the knowledge–environmental orientation and the knowledge–health orientation interactions are significant (95 and 90% confidence levels, respectively). But, those results show only support for knowledge–environmental orientation interaction (H3a). They do not offer support for H3b: the interaction between knowledge and health orientation is negative, opposite of our predictions. Hughner et al. (2007) note that most research identifies health as the main reason for buying organic food, but other authors, such as Williams (2002), find no conclusive evidence of an effect of organic food on people's health, compared with conventional food, which might explain this opposite result. That is, increasing knowledge does not reduce the incongruence between attitude and purchase behavior for health-oriented consumers. Instead, more knowledge of organic products strengthens the relationship only for environmental orientation in terms of the congruity in their attitudes and purchase behaviors. In brief, only consumers with environmental orientation will express more congruent attitudes and purchase behaviors when they know more about organic food, but consumers with a health orientation or hedonic orientation do not.

To check the heteroscedasticity of all estimated models, we ran a Breusch-Pagan test, one of the most common tests of heteroscedasticity. This test rejects homoscedasticity in the two models we used to test our hypothesis. Thus, we used robust estimations for the regression models appearing in the **Table 3**.

### DISCUSSION

The empirical results confirm both that knowledge and environmental orientation of consumers influence the congruity between their attitudes and purchase behavior when it comes to organic food. We also find significant interaction effects across these factors, such that more knowledgeable people with environmental orientation have more congruent responses, and their attitudes and purchase behaviors toward organic products in turn are more similar.

#### Theoretical Implications

In line with previous literature that suggests incongruities between attitudes and purchase behaviors toward environmental or green products (Kollmuss and Agyeman, 2002; Akehurst et al., 2012; Moraes et al., 2012), we find that consumers have very positive attitudes toward organic products, with a mean of 4.9 on a 0–6 scale, whereas their purchase behaviors are incongruent, with a mean of 1.3 on the same scale (see **Table 1**). This incongruity likely reflects the predictions of cognitive dissonance theory. This theory predicts that people can experience dissonance over a wide range of dimensions (e.g., behaviors, attitudes, beliefs), such that the set of possible mechanisms to mitigate the dissonance also is broad. Our results suggest that consumer orientations reflecting the environmental benefits of organic products, and consumers' knowledge about organic products both are mechanisms that can reduce incongruities between attitude and purchase behaviors in the organic market.

Consistent with Akehurst et al. (2012), who find a gap between purchase intentions and purchase behaviors for green products, which diminishes among people with greater ecological consciousness, we find a difference between attitudes and purchase behaviors. The experienced dissonance in turn can be explained by consumer orientations related to environment. Consumers with an environmental-protection orientation exhibit more similar responses, such that the gap between their attitudes and their purchase behaviors is smaller.

Knowledge also helps explain the dissonance resulting from this incongruity. This type of response is not only a mediator for the relationship between attitude, as a measure of overall belief about the product, and organic purchase as Pagiaslis and Krystallis-Krontalis (2014) propose for their relationship beliefsbehavioral intention. It is also a moderator. Knowledge helps transmit attitudes to purchase behaviors, overcoming several barriers to organic consumption, such as a lack of consumer confidence and high prices. Knowledge also moderates the effect of environmental orientation on the congruence between attitudes and purchases of organic food. For consumers with more information about the environmental benefits of organic products, the relationship between that orientation and the congruence between attitudes and purchase behaviors is stronger.

#### Limitations and Further Research

Some limitations of this study suggest directions for further research. We use only two items per factor about orientations. The results related to low Cronbach's Alpha of orientations could get better, for example, adding more items to measure these complex factors. On the other hand, we focus on factors that facilitate consumption and inhibit incongruity; we ignore factors that might inhibit consumption and facilitate incongruity, such as consumers' willingness to pay. Furthermore, we use consumers' assessments of their own purchases, rather than actually observed organic purchase data. This measure could lead to an underestimate of the actual level of incongruity between their attitudes and purchase behaviors. Other variables, such as social influence or social pressure, also could affect consumers' purchases of organic foods and their attitudes, such that they might increase or decrease the gap. Social influence also might be direct, or it could moderate the

### REFERENCES


effect of the consumer orientations on congruity between attitudes and purchase behaviors. Thus, additional research thus should consider factors that inhibit purchases, along with other variables that were not included herein, such as the effect of social pressure on organic purchase behaviors, to analyze the incongruity between attitudes and purchase behaviors in this market.

### AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication.

### FUNDING

This research was supported by Ministerio de Educación y Ciencia, Grant ECO2014-53060-R (Spain).


food and their implications for advertising strategists. J. Advert. 43, 18–32. doi: 10.1080/00913367.2013.799450


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Hidalgo-Baz, Martos-Partal and González-Benito. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Key External Influences Affecting Consumers' Decisions Regarding Food

#### María Pilar Martínez-Ruiz\* and Carmen María Gómez-Cantó\*

Facultad de Ciencias Económicas y Empresariales, Plaza de la Universidad, Albacete, Spain

Among the numerous internal and external forces that compete for consumers' attention in the context in which they buy their food, this paper will seek to provide a review of the most important external influences, such as the variables related to food itself. To this end, in addition to the food attributes traditionally identified in fields such as consumer behavior, it will give special consideration to the classification of food values. Although the influence of these variables on consumer decisions depends on the individual, analyzing them will undoubtedly increase understanding of consumers' decisions. Additionally, identifying and describing these variables will enable subsequent research on how they influence both consumer behavior and other key outcomes for producers, manufacturers, and retailers in the food industry, such as satisfaction, trust, and loyalty.

#### Edited by:

Alicia Izquierdo-Yusta, University of Burgos, Spain

#### Reviewed by:

Inés González, Escuela de Negocios de Navarra, Spain Blanca Garcia Henche, Universidad de Alcala, Spain

#### \*Correspondence:

María Pilar Martínez-Ruiz mariapilar.martinez@uclm.es Carmen María Gómez-Cantó carmenmaria.gomez1@alu.uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 24 September 2016 Accepted: 03 October 2016 Published: 18 October 2016

#### Citation:

Martínez-Ruiz MP and Gómez-Cantó CM (2016) Key External Influences Affecting Consumers' Decisions Regarding Food. Front. Psychol. 7:1618. doi: 10.3389/fpsyg.2016.01618 Keywords: food products, food values, consumer research, decision making

### INTRODUCTION

Recent decades have witnessed a number of changes in the buying habits and behaviors consumers have traditionally shown when purchasing food products (Pieniak et al., 2010; De Moura et al., 2012; Deloitte, 2015). To understand these changes, it is necessary to take into account the large number of forces, both internal (e.g., prior experience) and external (e.g., characteristics of the food products themselves), that compete for consumers' attention in the context in which they make their decisions (c.f. Garber et al., 2003; Mowen and Minor, 2003; Logue, 2015). Although the extent to which these influences ultimately affect consumers' buying behavior will depend on the individual, analyzing them will undoubtedly increase understanding of consumers' purchase decision processes with regard to food and, thus, facilitate proper planning for producers, manufacturers, and retailers in the food industry.

Consumer research is particularly difficult for food, among other things, because of the especially subtle and complex nature of food products as stimuli at points of purchase and during consumption (Garber et al., 2003). Consequently, among other external influences to receive attention, the relevant literature has focused in particular on variables related to food products themselves (Garber et al., 2003; Lusk and Briggeman, 2009; Lusk, 2011). Unsurprisingly, there is thus increasing interest in the field in trying to identify which food-related variables exert the strongest influence on consumer behavior (Deloitte, 2015; Logue, 2015).

Drawing on these ideas, this paper offers a synthetic review of those food variables that the relevant literature has identified as key external influences, including the latest developments. Specifically, it will address the research on food attributes and the food values identified by Lusk and Briggeman (2009), a significant development in the research in this field. The identification and

description of these variables is of great interest to enable subsequent work in this area, since: (i) it sheds light on which food variables have generally been considered key and should be taken into consideration in future research; and (ii) it facilitates the subsequent analysis of potential influences on and interrelations with various stages of the food-purchase decision process, as well as on other key results for retailers in the consumer goods industry, both financial (sales and profitability) and non-financial (satisfaction, trust, and loyalty).

### FOOD VARIABLES AS EXTERNAL INFLUENCES IN THE CONSUMER DECISION PROCESS: A LITERATURE REVIEW

In recent decades, numerous studies have sought to measure consumer preferences for certain food attributes over others (Lusk and Briggeman, 2009). However, some more recent work, such as that by Lusk and Briggeman (2009), has moved beyond the food attributes traditionally considered in the literature to propose a classification of food values, that is, a stable set of beliefs regarding the relative importance of the meta-attributes, consequences, and desired end states associated with purchasing and consuming food.

In light of the importance of both food attributes and food values as external influences affecting consumers' purchase decision processes with regard to food, the following sections will first review the food attributes generally considered in the relevant literature, in order to then use that discussion as a starting point to describe food values.

#### Product Attributes

Product attributes provide a basis both for marketers to differentiate and position existing products apart from those of their competitors and for the development of new products. This may be done based on a specific attribute or range of attributes or on several attributes at once (Belch and Belch, 1995; González-Benito et al., 2010).

A product's attributes influence consumers' product choices and are able to play a variety of roles (informational, communicative, symbolic, etc.). Industry operators must know what value consumers attach to those attributes, as well as how they factor into the purchase decision process. Moreover, companies must endow their products with the right level of attributes to meet consumers' expectations, without neglecting related managerial decisions, usually involving resource-allocation, cost, and price-setting considerations. Also, although decades ago the earliest work in the field tended to take into account only quantifiable product attributes that were objectively measurable, such as price, more recently, researchers have begun to include more subjective attributes in their work, such as quality (e.g., Kotler and Keller, 2012).

This broader and relatively more recent view of product attributes including attributes that are not only of an objective and measurable nature is clearly on display in the literature on food products. Indeed, until fairly recently, when choosing a given food product, consumers barely considered other types of issues, such as those related to good farming practices, food safety during the production process, nutritional quality, or the convenience or ease with which the product could be prepared and consumed (Berné and Martínez, 2007). In contrast, today's consumers have more and more information on these aspects and are thus more demanding when choosing the food they want to purchase. For instance, Robinson (2002) found that consumers supported sustainably produced food, although, paradoxically, they were not particularly likely to purchase it.

Nutritional aspects also generate considerable interest among end consumers, influencing their food choices. Consumers use this information to determine what nutrients they ingest, which largely affects their health (Kissileff and Van Itallie, 1982). In this regard, in their analysis of consumer orientations toward the health and hedonic characteristics of food products, Roininen et al. (1999) identified three health-related factors general health interest, light product interest, and natural product interest—and three taste-related factors—craving for sweet foods, using food as a reward, and pleasure. They also found that women were more interested in health- and tasterelated aspects than men, and that young people were less concerned with health and more interested in taste than older consumers.

Recent research by Deloitte (2015) confirmed the growing relevance of health-related attributes to consumers' food purchase decisions, noting that taste, price, and convenience are no longer the sole drivers of consumers' food and beverage purchases. The study further found that, in addition to these traditional drivers, more than half of American consumers now weigh the following drivers in their purchase decisions too: health and wellness, safety, social impact, experience, and transparency.

Food buying and consumption behavior has been widely studied from a psychological perspective, making it possible to focus on certain attributes over others (Logue, 2015). For instance, focusing on food reward, Berridge (1996) found that food consumption may be influenced by, among other things, certain taste-related psychological aspects, derived from the pleasure of the act of eating and the pleasantness of a food's taste.

Other lines of research in this area conducted from a psychological perspective include (Logue, 2015): (i) how people detect flavors; (ii) why people prefer some food to others; (iii) how people end up choosing certain foods over others; and, more specifically, (iv) how and why certain foods are able to influence consumers' choice behavior. All have considered numerous and diverse food attributes, to which they assign varying degrees of importance. Finally, from a perspective that transcends psychology, food provides people with opportunities to communicate and engage in a variety of socialization processes, allowing them to express and maintain their lifestyles, which are often linked to their individual cultures (Atkins and Bowler, 2001; Logue, 2015).

### Food Values

To understand how consumers evaluate food attributes and how they impact in the purchase decision process, it is particularly relevant to consider the article of Lusk and Briggeman (2009), who, in a key contribution, moved beyond the simple consideration of traditional food attributes. Although this work was published in the area of agricultural economics, the proposed values were developed based on a profound literature review on food preferences and human values, which enables to recognize these values as a cornerstone contribution in fields such as marketing and consumer behavior. Without doubt these values are closely related to the advent of the values-driven era (Marketing 3.0), that highlights the need to take care of customers not as mere consumers but as complex and multi-dimensional human beings with minds, hearts, and spirits. Because under this philosophy, customers choose those companies and products that satisfy their deepest needs for economic, social, and environmental justice (Kotler et al., 2010).

Rather than estimating consumers' preferences for certain specific food products and attributes, which consumers might have little knowledge of and/or experience with, Lusk and Briggeman (2009) identified a set of food values or metaattributes for which people might have better-defined preferences in order to gain greater insight into why consumers choose certain food products or attributes over others. To this end, these authors conducted a thorough review of the relevant literature on consumers' willingness to pay for food products and human values, which allowed them to apply the concept of overall life value, previously defined in pioneering work such as Rokeach (1973) and Schwartz (1992), to food. They were thus able to identify a set of food values likely to remain relatively stable over time. The aim was not merely to identify food attributes per se, but rather to identify more abstract attributes, consequences and end states of food consumption that might be applied to explain consumers' choices between a wide range of foods.

Specifically, they identified the following values (Lusk and Briggeman, 2009):


As Lusk and Briggeman (2009) explain, although some of these values may initially seem very similar to some of the classically considered food product attributes, they represent more abstract concepts, often encompassing numerous physical attributes at once. For instance, the value of nutrition can be considered more stable than a consumer's relative preferences for specific vitamin or fat contents. Likewise, while some of the proposed values, such as price, can be classified as personal (i.e., self-centered), others, such as tradition, origin, fairness, and environmental impact, can be regarded as social (i.e., societycentered).

To determine the relative importance consumers give to these attributes, Lusk and Briggeman (2009) conducted a survey of consumers in the US. They found that: (i) in general, there was significant heterogeneity across consumers in terms of the relative importance they assigned to food values; and (ii) on average, safety, nutrition, taste, and price were among the values consumers considered most important.

In a subsequent study, Lusk (2011) found that food values are significantly related to actual grocery store purchases, suggesting that the food values scale could potentially be used, among other things, to explain consumer choice and guide new product development and marketing decisions.

### DISCUSSION

This paper has offered a review of those food variables that, acting as external influences, impact consumers' foodpurchase decision processes. To this end, it first addressed the food attributes traditionally considered by the relevant literature, showing that, although early research tended to focus especially on objectively measurable attributes, more recently there has been a growing trend toward including more subjective ones. Indeed, recent data have confirmed the need to include these latter types of attributes, as, in addition to assessing easily quantifiable and objective attributes such as price, consumers increasingly also weigh other attributes in their decisions, such as those related to health or wellness.

This paper also considered the eleven food values identified by Lusk and Briggeman (2009), a significant advance in the research in this field. These authors suggest that consumers have intermediary values, consisting of a stable set of beliefs regarding the relative importance of the meta-attributes, consequences, and desired end states associated with food purchases. These values are intended to represent the main values influencing food choice and be comprehensive enough to cover the full breadth of issues that tend to drive consumers' decisions with regard to food.

The synthesis provided in this paper helps pave the way for future research in the field by identifying the food variables that are generally considered to be important and should thus be taken into account. This will facilitate their inclusion in subsequent studies, for instance, on potential influences and interrelationships in different stages of the food-purchase decision process or with regard to other key outcome variables.

Finally, one limitation of this article lies in the small number of external influences considered. Future research should thus expand the review to include other external factors, as well as address internal influences.

## AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication.

### REFERENCES


### ACKNOWLEDGMENTS

This paper was funded under Research Project No. ECO2014-59688-R, Planificación e Implementación de Estrategias de Gestión Óptimas del PDV Físico, Online y Móvil a Partir de las TIC y la Innovación (Planning and Implementation of Optimal ICT- and Innovation-Based Management Strategies for Physical, Online, and Mobile Retail Outlets), by the Spanish Ministry of Economy and Competitiveness.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Martínez-Ruiz and Gómez-Cantó. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Emotional Effects on University Choice Behavior: The Influence of Experienced Narrators and Their Characteristics

#### Ana I. Callejas-Albiñana<sup>1</sup> , Fernando E. Callejas-Albiñana<sup>2</sup> \* and Isabel Martínez-Rodríguez <sup>2</sup>

<sup>1</sup> Department of Psychology, University of Castilla-La Mancha, Ciudad Real, Spain, <sup>2</sup> Department of Spanish and International Economic, Econometrics and Economic History, University of Castilla-La Mancha, Ciudad Real, Spain

This study analyzes the influence that experienced users of university resources might have as narrative sources of information for other students in the process of choosing their schools. Informative videos about the benefits of studying at the university provide a reference model. In these videos, a group of young people present their views and explain their reasons for choosing the university in which they are pursuing their degrees; the various narrators detail all the resources available. This study investigates whether the individual identifiers of these narrators (e.g., gender, age, physical appearance, nonverbal gestures such as smiling, posture) influence perceptions of the credibility of the information they provide. Among a sample of 150 students in their last year of pre-university training, the results demonstrate that the students' ability to identify with the narrators provides information and arouses emotions that inform their perceptions of reliability and therefore their consumption choices. None of these predictors appear to serve as determinants that can be generalized, but if emotional attitudes in response to narratives about the topic (i.e., the university) are positive, then they prompt a change in attitude toward that reference topic too.

#### Edited by:

Ana I. Jiménez-Zarco, Open University of Catalonia, Spain

#### Reviewed by:

Victoria Labajo, Comillas Pontifical University, Spain David Cockayne, University of Huddersfield, UK

\*Correspondence:

Fernando E. Callejas-Albiñana Fernando.Callejas@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 08 March 2016 Accepted: 25 April 2016 Published: 24 May 2016

#### Citation:

Callejas-Albiñana AI, Callejas-Albiñana FE and Martínez-Rodríguez I (2016) Emotional Effects on University Choice Behavior: The Influence of Experienced Narrators and Their Characteristics. Front. Psychol. 7:689. doi: 10.3389/fpsyg.2016.00689 Keywords: emotions, informative video, consumption, university, narrative

## INTRODUCTION

The proliferation of private Spanish universities in recent years (of the 32 currently in operation, 14 are new since 2001), adding to the 50 public universities, all of which offer a wide range of degree programs, makes it imperative for these institutions to find new ways to attract consumers of their products, by encouraging diverse students to prefer their offerings to those provided in other, nearby regions. For example, the University of Castilla-La Mancha (UCLM) is a public university, located in the Secondary School of Spain, which manages to attract just 40% of the students in its surrounding community who qualify (i.e., those who pass an entrance exam). Yet data from the Ministry of Education, Culture, and Sports also reveal that UCLM includes students from other, nearby regions, such that overall, its average enrollment is comparable to that of other regions with similar characteristics.

These data in turn prompt an inevitable question: How do potential students in a particular region decide to study, or not, at the university located in their autonomous community? In its efforts to answer this question, and accordingly encourage more students to enroll, UCLM has undertaken a continual quest for improvement, seeking to advertise and diffuse its educational offerings, especially through online marketing strategies. For example, in online promotional videos, real UCLM students, graduates, or those who dream of attending the school describe their experiences at the university and their related views of its offerings. These videos then are available to be shared freely with potential enrollees. To test the effectiveness of this recruitment and marketing strategy, the current study seeks to analyze whether and how the narrators of these videos, who share their experiences with the product and thus provide signals regarding the prototypical users (i.e., students) of the university's offerings, affect the decision of students considering whether to enroll. In particular we consider whether the experienced narrators represent potential predictors of those decisions. To begin, we consider the different factors involved in this type of advertising and the reactions they might generate.

### CONCEPTUAL FRAMEWORK

The central variables underlying our empirical study reflect insights gained from research into electronic word of mouth (WOM), consumer behavior, and the relevance of emotions in decision making, as we review in this section.

#### Electronic Word of Mouth

Advertising through various media formats represents a common strategy for online marketing. For example, informative videos on websites can facilitate users' access to pertinent knowledge, and they can reach these videos through their smartphones, computers, tablets, or other devices. Such videos also do not require information gatherers to read extended texts or rely on just still images.

These videos also can enable a strategy by which real people, who have personal experience with a particular product, appear in the videos to share their "word of mouth," in which case it constitutes electronic word of mouth. According to the Association for Research of Media (2015), nearly 80% of Internet surfers read others' opinions or comments, and more than half of them express great confidence in this source of information. In particular, word of mouth affects 20–50% of all purchasing decisions (Marketing Directo, 2013). The increasing influence of eWOM suggests that it offers strongly compelling information to most consumers (Kozinets et al., 2010; Zhu and Zhang, 2010; Lim and Chung, 2011); it has long been well-established that the recommendations and opinions disseminated through word of mouth have far greater influence than other forms of paid advertising.

#### Consumer Behavior

Performance, from a psychological point of view, refers to the set of actions and behaviors that a subject exercises in response to external and internal stimuli. In a consumption context, this definition implies that consumer behavior is the set of actions conducted during a decision-making process, which arises due to some lack or perceived need. According to Wilkie (1996), consumer behavior involves the physical, mental, and emotional activities that people adopt to access the products and services they believe will meet their needs and desires. Arellano (2002) states that consumer behavior is related to the internal or external activity of the individual or group of individuals aimed at meeting their needs by purchasing goods or services.

Marketing researchers should be aware that there are many external variables that influence buying behavior (Rivas and Grande, 2004). To unify the various proposals we have to focus on connotations related to the common elements that can be seen in all of them, such as:


However, many decisions rely on quick thinking or emotions, such that they are relatively unconscious. Decision makers do not always undergo a rational process to arrive at conscious decisions. In some cases, others with some level of expertise or experience can prompt certain consumer behaviors, especially if they highlight needs that become personal aspirations. This effect implies a process of motivation, which refers to "the intensity, direction, and persistence of the effort to achieve a goal" (Robbins, 2004, p. 156).

Furthermore, two main streams of prior research identify the key predictors of consumer behavior. On the one hand, some studies focus on how consumers receive information and process it for use, premised on the notion of a fundamentally rational consumer (Solomon, 1998), who focuses on the decision process and brand evaluation. This stream of research primarily uses cognitive psychology as a foundation. On the other hand, other research notes the emotional, social, and cultural aspects of consumer behavior and therefore centers mostly on investigations of feelings and emotions. These two streams actually are complementary, in that they cover the various reasons people engage in the act of buying, and they span multiple cause-and-effect rationales for human behavior. As Bagozzi et al. (1999) suggest, a necessary condition for an emotional response to a situation or event is that the person already has some personal interest in the situation and is willing and able to assess the situation to support or mitigate this sense of interest or concern.

Focusing on the decision process involved in choosing a University, different stages of a clearly complex decision were analyzed, for example Chapman (1986) establishes five stages in a model of the college selection process: 1- Pre-search behavior, 2- Search behavior, 3- Application decision, 4- Choice decision, 5- Matriculation decision.

Another analysis of the stages of University choice is that made by Kotler and Keller (2009) establishing 5 such stages: 1- Need arousal, 2- information search, 3- alternative evaluation, 4- decision, 5- decision implementation, 6- post- purchase evaluation. The first three involve a search for a system of prior assessment based on the motivation and characteristics of universities that have previously been selected by the subject, including those given by word of mouth, or other factors that may be attractive, all of them leading to the establishment of criteria for optimal evaluation in the final decision (stage five), although the process does not end with the deployment option but rather advances toward the final phase (6), which involves evaluation of the success of the choice and the degree of satisfaction.

Using this as a reference we can emphasize that universities should establish attractive offers and maintain established services to the standard expected by its users.

We now analyze models of University choice based on factors that promote the final decision. Chapman (1981) performs a study about relatively fixed university in attracting students. Chapman describes two types of factors related to the choice of college students, based on variables such as skills and prerequisites of students and the student's choice of college, highlighting three types of variables based on external influences, relevant people within the family environment, friends or teachers; University features (localization, co-variation of degrees), and college efforts to communicate with student (written information, campus visit, admission).

Raposo and Alves (2007) establish a classification of models based on factors that influence the choice process of a college or university:


Following the exploratory study by Raposo and Alves (2007) whose claim is based on establishing a model of University choice that brings together all criteria derived from the models listed above in four dimensions: institution overall reputation, educational offer, previous knowledge about the institution, individual factors, influence of others. The results show the difficulty in determining generalizable and stable criterion factors for all persons and areas of knowledge as there is a significant connection between all these factors, but it is not decisive, although highly influential variables stand out as markers of choice of University: proximity to home, costs, parents, and secondary school teacher recommendations, etc. It nevertheless provides interesting data regarding the effort universities must make to attract students because it shows the need to address all dimensions referenced to the same degree of importance, in order to ensure that the choices students make will be based on one of these factors.

In our study, one of the objectives is based on the dimension of prior knowledge and the relationship with individual and influence of other factors, and promotion of institutions as prerequisites to then move into one of the modes of uptake data: the promotional video.

### Emotions in Advertising

Advertising uses various creative ways to win over consumers. Emotional advertising is an effective form of communication that companies can use to achieve differentiation from competitors, because this format awakens diverse sensations among audiences. (López, 2007). For example, The ARS Group (2010) uses the pleasure–arousal–dominance (PAD) theory to explain the role of emotion in advertising and how it generates preferences, sales, and market share. The Group has developed a technique to measure emotion and nonverbal responses, using visual techniques. The approach has been used successfully in more than 600 proprietary studies and 35 academic studies around the world. It also aligns with theory, as proposed by Gardner (1983) and subsequent (Clark and Isen, 1982; Burke and Edell, 1989) studies, that suggests that the affective responses influence cognitive processes, including evaluation, memory, and judgment.

Because so many decisions are subconscious, marketing strategies generally aim to provoke positive expectations among consumers, such as by forging positive emotions that evoke pleasant memories and benefits related to the consumption object. Emotional marketing acknowledges that emotions can satisfy consumers while also creating expectations about their feelings and sensations. To ensure that emotional marketing leads to consumption behaviors, marketers must establish a series of steps, such as identifying the wants and needs of consumers, establishing a link between their interests and intangible properties of the product, and creating a communication strategy that expresses the identified emotional perceptions, such that that both motivation and emotional processes influence consumers' psychic activity (Fernández-Abascal, 1997).

The basic process of communication theory emphasizes the importance of the source (sender of message) as persuasive element and the characteristics to be taken into account such as credibility, attractiveness, similarity, familiarity, and reputation. The most important is credibility, which depends on two fundamental factors: competence, understood as experience or reliable information which informs the skill in handling the transmitted message fluently (Briñol et al., 2001), and the sincerity that should allow the message to be sent with no profit motive and a lack of persuasive intent (Morales et al., 1997).

The attractiveness of the sender is one of the most persuasive elements in purchaser choice as it causes more attention to the message given, it influences the identification process because they may want to act the same way, and it can increase credibility by providing greater communication skills (Chaiken and Eagly, 1983). Furthermore, this attractiveness can be a critical component in making the message more effective (Schlecht, 2003).

However, the effects of source attractiveness can be overcome by the effects of credibility so that a highly credible source emitter with low appeal can be more effective than another with high attractiveness and low credibility (Wachtler and Counselman, 1981).

The credibility model of Ohanian (1990) relates these elements, indicating that purchase intention may be affected by the perceived attractiveness, reliability and level of experience of the source, draw attention to the positive characteristics of the communicator, affecting the acceptance of the message by the receiver.

The perceived similarity of the source by the receiver is also important especially if it refers to the same social group and provides attitudinal similarity (Briñol et al., 2001).

Both familiarity and reputation can be persuasive factors provided they do not become habitual, i.e., they change both the characters and the message (Briñol et al., 2004).

The choice of message transmission is one of the most interesting points in the process of persuading someone to purchase a product. The choices are to show directly, or to use a spokesperson or celebrity; however, in both cases the above factors should be taken into account. The study by Sertoglu et al. (2014) measures the differences in credibility perceived in Turkish society regarding the use of spokesman and celebrities and their impact on the purchase decision, obtaining that spokespersons are perceived as a more reliable and positive influence on purchase intent that celebrities, whereas celebrities are perceived as more attractive. This is in agreement with the studies of Van der Waldt et al. (2009). However, the society being studied should be considered when generalizing the results.

Long studies on the involvement of emotions in the process of purchasing decision have great relevance and Zajonc and Markus (1982) point out that there is a significant relationship between affective and cognitive reactions that affect the formation of the attitudes.

The theory of cognitive assessment studies the impact of emotions on behavior based on the cognitive assessment of the subject toward a certain behavior. According to Arnold (1970) initial evaluations begin the emotional sequence and arouse both appropriate actions and emotional experience, so that physiological changes accompany actions and experiences but do not start them. Lazarus (Lazarus, 1991) consider that the nature of the assessment of cognitions underlying emotional reactions is related, firstly, with a primary assessment linked to the importance of the event at the organic level, and then the organism's ability to face the consequences of the action should be assessed. However, Taylor (2009), believes that emotions can exist throughout the decision process with cognition.

The regulatory role of conduct in relation to behavior is also determined by the prospect of the cognitive assessment process which leads to emotional experience as the level of arousal. Di Muro and Murray (2012) examine how the interaction between the level of excitement and the valence of the emotional state affects consumer preferences, so that a positive mood tends to the choice of products according to their level of arousal, while a negative mood will choose products unsuited to the current level of excitement. Research Carrera and Oceja (2007) indicate that level of emotional intelligence is significant for handling negative emotions in the selection process as it can achieve positive results.

The impact of emotions on decisions related to consumption has acquired great significance today (Kotler et al., 2010; Peter and Olson, 2010), which leads to the need to develop marketing strategies based on the consumer rather than holistic strategies.

Finally, the new discipline of neuromarketing is based on the study of the brain in an effort to understand unconscious patterns that govern the buying process. Researchers using these techniques argue that consumers' attention can be captured through by images that excite them, not by rational argument. The more intensely deeper emotions are generated, the better the neurological brain connection encouraged by consumer advertising can reinforce these neural networks. We can point to many examples of the application of this notion, such as olfactory perceptions prompted by the smell of perfumes or cakes in bakeries, auditory perceptions sparked by contemporary music played in stores that target young consumers, or visual perceptions resulting from lighting that emphasizes a particular product. This study seeks to identify relevant variables in promotional videos that evoke emotional connotations that also can predict the consumption decisions among their audiences.

### RESEARCH

Concerning the information about the ethics approval, we would like to highlight that: We always give participants detailed and written information about the study and the procedure. We do not collect data related (direct or indirectly) to the subject health at all and usually do not refer to the declaration of Helsinki when informing the subjects. We always guarantee the anonymity of the data collected. Consent was implied by voluntary completion of the questionnaire.

Several research goals drive this work. In particular, we seek to:


To achieve these research goals, this study uses six independent variables related to the study participants (i.e., potential students at UCLM) age (16–18 years), gender, high school course (controlled for, second baccalaureate<sup>1</sup> ), residence (i.e., city

<sup>1</sup>The Spanish education system includes two high school courses, defined by they specialty, for students who plan to attend college. These studies are not binding as basic education, but they are required for university entrance.

#### TABLE 1 | Narrator traits.


A&H, Arts and Humanities; Sc, Sciences; H, Health Sciences; SS&L, Social and Legal Sciences; E&A, Engineering and Architecture.

#### SAMPLE


with <10,000 inhabitants, city with more than provincial capital), course of prior studies.

High school specialty (science, humanities/social sciences, or arts), areas of University knowledge, arts and humanities, social sciences, health sciences, social/legal sciences, or engineering and architecture). The dependent variables describe the character of the narrator in the video. They include the narrator's age (15–30 years), gender, and degree.

Sample of secondary schools in Ciudad Real province, we used stratified random sampling, based on the number of inhabitants. Accordingly, we defined three groupings based on residence, reflecting smaller cities, larger cities, and the capitol. We randomly sampled people who lived in all three locations. Localities included secondary schools: 16 in smaller cities (<10,000 inhabitants), 12 in larger cities (>10,000 inhabitants), and 1 in the capitol. The random draw then led us to select populations from Argamasilla de Alba (7201 inhabitants), Puertollano (50,608 inhabitants, seven centers of which randomly chooses one of them), and Ciudad Real (74,960 inhabitants). This distribution produced the sample detailed in **Table 1**. However, the information in **Table 1** does not include the type of bachelor's degree, because this option is not available to customers in any of the selected Secondary Schools, and one of the preconditions for sample participants was that they were attending their last year of their baccalaureate degree.

Complete a questionnaire to provide the data for our analysis. We solicited cooperation from the teachers, who agreed to devote some class time to allowing the students to complete the questionnaires, and the schools which made computers available for the students to complete the surveys. All schools had dedicated computer rooms or spaces, so this requirement was not a limitation.

The questionnaire (see the Appendix) consisted of two parts: a first part to collect demographic data about the participants and a second part that required participants to visit the website of the University of Castilla-La Mancha and locate a banner entitled "ASI all started," which linked to the informative video. The video featured 12 narrators, describing their first-hand relationship with UCLM from different perspectives, representing different campuses, degrees, and levels (undergraduate, graduate), as well as some people who simply expressed their interest in studying at the university. **Table 1** summarizes their characteristic features. On the website, participants also could access more information about each narrator by clicking on a related title, which revealed that person's history. Each participant could choose to read or listen to the personal information. After receiving these details, the participants completed several items related to their resulting impressions. Requested issues are: (1) which character you've chosen? (2) Why did you choose? (3) How old you are or think has the character you've chosen? (4) What do you like best of what it has been told? Why? (5) What it is the least liked you? Why? (6) It has been clearly expressed? Highlights some phrases or comments that have caught your attention; (7) Indicates aspects of his gestures, look, smile, posture, physical appearance; (8) It has shed new light on the University of Castilla La Mancha; (9) has your opinion changed about making your studies at this University?

Ultimately the idea is to confirm whether the information received from agents that represent similar characteristics or an expectation can be highly efficient for the choice of university studies. The work of Hung and Mukhopadhyay (2012) examines how visual perspectives taken to assess a situation are related to seeing themselves as actors in that situation influencing the intensity of the experience (related to the characters in the video).

The focus is on identifying the factors that are decisive in the choice of University and the motivation behind them, for example Keshishian and McGarr (2012) note that, for New York students, intrinsic motivation, social relationship, and the possibility of achieving the goal set by the student, act as important predictors and may be helpful when preparing recruitment strategies. However Mohamad et al. (2012) found that for Malaysian students, the fees (academic cost) were the most representative factor when choosing university.

Another notable factor is that the protagonists are natural storytellers, not celebrities. The fact that it is the image and the slogan which decide the choice or otherwise of the character is due to the claim that the greater personal involvement, the greater the tendency to pay attention to the message, to the arguments and not to peripheral signals such as the attractiveness of the source (Morales et al., 1997).

Gender differences regarding the choice of the character are also a differentiating factor when choosing and defining information about search behavior as a criterion for guidance through the media (Barber et al., 2009; Meyers-Levy and Sternthal, 1991).

In these terms, we want to know what the relevant factors for students in Castilla La Mancha (Spain) would be, and whether the marketing strategy is suitable or new methods may be found.

#### RESULTS

The data analysis relied on the software package SPSS, version 22. Our analysis revealed insights that reflected our research objectives. First, the participants clearly preferred eWOM, such that all of them (n = 150, 100%) chose to watch the video of the narrator versus reading their history.

These results are consistent with related studies by different authors (Chaiken and Eagly, 1983; Meyrowitz, 1985) on preference in video or audio media over printed sources.

Second, to investigate the links among the study variables, we performed a preliminary analysis of the frequencies of the characteristics of the narrators, to test its effect on the participants' choices. With the significance of the variables and their χ 2 -values, we built a cross-contingency table related to the features of the narrators and the personal characteristics of the participants. Then, we used a logistic regression to determine the predictive ability of these variables for participants' choice to study at UCLM.

Third, we consider the frequencies of the variables related to personal characteristics: age, gender, course of studies, situation, and narrator chosen. These descriptive statistics, with the exception of the age variable, are categorical variables. Therefore, it is important to acknowledge the descriptive statistics (Mo), or the value with the greatest frequency distribution data, without ignoring that categorical variables are inherently encoded. We find that the age of the narrator that earns the highest percentage rating is 30 years (29%); ages older than 25 years account for 80% of the total (**Table 2**).

However, the character most frequently chosen has the highest age (30 years), followed by 28, 26, and 23 years, so it is understood that the preference is in the expectation of achievement, and that in the attractiveness of the slogan in these cases the coincidence is in the mix of interests, not only academic but also personal (see "character traits").

In addition, as we show in **Table 3**, male narrators are more likely to be selected than female narrators. The number of male narrators in the video was greater, by more than 50%, though this statement does not need to be relevant. However, we also uncover an apparently inverse relationship with regard to gender: Greater percentages of female (n = 105) and male (n = 49) participants in the sample did not choose a narrator of the same sex. With regard to the course of study, the most mentioned focus of the narrators was social and legal sciences (37%, **Table 4**). Regarding the focus of the participants in their baccalaureate studies (science or humanities), somewhat surprisingly, more students had chosen to study science (n = 108). Even if we include arts and humanities, the percentage value for science

#### TABLE 2 | Frequency of the age of the narrator.


TABLE 3 | Frequency of the gender of the narrator.


#### TABLE 4 | Frequency of the knowledge of the hero Rama.


TABLE 5 | Frequency of situation of the protagonist.


remains greater. This finding makes more sense in relation to the findings regarding their planned university studies, as we discuss subsequently. Furthermore, as **Table 5** shows, the participants preferred narratives by people who currently are students at the university (80.9%), followed in much smaller percentages by students who had completed their undergraduate degrees and were continuing on to postgraduate studies (17.9%), and then by other prospective students (0.7%).

**Table 6** contains the frequency of selections of narrators representing different fields of study. 30-year-old engineer is engaged in his postgraduate studies. Thus, he effectively represents our findings regarding the most preferred gender, age, and course of studies. However, his degree, engineering, represented only the penultimate position in terms of choices. The next most frequently chosen narrator was 27-year-old Rocio, a student of environmental science, who also represents the most preferred features in terms of age. In contrast, two narrators did not prompt any significant selections. Both of them were younger than 25 years of age and did not currently attend UCLM but instead described their desires to go there (i.e., Juan Miguel and Lucia).

#### Cross-Table Analysis

Fourth, the cross-tabular analysis details the relationships across variables. A relationship between two categorical variables cannot be established simply by observing the percentage frequencies; rather, we required some measure of the association and its corresponding significance. Therefore, working with statistical non-parametric tests, we relied on Pearson chi-square (χ 2 ) values to determine whether a relationship exists between two categorical variables, reflecting the difference between the expected and observed frequencies. For the hypotheses tests, we note the degree of significance, which indicates the certainty associated with the relationship, though not of its strength. For that measure, we turn to the V Cramer and Pearson Phi, which indicate relationship intensity. The contingency tables accordingly present the independent categorical variables in rows

TABLE 6 | Frequency of choice of the protagonists.


and the categorical dependent variables in the columns. Using the statistical χ <sup>2</sup> measure also requires that the expected frequencies are not too small. That is, if frequencies of <5 are expected, they must not exceed 20% of all expected frequencies. If the percentage exceeds 20%, then we must interpret the Pearson statistic with particular care. In **Table 7**, we denote these assumptions with an asterisk. Finally, the university course chosen by the participants was not estimated, because it remains a constant factor for these baccalaureate students.

To interpret the data in **Table 6**, we use reference variables, associated with each narrator feature and its association with the participants' characteristics. For example, the gender of the narrator had no association with age. Nor did we find any effect of the narrator's gender on the choice to study at UCLM. The correlations of the genders of the narrators and the participants also revealed no evident relationship [χ <sup>2</sup> = 7360 (1, N = 150) = 0.007, p > 0.05]. Of the 49 male respondents, 42 chose a narrator of the same sex, and 7 chose a female narrator. However of the 107 women surveyed, 65 chose men and 36 women, such that more than 60% of these participants opted for narrators of the opposite sex. In addition, the type of facility is significant [χ <sup>2</sup> = 13,629 (2, N = 150) = 0.00, p < 0.05], and the degree of intensity of.301 V Cramer indicates a moderate relationship. The type of baccalaureate [χ <sup>2</sup> = 35,912 (1, N = 150) = 0.00, p < 0.05] is statistically significant, but the intensity ratio (V. indicates a strong degree of moderate intensity). The relationship, according to the ratio phi = −0.489, also is reversed, such that the variables fluctuate and when one increases, the other decreases in value. The type of area of university knowledge future of the students surveyed, in relation to the gender of the narrators, indicates that χ <sup>2</sup> = 45,789 (6, N = 150, p < 0.05), but 21.4% count <5, so the significance of these variables is not sufficient.

The age of the narrator relates significantly to several other variables (high school specialty, secondary school, areas of university knowledge, study at UCLM) but cannot be regarded as a reliable predictor, because the frequencies of <5 exceed 20%. The age and gender of the participants offer no significant results.

#### TABLE 7 | Contingencies.


\*Expected frequencies under 5 to over 20%.

The field of study that the narrator is pursuing correlates positively with the gender of the participants [χ <sup>2</sup> = 9767 (4, N = 150) = p < 0.05], but the intensity Cramer V = 0.045 indicates a discreet relationship. The type of baccalaureate degree and the narrator's field of study showed a positive [χ <sup>2</sup> = 38.879 (4, N = 150) = p < 0.05] and intense (Cramer V = 0.516) correlation. The degree also relates related to the study at UCLM [χ <sup>2</sup> = 16,077 (4, N = 150) = p < 0.05], with a relatively high intensity. The remaining variables have no influence. Regarding the situation of the character, not influence results are statistically significant in the variables analyzed regarding subjects surveyed.

Although we have presented the cross-table results for the narrators, it was expected that the results were not reliable regarding expected frequencies as they have been fractionating in features that they can be representative (i.e., age, sex, type of area of knowledge of university qualifications, etc.).

Study pertains to the relationship between the variables reflecting the intended choice to study at UCLM or not. The contingency tables between choosing UCLM or not and other student characteristics (age, gender, high school specialty, secondary school, area of university knowledge future) indicate significant relationships only between the choice of UCLM and the type of school (**Table 8**), which in turn reflects the number of inhabitants in the region. Starting with the frequency determination, we find that 55.3% of participants indicated that they would not chose to study at the UCLM, whereas 44.7% prefer to study there, though men more frequently indicated that they would not choose to study at UCLM (61.2%) than did women (52.5%). The statistical results also indicate a χ 2 value of 10,848 [(2, N = 150) = p = 0.004]; moderate relationship between the two variables. Ultimately these data appear to exert significant influences on Secondary School in the choice variable UCLM. This relationship is reflected in the

#### TABLE 8 | Cross-Tabulation UCLM \* Secondary School.


percentages of frequency expressed: students from towns with fewer than 10,000 inhabitants are more likely not to choose UCLM as the location for their higher education studies.

## Logistic Regression Analysis

The logistic regression model seeks to predict the estimated probability of a dependent variable and present some possible values (1 = yes, 0 = no), depending on the different values for the independent variables. That is, it reveals how dependent variable relates to one or more independent variables multinomial logistic regression and categorical variables that are not dichotomous, so that we can classify the participants according to the values of a set of predictors. Because our research goal is to analyze the effectiveness of a promotional device, according to factors involved in it, we define the variable Y as interest in studying at UCLM or not and variables X as the gender of the narrator (GENDERNA), the narrator's areas of university knowledge (AREANA), and the environment of secondary school (SECONDSCHO). These X variables thus reflect the results of the observed factorial relationship variables. The test of the predictive nature of this model depends on the extent to which the logarithm of the odds ratio increases when we observe the presence of these variables.

To test this relationship between predictor variables has made an analysis of the Hosmer-Lemeshow (B) to evaluate the goodness of fit and check for predictive relationship between selected as independent and dependent variables, and the Wald test (W) to check the plausibility of the model.

**Table 9** contains the estimates of the parameters for the final model of METHOD = ENTER. SEXOPE. RAMAPE. CENTER. The results reveal that χ <sup>2</sup> = 26,356 [(7, N = 150) =, p < 0.00]. The lack of consistency of the predictive model means that we cannot establish its significance, despite some suggestion that the key character is a woman with some degree of knowledge about science.

### Analysis of Open Questions

By studying participants' responses to the open questions in the questionnaire, we can address some alternative explanations and test some related concepts. In particular, we gauge the participants' first impressions of the narrators, their perceptions of the narrators' communicative ability, and their physical traits. Then we investigate if any of these perception change their view of the information received about UCLM idea and their choice of whether to study at UCLM or not. In this section, we present the most relevant findings, in the order in which they appeared in the questionnaire (see the Appendix).


Despite acknowledging that they received broad, attractive information about the UCLM, 68.2% of the participants noted that they did not change their initial intentions with regard to choosing the UCLM for their higher education; others often expressed basic doubt in categorical responses, such as "I don't know. Maybe. I'll think about it."

## DISCUSSION AND CONCLUSIONS

The study data reveal that the promotional video for UCLM meets the standards for a marketing strategy that leverages eWOM, in both format and the argument set out, and it helped increase positive attitudes among the audience who read or heard the stories of the different narrators. These results emphasize the influence of the communication channel in decisions and persuasive capacity. Through the video as a resource, the message becomes visually important, is expressive and encouraging direct printing and images, facilitating perceptions of more intimate relationships with the characters (Meyrowitz, 1985). These reactions immediately follow the viewing of the information through the source, since the actual decision is often not present to act as an explicit signal for consumer choice (Monroe and Chapman, 1987).

In short, positive attitudes toward the decision to study at UCLM increased after viewing, although this does not ensure that the time that elapses until the actual decision cannot vary based on perceptual factors and the magnitude of emotional responses (Meyers-Levy and Sternthal, 1991).

Our findings also reveal some links across the variables studied. Considering the demographic data (age, gender, degree), it seems logical to predict that participants would prefer to receive eWOM from narrators whose characteristics are similar


#### TABLE 9 | Estimates of the model.

<sup>a</sup>The category of reference is: YES.

<sup>b</sup>Set to 0.

to their own. However, we determined that female participants chose more narrators of the opposite sex, whereas men chose same-gender narrators.

Although the literature based on the difference in information sources between men and women warns that the discrepancy is related to interests or knowledge (Holbrook, 1986; Fischer and Arnold, 1994), it is possible to explain it in terms of differences in information processing (Meyers-Levy, 1989). Women are more thorough when coding nonverbal messages (Everhart et al., 2001), they are also more affected by visual stimuli and are more emotional (Dittmar et al., 2004). Men are more analytical and therefore more objective (Kim et al., 2007).

We must not forget that in our study, the number of male narrators is greater than that of women as an influencing factor when choosing, as are characteristics related to the source, mentioned above, and the prior credibility of the character to the argument.

A similar result applies to the choice of degree, which revealed an effect not of the similarity of the degree or the focal field of study that the participants aspired to but rather the influence of the suggestive content of the slogan used to attract their attention to certain stories—namely, those involving personal achievement, especially in relation to art and engineering studies that might lead to the fulfillment of a dream. The participants also preferred stories told about the present or past, reflecting lived experiences or experiences in progress. In summary, the interest in the product (university) is most closely linked to experienced users, who continue to study, who are men, and whose narratives suggest a history of achievement, which might be related to their future aspirations as academics. In this point, gender differences can also be interpreted. Men are less likely to get involved in the stories, but women tend to enter into the story from personal experience (Kim et al., 2007).

However, we cannot establish any predictive traits of the narrators, such that we could anticipate which of them lead audiences to choose UCLM.

This study also helps identify some possible causes of "leakage" of students to other universities. The results indicate that the most significant cause is due to location factors, reflecting both the actual region and the size of their home towns. For example, people from smaller cities appear most interested in "flight" from their location region.

In short, the use of the promotional video to share information about UCLM is effective, if discreet, and provides a resource that can capture the attention and spark positive attitudes; in no case did it generate negative reactions (e.g., putting positive intentions to choose UCLM in doubt). To advance and continue this investigation, it would be interesting to track the students who participated in this study to specify their final decision. Because these students often are the consumers but not necessarily the buyers of the university product, it also would be interesting to determine if the preferences and choices of parents (buyers) match those of the students (consumers).

### AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct, and intellectual contribution to the work, and approved it for publication.

#### ACKNOWLEDGMENTS

This work was funded by the Ministry of Economy and Competitiveness (Spain), through Research project ECO2014- 59688-R, Programa Estatal de Investigación, Desarrollo e Innovación Orientada a los Retos de la Sociedad, Plan Estatal de Investigación Científica y Técnica y de Innovación 2013–2016.

### REFERENCES

Arellano, R. (2002). Consumer Behavior, in Focus Latin America. Mexico: McGraw Hill.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Callejas-Albiñana, Callejas-Albiñana and Martínez-Rodríguez. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

### APPENDIX


# Person-Organization Commitment: Bonds of Internal Consumer in the Context of Non-profit Organizations

Emma Juaneda-Ayensa<sup>1</sup> \*, Mónica Clavel San Emeterio<sup>1</sup> and Carlos González-Menorca<sup>2</sup>

<sup>1</sup> Departamento de Economía y Empresa, Universidad de La Rioja, Logroño, Spain, <sup>2</sup> Facultad de Ciencias Jurídicas, Sociales y Humanidades, Universidad Internacional de La Rioja (UNIR), Logroño, Spain

From an Organizational Behavior perspective, it is important to recognize the links generated between individuals and the organization that encourage a desire for permanence. After more than a half century of research, Organizational Commitment remains one of the open questions in the Psychology of Organizations. It is considered an essential factor for explaining individual behavior in the organization such as satisfaction, turnover intention, or loyalty. In this paper, we analyze different contributions regarding the nature of the bond between the individual and the organization. Taking into account the peculiarities of Non-profit Organizations, we present different interpretation for later validation, comparing results from the Confirmatory Factor Analysis of the four models obtained using exploratory factor analysis, both conducted on a sample of 235 members of Non-profit Organizations.

#### Edited by:

Maria Pilar Martinez-Ruiz, Universidad de Castilla La Mancha Albacete, Spain

#### Reviewed by:

Santiago Gutiérrez-Broncano, University of Castilla-La Mancha Toledo, Spain Jorge Linuesa-Langreo, Universidad de Castilla-La Mancha Cuenca, Spain

#### \*Correspondence:

Emma Juaneda-Ayensa emma.juaneda@unirioja.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 29 March 2017 Accepted: 04 July 2017 Published: 20 July 2017

#### Citation:

Juaneda-Ayensa E, Clavel San Emeterio M and González-Menorca C (2017) Person-Organization Commitment: Bonds of Internal Consumer in the Context of Non-profit Organizations. Front. Psychol. 8:1227. doi: 10.3389/fpsyg.2017.01227 Keywords: non-profit organizations, affective commitment, continuity commitment, normative commitment, confirmatory factor analysis

## INTRODUCTION

Uncertain and changing environments characterize the context which organizations have to develop their activity currently. In order to face such difficulties, one of the key aspects to win competitive advantage is to ensure that employees are committed, that they identify with the organization and accept its values and objectives as reflecting their own interests. Staff members of an organization prove fundamental to its success, especially where their satisfaction impacts on client satisfaction levels, it is necessary for organizations to perceive their employees as their first clients (Alves et al., 2015). Internal marketing should take corresponding priority over any external marketing processes (Ahmed and Rafiq, 1995; Flipo, 2007; Alves et al., 2015).

Organizational commitment has been highlighted as the primary attitudinal variable in the development of volunteer commitment and long-term retention (Stirling et al., 2011; Vecina et al., 2012) which are held to provide motivation (McCormick and Donohue, 2016). Although, there are few works, the obtained results in the frame of the Third Sector show differences with regard to the general multidimensional model as defined by Meyer and Allen (1984, 1991). One of the reason to consider is the differences between the organizations that operate across the sectors. The first is an important difference in the defining source of revenues.

The study of Commitment began with sociological theories that analyzed the impact of penalty systems on socially accepted values (Becker, 1960). But the work of Porter et al. (1974), which takes a sociological and psychological approach, was probably the origin of the study of links between the individual and the organization from the perspective of organizational behavior. Several decades later, Organizational Commitment is a complex concept that continues to be actively researched (Meyer et al., 2002, 2004; Allen, 2003; Cohen, 2003, 2007; Wasti, 2005; Ashman and Wintanley, 2006; Bergman, 2006; González and Guillén, 2008; Iqbal, 2010; Walumbwa et al., 2010; Stazyk et al., 2011; Klein et al., 2014; Reevy and Deason, 2014; Zayas-Ortiz et al., 2015; Hansen and Kjeldsen, 2017; Idris and Manganaro, 2017; Jaros, 2017; Wang et al., 2017) in attempting to define the relationships established between a person and the organization in which he/she works, due to the importance of the construct for employees and employers (Yousef, 2017). Although recently new approaches have appeared (e.g., Klein et al., 2014), the majority of researchers agree that organizational commitment should be treated as a multidimensional construct (Back et al., 2011) and that consistent correlations with other concepts vary with respect to dimensions. Notwithstanding, due to the use of different measurement scales and results regarding the internal structure there is no consensus regarding their interpretation hence the debate is still open (e.g., González and Guillén, 2008; Klein et al., 2014; Jaros, 2017).

Due to the aforementioned lack of consensus and the different results found in the body of literature regarding the Third Sector, the aim of this study is to gain an in-depth understanding of the internal structure of Organizational Commitment for Non-profit Organizations (NPOs), in order to understand the reason why people become involved in an organization and the types of bonds established between people and the organization. Considering the different models and theories concerning commitment, this work compares four structural models and the present paper is structured as follows. The first section reviews the main theories relating to organizational commitment in the body of scientific literature. Subsequently, we present our definition of commitment with different dimensions that are adapted to the field of NPOs for further analysis. Then, we present the results of the structural analysis of dimensions. Finally, the last section presents the main conclusions, taking into account the limitations of the study and the different implications for human resources management for NPOs.

### LITERATURE REVIEW

### Organizational Commitment: An Open Debate about the Nature of the Bond

During recent decades, Commitment has been defined and measured in different ways (Gupta, 2017). However, the lack of consensus regarding its definition does not imply the lack of a common body of knowledge that allows us to "distinguish it from other related constructs, such as satisfaction, motivation, implication" (Liou and Nyhan, 1994, p. 100).

Organizational commitment is defined as an emotional, moral and rational phenomenon (Ahmad and Oranye, 2010). Taking into account the different connotations as to the origin of the bond and the main differences among the different contributions, it can be seen in **Table 1** that all the authors agree that commitment is a link with the organization that involves either behavior or attitude.

Therefore, it is necessary that the link involves particular behaviors or a positive attitude toward an organization that predisposes the individual to benefit the organization (Meyer and Herscovitch, 2001). Organizational commitment is the extent to which individuals psychologically identify with their work organizations (Idris and Manganaro, 2017). The nature of the links can vary and they include desire, perceived cost and obligation to continue a course of action (**Table 1**). Although it is useful to consider commitment as spread over a range from the emotional to the instrumental perspective, these approaches reflect different underlying components of commitment and therefore scales of measurement for explaining the construct dimensions will continue to be developed. But, the generally accepted feeling seems to indicate that the consequences of the multidimensional construct links lead us to links of different individuals. These types of links could be set as follows: (1) Affects or an affective link as an affective feeling or emotional link; (2) Fear or a repressive link as a feeling of being trapped; and (3) Normative links as feelings of obligation.

The studies relating to organizational commitment from an empirical point of view and the different interpretations of the links require us to rethink their meanings, especially within the scope of third sector NPOs and particularly, considering that they have volunteer resources. NPOs do not have financial wherewithal to implement human resources policies for promoting involvement and motivation, either for workers or for volunteers (Pearce, 1993; Boezeman and Ellemers, 2008), hence they have to use other tools to attract and retain human resources in the organization. It can be one of the tools that human resources managers utilize in order to analyze employee identification with organizational goals, and loyalty linking employees to their workplace (Zayas-Ortiz et al., 2015). In the next section, taking into account the different contributions made and organizational characteristics, we present the various bonds considered for these organizations.

### Definition of Commitment: Dimension Content

In general, there is a scientific consensus regarding this conceptual delimitation. But, it diminishes as far as the taxonomy of links is concerned, or the way in which individuals feel tied to the organization. There have been several attempts to classify organizational commitment (Ahmad and Oranye, 2010). After reviewing the different contributions in the body of literature, we raise the idea that the nature of the bond generates different dimensions that, as aforementioned, can be summarized as affections, fears and obligations.

Among the different methods recognized in the field of psychology, our revision of the presented literature indicates that the previous studies usually work with models having different numbers of dimensions (**Table 2**). To the best of our knowledge, previous studies collected at least six meanings that are viewed as relevant to the level of effort demanded of the individual and his/her own acceptance (attitudinal aspects), and relate this and other factors of individual motivation with a decrease in probability of abandoning the organization. For that reason,

#### TABLE 1 | Definitions of the commitment dimensions.


#### TABLE 2 | Dimensions of the Construct Organizational Commitment.


exploration of the nature of the relationship and motivations of the individual become the fundamental factors. Next, we present the factors considered and justification for the most relevant factors selected.

#### Affective Bonds. Affections

In this first interpretation of commitment, we aim to gather the attitudes of an individual that link him/her to the organization, either because he/she appreciates the entity values, or because he/she identifies with them. It has been identified as an antecedent of organization citizenship behavior (Wang et al., 2017). O'Reilly and Chatman (1986) show that this type of attitude arises when workers behave in a certain way, because they want to remain in the organization due to its attractiveness (values and goals), even though its values may not be those that the person would adopt. The individual accepts the influence of values to establish or maintain a satisfactory relationship, hence this dimension emphasizes an aspect of socialization of an instrumental nature. The affection is shown by a feeling of pride that is generated outwards toward the reference group, which finally generates self-esteem owing to the sense of belonging within the organization. Feelings of pride and respect are seen as important motivators in the field of voluntary organizations (Boezeman and Ellemers, 2008). The type of commitment (compared to continuance and normative commitment) that is expected to be most clearly related both to organizational issues would be for example attendance, performance, and organizational citizenship behavior (Hansen and Kjeldsen, 2017) or dimensions of attitudes toward organizational change (Yousef, 2017). The most used and validated measure of organizational commitment in the body of public management literature (Hansen and Kjeldsen, 2017) is affective commitment, and it is especially relevant to volunteering, given the intrinsic motivation, non-monetizable, socio-emotional need of fulfillment and positive work experience ascribed to voluntary contributions to an organization of time, energy and expertise (Ohana and Meyer, 2016).

Although, non-profit employees often feel they are underpaid (Light, 2003; Kim and Lee, 2007; Handy et al., 2008), they may be willing to sacrifice some money (from wage, income) in order to serve a cause or specific social mission (Ohana and Meyer, 2016). In this way, the organization's affective commitment is important in a context of scarce financial resources and can help to resolve the dilemma faced by non-profit managers of how to keep employees committed without offering them the highest possible salary (Ohana and Meyer, 2016).

Secondly, and still referring to this dimension, other studies show an individual's identification with the value system of the organization. In this case, the identification reflects a behavior that is supported by internal values and goals which are also connected with those adopted by the organization. In this case, the concept involves more than mere loyalty. It implies an active relationship with the organization in reaching its goals, as a way to serve one's own interests. In this regard, we will adopt the meaning of Affective from the model of Meyer and Allen (1984, 1991), the significance of the definition given by Mowday et al. (1979), the Internalization dimension of O'Reilly and Chatman (1986), and the meaning compiled under Moral Commitment from Jaros et al. (1993).

The reason why we question the breakdown into these two subdimensions is due to the diversity of interpretations of the emotional aspect of the term. This difference can be seen in the model developed by Jaros et al. (1993) under the itemization of the affective dimension and the meaning given to the moral dimension, or in the O'Reilly and Chatman model (1986) as comprised in the dimensions of identification and internalization.

Affective commitment has been linked to measured involvement in organizational activities, a strong willingness to contribute to achieving organizational goals and a strong desire to remain with the organization (Walumbwa et al., 2010; Idris and Manganaro, 2017). It is found to be strongly related to important organizational outcomes such as attendance, turnover, performance, and organizational citizenship behavior, as well as individual outcomes such as stress and work family conflict (Meyer et al., 2002; Stazyk et al., 2011; Chordiya et al., 2017).

### Continuity Bonds. Fears

The second attitudinal dimension explains the relationship between an individual and the organization as a sense that his/her withdrawal would imply the loss of some acquired conditions or rights, or that he/she has no other labor alternative (Tekingündüz et al., 2017). According to the definition by Becker (1960), organizational commitment is associated with the assessment made by an individual of the costs involved in the abandonment of the organization where he/she works and/or the costs of renouncing a situation or status resulting from his/her efforts. Commitment is thus defined as a willingness to deploy a determined consistent line of behavior as a result of the accumulation of investments that could be lost if that line of action was abandoned (González and Antón, 1995). Consequently, in the light of commitment related to the investments made, all actions performed by an individual after becoming part of an organization will lead to attempting to justify his/her continuance (Becker, 1960; Salancik, 1977).

Along with the prior aspects, the explanation based on the continuity/fear argument, another key factor stands out: the lack of alternatives. McGee and Ford (1987) pioneered the study of commitment bi-dimensionality based on Becker's theory. These authors suggested the existence of two interpretable factors: the perceived sacrifice associated with neglect and the lack of alternatives.

The results obtained that split Continuous commitment into two dimensions are supported by studies that use discriminatory factor analysis, such as studies by Allen and Meyer (1996), Hackett et al. (1994), Iverson and Buttigieg (1999), Meyer et al. (2002), Meyer et al. (1990), and Somers (1993). Although there are other results that indicate uni-dimensionality such as those results obtained by Dunham et al. (1994), Kou et al. (1997), and Powell and Meyer (2004).

Moreover, works in the field of voluntary organizations demonstrate empirically that normative commitment does not imply a significant relationship for this type of personorganization relationships (Liao-Troth, 2001; Stephens et al., 2004; Dawley et al., 2005).

### Normative Bonds. Obligations

For this third and final dimension of reference, there is less research and empirical contributions in the body of literature than for those previously mentioned dimension, despite its importance in explaining Organizational Commitment. One of the main problems found in the definition of this dimension is the lack of consensus as to its meaning, although the work goes on to consider that the policy linkages reflect a sense of obligation (Meyer and Herscovitch, 2001). As such, a committed person will feel compelled to stay connected with the organization (normative linking). One of the factors that has contributed to the lack of clarity with respect to the affective dimension relates to the high levels of correlation with this dimension, as obtained in previous studies. The person has normative commitment while showing loyalty to their organization and they present suitable behavior and conduct with motivation for doing good for the organization (Tekingündüz et al., 2017). This dimension is still unknown and probably the most controversial dimension (Jaros, 2017) bearing in mind the implications included in previous works (Juaneda-Ayensa and González-Menorca, 2007; González and Guillén, 2008; Grant et al., 2008; Meyer and Parfyonova, 2010), we break this dimension into two types of link: Moral Obligation and Gratitude.

The need for internal consistency presented in the cognitive dissonance theory of Festinger (1957) is reflected in the concept of Moral Obligation. It describes the tendency of individuals to reconcile internal inconsistencies. Because of this desire for internal consistency, attitudes and beliefs of an individual may not only be determinants of their behavior, but a result of it. This approach assumes that attitudes are relatively private, malleable, and not always clearly identifiable. In comparison, behaviors are more public, and once acted out, irrevocable (although the consequences are not). Thus, attitudes, which are easily modifiable elements, will be molded around the least malleable factor, or behaviors, the mechanism by which the individual strives to maintain consistency between them (Oliver, 1990). This would mean that an action taken modifies an attitude if there is any inconsistency between them. According to this theory, people who are committed to participate/collaborate on a project should continue with it in order to avoid contradicting his/her line of action that was already begun. If they were to decide to interrupt it, it would be a contradiction and therefore internally inconsistent.

The second meaning of this dimension is Gratitude, defined as a sense of obligation due to the feeling of having received more than what has been given. In this case, social exchange theory is of significant importance. According to Blau (1964), relationships between two parties, when one of them provides benefits to the other, the imbalance in the relationship confers an obligation on the second party. The person feels there is an imbalance regarding benefit, and this fosters a sense of debt to the organization, i.e., strengthens the feeling of obligation to the organization and attempts to balance it, believing this behavior to be appropriate to contribute to balance in the exchange.

We believe it is important to look at this dimension in depth in the context of NPOs. In an attempt to enhance the theory of normative commitment, González and Guillén (2008) consider that the normative dimension should be grouped with rational judgments regarding the moral sphere of the individual, and include whatever has to do with moral judgments (fairness) and moral practice (responsibility), which to some extent is what we intend to compile under the dimensions of Gratitude and Moral Obligation. On the other hand, we consider that in the NPO, this moral connotation is relevant and it is particularly influenced by NPO characteristics, including the role of demands and political pressure aimed at ensuring a level of rights for a group or society in general and the role these activities play in transforming individuals. From this point of view, linking to one of these organizations is a public demonstration of certain convictions, and activities performed under the auspices of the entity support identification of the individual with the organization's ideology. In the event that there is incomplete identification between the value and belief systems of the organization and the individual, the size of the organization, and the difficulty to model this value system will indoctrinate individuals as a tool of socialization. The result will be a link with the organization as a consistent behavior based on public demonstration of an ideological system. However, we must include ideological or ethical issues in this section [the significance of the normative commitment of the Allen and Meyer (1990) scale] in which the relationship is reflected and manifested in the permanence within the organization that are generated from the value system of the individual who becomes "tied" to the acquired commitments.

When gratitude is felt toward the organization, the more relevant argument is based on the fact that most of these organizations provide services that are not provided by the public or private sector, and the organizations offer solutions to problems for people in situations of some complexity, helping them feel connected to the organization as a show of gratitude for the support received.

In addition, this type of link may have a particular impact on NPOs and as suggested by Meyer and Allen (1997), commitment characterized by a sense of obligation might be a better predictor of employment outcomes in collective contexts where social ties and regulatory obligations are most relevant.

Then, we propose a concept as a three-dimensional model which, depending on the different connotations attributed by the authors, can have each of its dimensions broken down into two sub-dimensions (**Figure 2**).

Affective commitment has been linked to increased involvement in organizational activities, involvement in organizational activities, a strong willingness to contribute to achieving organizational goals and a strong desire to remain with the organization (Walumbwa et al., 2010; Idris and Manganaro, 2017). Continuance and normative dimensions of commitment have been critiqued for their inconsistencies with affective commitment (Chordiya et al., 2017). Normative commitment is usually strongly linked to affective commitment (Guerrero and Herrbach, 2009) and is linked to individuals' sense of obligation to stay in the organization (Wang et al., 2017).

### MATERIALS AND METHODS

As stated previously, there is a question regarding the link between the individual and the organization and, although most empirical studies are based on the one-dimensional Porter et al. (1974) model or the three-dimensional Allen and Meyer (1990) model, in our case we decided to rethink the links and try to show the existence of links that have been overlooked or have had lesser attention in the body of literature on organizational commitment. The main reason for this decision was that this study was undertaken in a particular environment within which the relationships between members of the organization and the organization itself showed the characteristics we have already mentioned. Because of the complexity of the construction and due to the lack of works in the Spanish language that are adapted to Third Sector organizations, we decided to develop an ad hoc measurement tool that would allow us to achieve our goals. Hence, we developed a measurement scale based on the main previous works, in order to obtain a reliable measurement tool that would enable us to assess most accurately the level of commitment to an organization. First, we defined the dimensions and later allocated the items (**Table 3**) to the different connotations set out. We obtained a 20-item scale from experimental analysis, although we decided to add a control item that enabled us to gather the perceptions of individuals regarding their level of commitment to the organization (Sánchez TABLE 3 | Definition of organizational commitment scale.



OCQ, Organizational Commitment Questionnaire (Mowday et al., 1979). AMM, Allen and Meyer Measurement (Allen and Meyer, 1990).

and Sarabia, 1999). We used a Likert scale of 0–10 for measurement.

The analysis was performed using IBM SPSS 24 for the exploratory factor analysis and using AMOS 24 for the confirmatory factor analysis. Validation of the tool was undertaken by following considerations provided by work on testing sociometric properties required by the scales of measure, an issue that enjoys broad endorsement in the body of marketing and organization literature (Lévy and Varela, 2006; Camisón and Cruz, 2008; Hair et al., 2010).

Exploratory factor analysis was applied to the principal axis factoring method with Varimax rotation in order to compare the underlying structure of empirical data with the theoretical structure of the resulting models from the literature review and which are presented in the above. In the first extraction, the result obtained for the analyzed sample was five dimensions, which were automatically determined. But, as we wanted to test the adequacy with respect to different theoretical proposals, we decided to obtain the factor analysis for models with 3, 4, and 6 dimensions (**Table 4**). Following the recommendations of Hair (Hair et al., 2010) to develop the most suitable, commitment model, and in order to facilitate our work on the confirmatory factor analysis, we proposed comparisons among the four structures with 3, 4, 5, and 6 latent dimensions as a rigorousness test, by comparing competing models (Bentler and Bonett, 1980; Hair et al., 2010).

Parameters (standardized factor loadings) for the confirmatory factor analysis construct elements were obtained through structural equation systems. To carry out the estimation of model parameters, we used the original data matrix instead of the correlation matrix as input, because of the information available to us and because our desire is to explain the nature of the latent construct. Moreover, taking into account the lack of multivariate normality, we decided to use the Asymptotically Distribution Free (ADF) method of estimating function—guided by the considerations presented by Hair et al. (2010). One of the main drawbacks of the method is its higher demand with regard to sample size (Hair et al., 2010). In order to minimize the number of model parameters and to increase the degrees of freedom, we decided to group the observed measures of the same latent variable, the dimensions of commitment, in a composite score, the arithmetic mean, depending on the results of the exploratory factor analysis.

#### Sample

The information was obtained from 14 non-profit organizations that operate in various fields of activity, but all of them are characterized as direct services. The areas of activity are services for children, mental disability, and the promotion of employment, physical disability, and a volunteer organization. The samples were obtained by the directors of the entities. The directors distributed the questionnaires and they committed themselves to respecting the anonymity of those respondents wishing to undertake the completion of the questionnaires.

The final sample consists of 235 questionnaires, of which 156 (66.38%) pertain to workers and 79 (33.61%) pertain to volunteers. The average seniority in the organization is 4.51 years with a standard deviation of 4.67 and a maximum of 26 years and a minimum of 6 months. Sixteen percent of the sample consisted of people who stayed less than a year in the organization, 46.6%


TABLE 4 | Definition of dimensions.

were there for a period of 1–5 years, 27.7% for 5–10 years, and only 9.7% spent over 10 years in the organization. As for the level of education, the largest proportion of the sample comprised people with a mid-level university education (34.7%) followed by graduate degrees (25.4%), and only 10.9% have a level of basic studies.

#### RESULTS

The first step in the analysis was the validation of the measurement tool for the configuration proposed in **Figure 1**. The Bartlett sphericity test (χ <sup>2</sup> = 2828.781; p < 0.001) and the Kaiser–Meyer–Olkin statistic (KMO = 0.861) report that the matrix of correlations for the exploratory factorial analysis factors proves to be very good, hence, appropriate to describing the data structure (p < 0.001; KMO > 0.8; Hair et al., 2010; Lévy and Varela, 2006; Camisón and Cruz, 2008). For reliability and validity of the scale, we calculated Cronbach's α of the composite reliability and the extracted variance (Camisón and Cruz, 2008). Cronbach's α-value obtained was 0.906 (Cronbach's α > 0.7; Lévy and Varela, 2006; Camisón and Cruz, 2008; Hair et al., 2010), which is high enough to believe that our scale is reliable and the factors account for 59.6% (3 dimensions), 65.9% (4 dimensions), 71.7% (5 dimensions), and 75.8% (6 dimensions) of the variance in the original data.

Once it was verified that the requirements for using factor analysis (Hair et al., 2010) were fulfilled (**Table 5**), we performed exploratory factor analysis using the extraction method of

Principal Component Analysis with scale items that reflected the different connotations regarding the individual-organization links as previously defined.

Exploratory factor analysis showed us the different configurations in models of 3, 4, 5, and 6 dimensions (**Table 6**).

Once the factors were extracted and the models to be evaluated were defined, the confirmatory factor analysis was conducted. Before examining the estimated parameters, model adjustments were checked (**Table 7**).

As can be seen, the data show that all models have an adequate level of measurement reliability (composite reliability > 0.7; AVE

TABLE 5 | Exploratory factor analysis.


> 0.4; Cronbach's α > 0.7; Lévy and Varela, 2006; Camisón and Cruz, 2008; Hair et al., 2010). In relation to absolute fit, we can see that the 3-D model has a Chi-square value equal to 0, meaning that this would be the model that enables us to ensure the ability to reproduce the observed matrix, although this feature is not useful for us due to the difficult generalization of the models that were identified as having no degrees of freedom (Hair et al., 2010). The 4-D model also has quite a high level of global adjustment (GFI > 0.9; Chi-square sig > 0.05; Lévy and Varela, 2006; Hair et al., 2010), followed by the 5-D model which does not allow us to accept the null hypothesis of equality between the observed and reproduced matrices (Chi-square sig < 0.05), although the value of the Chi-square standard indicator shows an acceptable value. The 3-D model provides a good fit although the RMSEA value behaves with values above those indicated as appropriate (RMSEA < 0.08; Hair et al., 2010; Lévy and Varela, 2006). The indicators of incremental and Parsimony adjustment (AGFI > 0.9 and 1 < Normalized Chi-square < 5; Hair et al., 2010; Lévy and Varela, 2006) of the 4-D model are the best (AGFI = 0.996; Normalized Chi-square = 0.163) followed by those of the 5-D (AGFI = 0.92; Normalized Chisquare = 2.524) model which are also acceptable. The 6-D model does not present a good fit in any of the analyzed adjustment dimensions.

In **Table 8**, the results of confirmatory factor analysis of the presented models can be seen. The 3-D model shows two main dimensions of organizational commitment, the size of links of Affective and Regulatory type (desire of belonging and sense of obligation), both of which have Standardized Regression Weights (SRW) = 0.795, while the continuous dimension (duty to stay) is not confirmed (SRW < 0.7; Lévy and Varela, 2006; Hair et al., 2010). This model is that which is usually considered and it has been compiled from the contributions of previous work (Allen and Meyer, 1990; Meyer et al., 2002). We aimed at collecting different interpretations offered in accordance with the context to apply, although in our case the results do not confirm the continuity dimension in the way they confirm the normative dimension. In the case of the 4-D model, the Moral Obligation (SRW = 0.723) and affection dimensions (SRW = 0.719) are confirmed, followed by values close to 0.7 of the Gratitude dimension (SRW = 0.616), and again the continuity dimension (SRW < 0.7; Lévy and Varela, 2006; Hair et al., 2010) does not show a high enough value. In the following model, continuing with the breakdown of the affective dimension, we find that both dimensions are confirmed, leaving all other dimensions with standard load values below the minimum (0.7), the Continuity Factor being that with the lower value (close to 0.5).

Finally, the 6-D model shows Commitment factors as the Pride dimension, followed by the Lack of Alternatives, the Costs of Abandonment and we could accept as confirmed both Identification (0.679) and Moral Obligation (0.655), although this model cannot be considered valid and replicable due to the adjustment problems it presents.

### DISCUSSION

The results of this research make theoretical contributions to understanding the underlying nature of links between individuals and Non-profit Organizations (NPOs) and the reason why people (workers and volunteers) become involved in an organization, and managerial implications to improve the human resources management in organizations, especially in NPOs.

#### Theoretical Contributions

One of the main drawbacks in the organizational commitment study is the use of different measurement scales, and the problems inherent to cultural questions, and the linguistic adaptation of questionnaires to specific contexts. In this sense this work analyzes the meaning of the construct identifying different structures and comparing among them. From the comparative analysis of structural models, we can highlight several aspects.

First, in relation to the interpretation of the composition of the latent variable Organizational Commitment, it consists mainly of contributions from the individual's emotional bonds with respect to the organization, followed by normative and yet, the Continuity dimension values are not enough to ensure that this type of relationship creates a relevant link with respect

#### TABLE 6 | Rotated component matrix.



The bold values mean the factor loads on the factor to which they are assigned after the factorial analysis performed.

to the entity. Therefore, we cannot accept the existence of the Continuity dimension as reflected in other studies (Meyer et al., 1990, 2002; Somers, 1993; Dunham et al., 1994; Hackett et al., 1994; Iverson and Buttigieg, 1999; Powell and Meyer, 2004), but it confirms, to the best of our knowledge, the results obtained in previous research relating to the scope of voluntary activities (Liao-Troth, 2001; Stephens et al., 2004; Dawley et al., 2005).

According to previous studies (i.e., Meyer and Allen, 1984, 1991; O'Reilly and Chatman, 1986; Jaros et al., 1993; Boezeman and Ellemers, 2008) the breakdown of the Affective dimension is confirmed and we obtain empirical evidence of the difference between the links generated by feelings of Identification and those generated by feelings of Pride of ownership.

The Commitment configuration to reflect upon is related to the breakdowns made in the normative dimension. This dimension is an important motivational force that has been overlooked and underutilized (McCormick and Donohue, 2016; Meyer and Parfyonova, 2010) and as McCormick and Donohue (2016) point out, they have been the object of conceptual reconfiguration over time (Wayne et al., 2009). In recent years, it has become a moral obligation (Meyer and Parfyonova, 2010). The second theoretical implication is that feelings promoted by


<sup>a</sup>The Chi-square text (p < 0.05) for df = 2 its value is 5.991; df = 3 is 7.815, and df = 4 is 9.488 (Malhotra, 2008).

TABLE 8 | Confirmatory factor analysis.


\*\*\*p < 0.001. C.R., Critical Ratio; SRW, Standardized Regression Weights.

the need for internal consistency and meeting responsibilities acquired (Moral Obligation dimension), based on cognitive dissonance theory of Festinger (1957), are stronger links with respect to the entity than those that may be generated by feelings of gratitude toward the organization, feelings that are reflected in the Gratitude dimension and based on the social exchange theory of Blau (1964), which does not show a sufficient level of variance and hence reliability.

Finally, in this study we have aimed at adapting organizational commitment to a specific field, and one which is as peculiar as voluntary organizations. The main theoretical contribution is, regarding NPOs, the most significant links to people who work in an NPO are those that are related to affective ties. In conclusion, this theoretical implication is aligned with motivational theory on intrinsic/extrinsic motivational factors and contribute to demonstrate how intrinsic motivations are more relevant than extrinsic motivations.

#### Practical Implications

Research on commitment in NPOs underlines the role that management and human resources practices can play in fostering employee commitment (Cunningham, 2001; Alatrista and Arrowsmith, 2004). To undertake such research it is essential to identify the factors that promote the commitment and retention of NPO employees (McCormick and Donohue, 2016).

The first managerial implication is the individual's emotional bonds to the organization, defined as affective commitment, which are of great importance to NPOs. Employees and volunteers of NPOs are highly sensitive to the organization's mission and values and they strongly identify with the organization's social mission. NPO internal consumers need these affective bonds to feel committed to the organization and to bring the best to the organization (Ohana and Meyer, 2016) and these bonds influence motivation (Somers, 2010; Cohen, 2011). In considering individual links, this commitment is key to attitudes and behaviors, including higher performance, organizational citizenship behavior, as well as lower levels of turnover and absenteeism (Cunningham, 2001; Meyer et al., 2002; Ridder and McCandless, 2010). Strong employee engagement with values, missions, and goals is therefore essential to organizational success and organizational survival (Ridder and McCandless, 2010). Incorporated into the stated mission, organizational values provide guidance and justification for the decisions and behavior of members of the organization (O'Reilly et al., 1991; Lawrence and Lawrence, 2009). In this sense, in accordance with our results, the main implication for managers is that they have to declare publicly and clearly the organization's mission and values, and this declaration is a key factor in addressing the following:

Regarding volunteer engagement, this is the main aspect with which to attract and retain motivated volunteers because those volunteers feel that they identify with the organizational mission and values. Often, volunteers are motivated to join organizations on the basis of the compatibility of their individual beliefs and values with the organizational values that are adopted (Amos and Weathington, 2008; Van Vuuren et al., 2008).

Motivational aspect for employees: Non-profit employees often feel they are underpaid (Light, 2003; Kim and Lee, 2007; Handy et al., 2008) but they may be willing to sacrifice some money in order to serve a cause or specific social mission (Ohana and Meyer, 2016). In this way, in a context of scarce financial resources, our second recommendation is that human resources managers should consider the definition of job position and must establish mechanisms to promote appropriate activities in considering how they contribute to the mission of the organization. This is more effective than other retention systems such as reward systems (continuance commitment).

It is thus crucial to stress those work experiences that contribute to the feeling of belonging, and to develop an organizational culture based on common values and goals rather than on economic rewards, which can be impractical and unusual in most organizations. There is also evidence of links

#### REFERENCES


related to feelings of obligation and normative commitment, which have come to be associated with an accountability and a greater control of one's own activity that promotes long-lasting behaviors, although concerning this type of relationship there is little empirical support that backs previous results of our work. Therefore, we consider it important to deepen the knowledge of elements linking people with the organization in the context of NPOs and the relationship among them.

#### Limitations and Future Research Lines

Although this work contributes to a better understanding of the nature of links between individuals and NPOs, it has also some limitations. The main limitation of this research is the size of the sample. In future works, it should be interesting to use larger and representative samples to deepen research into the differences between employees and volunteers. Although, the internal meaning of each dimension is still an open debate, and is necessary to clarify the relation between Affective and Normative dimensions, another future research could aim to analyze the relations of each dimension with other variables (e.g., satisfaction, social performance, organization culture) and their effects on them over time. We hope new works will explore these future research lines because are essential to develop mutually beneficial and satisfactory relationships between organization and person.

### AUTHOR CONTRIBUTIONS

The three authors have equally participated in literature review, data analysis and writing of the paper. All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewers JL, SG and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.

Copyright © 2017 Juaneda-Ayensa, Clavel San Emeterio and González-Menorca. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Consumer Behavior in the Choice of Mode of Transport: A Case Study in the Toledo-Madrid Corridor

Ana I. Muro-Rodríguez <sup>1</sup> \*, Israel R. Perez-Jiménez <sup>1</sup> and Santiago Gutiérrez-Broncano<sup>2</sup>

<sup>1</sup> Econometrics Area, Department of Spanish and International Economy, Econometrics, History and Economic Institutions, University of Castilla-La Mancha, Cobertizo San Pedro Mártir, Toledo, Spain, <sup>2</sup> Business Administration Department, University of Castilla-La Mancha, Talavera de la Reina, Toledo, Spain

Within the context of the consumption of goods or services the decisions made by individuals involve the choice between a set of discrete alternatives, such as the choice of mode of transport. The methodology for analyzing the consumer behavior are the models of discrete choice based on the Theory of Random Utility. These models are based on the definition of preferences through a utility function that is maximized. These models also denominated of disaggregated demand derived from the decision of a set of individuals, who are formalized by the application of probabilistic models. The objective of this study is to determine the behavior of the consumer in the choice of a service, namely of transport services and in a short-distance corridor, such as Toledo-Madrid. The Toledo-Madrid corridor is characterized by being short distance, with high speed train available within the choice options to get the airport, along with the bus and the car. And where offers of HST and aircraft services can be proposed as complementary modes. By applying disaggregated transport models with revealed preference survey data and declared preferences, one can determine the most important variables involved in the choice and determine the arrangements for payment of individuals. These payment provisions may condition the use of certain transport policies to promote the use of efficient transportation.

Keywords: consumer behavior, choice of service, transportation, modeling, discrete choice, logit, willingness to pay

## INTRODUCTION

Transport is one of the most important services of a developed society. The growing need for mobility of people motivated mainly by the spatial differences of the locations to which people need access to makes transport modeling a matter of spatial importance for any developed country. One of the main problems to be analyzed is how individuals move or what their mobility patterns are. This issue is fundamental to the proper planning of the transport system. An efficient transport system must serve the mobility needs of individuals, for this it must use the necessary tools to be able to plan this mobility.

The individual faces, decisions daily between different alternatives of choice, whether goods or services, conditioned by the qualities or attributes of the different options available

#### Edited by:

Alicia Izquierdo-Yusta, University of Burgos, Spain

#### Reviewed by:

Cristina Peñasco, Consejo Superior de Investigaciones Científicas (CSIC), Spain Victor Martin, Complutense University of Madrid, Spain

\*Correspondence:

Ana I. Muro-Rodríguez anaisabel.muro@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 25 March 2017 Accepted: 01 June 2017 Published: 20 June 2017

#### Citation:

Muro-Rodríguez AI, Perez-Jiménez IR and Gutiérrez-Broncano S (2017) Consumer Behavior in the Choice of Mode of Transport: A Case Study in the Toledo-Madrid Corridor. Front. Psychol. 8:1011. doi: 10.3389/fpsyg.2017.01011 (McFadden, 2001b; Ye et al., 2007; Chowdhury and Ceder, 2016; Houdek, 2016). In order to determine the variables that affect their choice and to determine the probability of choosing between different available options, disaggregated demand models are used (McFadden, 1981; Schakenbos et al., 2016).

Disaggregated demand models are also called discrete choice models, due to the characteristics variable of choice that can be the mode of transport; this variable is discrete, due to the qualitative nature of the individual's response and also a finite number of responses are shown (Ben-Akiva and Lerman, 1985; Train, 2003; Navarette and Ortuzar, 2013). These models have the ability to predict individual decisions and joint decisions and thus serve as a basis for policy planning (St-Louis et al., 2014).

In order to try to predict the behavior of users regarding modal choice, we will focus on the disaggregated demand models based on the Theory of Choice, in which the traveler maximizes its utility (McFadden, 2001a; Train, 2003). These disaggregated demand models are based on the theory of discrete choice, to determine the probability of choosing the different alternatives which the individual counts with (Martín et al., 2011).

Most of the models used for travel behavior applications are based on utility theory (McFadden, 1974; Domencich and McFadden, 1975; Manski, 1977; Williams, 1977; de Dios Ortúza and Willumsen, 2001), which assumes that the preference of choice of an alternative is captured by a value, called utility, and decision-making selects the alternative in the set of choices most satisfactory (Taniguchi et al., 2014).

This concept, used by the microeconomic theory of the consumer, presents strong limitations for practical applications, since the complexity of human behavior suggests that the decision rule must include a probabilistic dimension (Simon, 1957; Sandoval, 1994; McFadden, 2001b).

Some models assume that the decision rule is intrinsically probabilistic, and even a complete understanding of the problem does not surpass uncertainty (Thurstone, 1927; Quandt, 1956; Luce, 1959; Tversky, 1969; McFadden, 1981). Others consider that the decision rules of individuals are deterministic and motivate uncertainty to the analyst's limited ability to observe and grasp all the dimensions of the election process because of its complexity (Aguado Franco, 2012).

A fundamental assumption in the choice process is that decision making is assumed to have "rational behavior" and implies that a decision maker is a "maximize" of utility and, in the same circumstances, will repeat the same choice, in addition to the transitivity of preferences (Domencich and McFadden, 1975; McFadden, 2001b; de Dios Ortúza and Willumsen, 2001; Train, 2003).

For the estimation of disaggregated demand models, survey data are required. Two types of surveys, revealed preferences, and declared preference surveys can be differentiated. The revealed preference surveys, which try to reflect the current behavior of individuals in their travel decisions and the declared preference surveys, provide us with data that try to reflect how travelers share a certain hypothetical situation (Espino, 2003).

The Toledo-Madrid corridor has HST, bus and car, as alternative modes of transport. In addition, the completion of the Atocha-Chamartín-Airport railway corridor will allow the connection between the Spanish high-speed network and Madrid's international airport. This will favor the cooperation of operators to offer better transport services to travelers (Muro, 2012): new figures of transport tickets, shorter connection times, integrated baggage management, information visualization, etc. (Eurocontrol, 2005b).

In this context, the research work on the "Toledo case" analyzes the current state of mobility in the Toledo-Madrid corridor and, on the other hand, changes the situation in a scenario of new transport infrastructures. This makes it possible to formulate policies at the level of passenger mobility to mitigate the most harmful effects of transport, such as congestion, environmental damage (INFRAS IWW, 1995, 2000, 2004; Intergovernmental Panel on Climate Change (IPCC), 1999; Aviation Environment Federation (AEF), 2000; Whitelegg and Williams, 2001; IATA, 2003; Steer Davies Gleave, 2006; CE Delft, 2008, 2011), by promoting complementarity between modes (Watkiss et al., 2001). Specifically with the promotion of intermodality between aircraft and train, the European Commission aims to address some of the negative effects of transport (Eurocontrol, 2004b).

## THE INDIVIDUAL BEHAVIOR IN THE CHOICE OF A SERVICE

One of the main objectives of social research is to determine the behavior of the consumer in the decision-making process and this is especially relevant when we are talking about services (McFadden, 2001b), whose supply is neither storable nor cumulative such as transport (Domencich and McFadden, 1975; de Dios Ortúza and Willumsen, 2001).

According to Ben-Akiva and Lerman (1985), an election can be considered as the result of a sequential decision-making process, which includes the following steps: definition of the problem of choice, generation of alternatives, evaluation of the attributes of each alternative, choice, and application of choice.

A specific theory of choice is therefore a collection of procedures and elements based on some general hypotheses. The elements that define the process of decision making are the decision maker and its characteristics, the alternatives or the determination of the options available for decision making, attributes, and decision rules, which describes the process used to choose a decision alternative (McFadden, 2001b; Ye et al., 2007; Martín et al., 2011; Chowdhury and Ceder, 2016; Houdek, 2016).

The decision-making or decision-making unit can be a person or a group of people, such as a family or an organization. If one considers a group of people as a decision maker we can ignore all the internal interactions within the group, and consider only the decisions of the group as a whole. In transport modeling, it has been customary to use aggregate models, which are calibrated with data that has been grouped or aggregated in some form (e.g., using average income per zone), although in the last decades a more disaggregated focus has been considered taking into account the decision unit, the individual (Domencich and McFadden, 1975). In this case, to explain the heterogeneity of preferences among decision makers at the individual level, a disaggregated model should include the socioeconomic characteristics of each decision maker, such as age, sex, and income (Collins and Chambers, 2005).

The analysis of individual decision-making requires not only the knowledge of what has been chosen, but also of what has not been chosen. Therefore, it is necessary to make hypotheses about the available options, or alternatives, that an individual considers during the election process. The set of considered alternatives is called the set of choice. We can differentiate two types of set of choice: continuous and discrete choice sets (Ben-Akiva and Lerman, 1985). The first case defines the set of choice taking into account all possible combinations of goods and services available to the individual<sup>1</sup> . On the contrary, a discrete choice set contains a finite number of mutually exclusive alternatives that can be specifically enumerated. The choice of a means of transport is a typical example of a choice of a discrete choice set<sup>2</sup> (Taniguchi et al., 2014).

The identification of the list of alternatives is a complex process, since not all alternatives will be available all the time for all decision makers. Therefore, it is necessary to differentiate between the universal set of choice, which contains all possible alternatives and the set of choice. The latter is a subset of the universal set of choice available to a particular individual, since alternatives to the universal choice set not available to the individual are excluded.

Normally, to determine these alternatives, deterministic criteria of availability of alternatives are used (for example, having a driving license determines the availability of the alternative of car). In addition to availability, knowledge of the existence of an alternative is a very important factor for the choice made. In this sense, Ben-Akiva and Lerman define that the causes of an alternative is feasible depending on availability in the market, budgetary constraints, available time or other informal factors such as knowledge of the service of a given alternative (Ben-Akiva and Lerman, 1985; Chowdhury et al., 2015; Schakenbos et al., 2016).

Behavior aspects sow uncertainty in the production of alternatives and motivate the use of probabilistic models of choice generation that allow us to obtain the probability of each alternative in the universal set (Swait, 1984).

An attribute is a trait that characterizes an alternative, with a certain value for individuals. The neoclassical economic theory considers that the individuals choose an alternative, from the different amounts of the goods that are included in them. In contrast, the discrete choice theory considers each alternative as a set of attributes and in function of these the attraction of an alternative is expressed by an attribute vector, called the utility function (Ben-Akiva and Lerman, 1985). Moreover, each alternative has its own attributes, and these can be generic for all alternatives or specific for a specific alternative. An attribute is not necessarily an objective quantity that can be measured, but can derive from a subjective measurement from perceptions. Just as the characteristics of the decision maker are taken into account, the analyst must include the attributes of each alternative (Sandoval, 1994).

In decision making, the last of the elements is the decision rule, defined as the process used in decision making, to evaluate the information available on the attributes of each alternative, in the set of choice and to determine a single choice. There are numerous decision rules in the literature that can be grouped into the following four categories (Ben-Akiva and Lerman, 1985):


These four decision rules, at the same time can be grouped into two types of behavior: compensatory or non-compensatory behavior. The compensatory type takes into account the set of all the attributes, so that changes in one or more attributes can be compensated by others, and in case of a decrease in one attribute can be compensated for an improvement in another, as is the case of the utility rule. In non-compensatory behaviors, rules or levels are defined to restrict the choice of some alternatives. Among these are the first three decision rules described (McFadden, 2001b).

The classical (neoclassical) and discrete-choice theories analyze individual consumer behavior (Ben-Akiva and Lerman, 1985), from a utility function that individuals must maximize, based on the analysis of revealed preferences, although with important differences between them, in the definition of decision rules and alternatives.

Most transport choice models have been based on this behavior (utility maximization), although other decision rules have been tried (Cantillo and Ortúzar, 2005), especially for the case of the declared preferences, where the choices that do not follow the principles of maximum utility are measured (Sælensminde, 2002; Rouwendal and Blaeij, 2004).

The mobility of people is a complex phenomenon, due to the large number of factors that influence the decisions

<sup>1</sup>Based on the neoclassical theory of consumer behavior.

<sup>2</sup>From this set of decision we derive discrete choice models, which we will see below.

of individuals. For this reason, it is necessary to analyze the displacements according to three interacting aspects a spatial approach, a social approach and a perspective approach (Rodríguez, 1991).

The spatial approach focuses on the different land use and the distribution of activities in space. This creates the necessity of the individual of the individual to move, depending on the location of the activities you want to perform.

The social approach refers to the fact that an individual's movements are exclusive of their characteristics and are therefore the result of the socioeconomic characteristics of the individual who performs it. Because of this it is necessary to have very present variables like the sex, the age, cultural level, income, health, etc.; for the analysis of the mobility.

Finally, we can speak of a perceptive approach, which refers to the image that the individual has formed. This image varies in each individual or group of people, leading to different assessments of the same data in decision making.

In the classic model a sequence of decisions of the individuals is done in four stages which are the generation of trips, zonal distribution, modal distribution, and assignment, but another sub-model has recently been included that is the last one mentioned, the election of the hour that gives place to the models of time distribution.

At present, this method is recognized to be too strict, since the decisions of the individuals are not taken following this sequence, but each sub-model depends on the type of the function of utility assumed to explain all these choices of trip. There are some current approaches that differ from this four-step sequential methodology and simultaneously address the stages of choice of frequency, destination, and mode of travel, but they are still at a level of research and have not been implemented. In addition, we must highlight models based on the activities of families or decision-making centers, which take into account constraints on budget and time choice (de Dios Ortúza and Willumsen, 2001) stand out.

### METHODOLOGICAL FRAMEWORK FOR MODELING CONSUMER CHOICES

From an economic perspective, consumer theory is the economic modeling of the behavior of an economic agent that consumes goods and services. This theory relates preferences, indifference curves and budget constraints to consumer demand curves.

The fundamentals of the individual choice theory of the neoclassical model of consumer behavior are based on the fact that the individual chooses an amount of goods, which form their basket of goods, in order to maximize their utility which translates into their level of satisfaction, subject to their income restriction (Torres, 2008). Therefore, given a consumer whose preferences fulfill these assumptions, there is a deterministic utility function, U, which represents these preferences that will be ordered so that the individual will choose the alternative that gives him greater satisfaction<sup>3</sup> .

Given that consumer economic theory has been developed without taking into account the nature of the alternatives, some extensions have been made such as Strotz (1957, 1959), which analyzes consumer behavior by introducing the concept of "utility tree" (Strotz, 1957, p. 270). Muth (1966) proposes that the goods and services of the market are considered by the consumer as inputs of a domestic production function, whose production is the satisfaction of the consumer needs derived from consuming certain goods or Becker (1965) that extends the traditional theoretical formulation by adding the time constraint, in addition to the budget, to the production function.

Finally, one of the most important extensions is that one made by Lancaster (1966). This author considers three assumptions that break with the traditional approach and states. That (1) goods per se do not contribute to usefulness, but are the characteristics they possess, which provide usefulness to the consumer, (2) a good has more than one characteristic, and many of these characteristics can be shared by more than one good; (3) and through the combination of different goods different characteristics can be obtained from those corresponding to the separate goods. These assumptions represent the definition of utility in terms of the attributes of the goods.

Taking into account all these assumptions, consumer economic theory has important limitations to explain the consumer behavior, which justifies models based on Discrete Choice Theory. This reformulation of the behavior of the consumer will not occur until 1981, with the inclusion of consumer goods of a discrete nature (McFadden, 1981), with which the Theory of Discrete Choice begins, whose characteristics are that:


The problem that arises when using Discrete Choice Theory is that it represents a mechanism of deterministic choice and these deterministic election mechanisms do not fit the analysis of choice problems in real situations (Block and Marschak, 1960). In this line of study, this approach has been widely criticized, both in the field of psychology by Thurstone (1927), Luce (1959), and Tversky (1969), as in the field of economics by Quandt (1956) and McFadden (1981). These authors justify the inclusion of uncertainty in modeling, being the source of randomness determinant when specifying the different models.

Therefore probabilistic mechanisms of choice are required in order to analyze individual options. In order to do this, we use Probabilistic Election Theory, specifically to the Theory of Random Utility formalized by McFadden (1974, 1981), Domencich and McFadden (1975), and Manski (1977) and whose most recent developments are by (de Dios Ortúza and Willumsen, 2001).

The Random Utility Approach assumes that the individual always selects the most useful alternative, although the utilities are not known to the analyst and are treated as random variables. In mathematical terms, this is expressed by separating the total

<sup>3</sup>This is possible under the assumption of rational consumer behavior, namely perfect rationality (Simon, 1957).

utility Uin into a deterministic component Vin, called Systematic Utility and a random component ε<sup>i</sup> that captures uncertainty, called Random Perturbation:

$$U\_{jq} = V\_{jq} + \varepsilon\_{jq}$$

where:

Vjq is the systematic component that represents an appreciable part of the utility of the individual q to choose alternative j. Systematic utility depends on the attributes of the alternative and the socioeconomic characteristics of the person.

εjq is the random term that accounts for unobserved factors.

The individual chooses the most useful alternative. Alternative i is chosen, if and only if:

$$P\_{iq} = \operatorname{Prob}\left(U\_{iq} > \left. U\_{jq} \right|\right) \forall j \neq i$$

The analyst can only obtain the probability of choosing alternative i as:

$$\begin{aligned} P\_{i\dot{q}} &= \mathit{Prob}(V\_{i\dot{q}} + \varepsilon\_{i\dot{q}} > V\_{j\dot{q}} + \varepsilon\_{j\dot{q}} \forall j \neq i) \\ P\_{i\dot{q}} &= \mathit{Prob}(\varepsilon\_{j\dot{q}} - \varepsilon\_{i\dot{q}} < V\_{i\dot{q}} - V\_{j\dot{q}} \forall j \neq i) \end{aligned}$$

This is the probability that each random term εnj – εni, is less than an observed quantity Vni − Vnj, so it is a cumulative distribution function. From the density function f(εn) we can calculate this probability as:

$$P\_{ni} = \operatorname{Prob} \left( \varepsilon\_{nj} - \varepsilon\_{ni} < V\_{ni} - V\_{nj} \,\forall \, j \neq i \right)$$

$$= \int\_{\varepsilon} I \left( \varepsilon\_{nj} - \varepsilon\_{ni} < V\_{ni} - V\_{nj} \,\forall \, j \neq i \right) f \left( \varepsilon\_{n} \right) d\varepsilon\_{n}$$

Being I (·) a function that has a value of one if the term given in brackets is true (if the individual has chosen alternative i) and zero in another case. According to the different hypotheses formulated about the distribution, εnj will have a closed value of this integral (Simple Logit or Hierarchical Logit) or it will have to be evaluated numerically by simulation (Probit or Mixed Logit), and therefore the different Models of discrete choice (Train, 2003).

Defined the set of choice, we can proceed to define the utility function that will give rise to different types of models. The typology of this type of model is broad and can be classified (Medina, 2003 according to the number of alternatives of the endogenous variable (dichotomous response models and multiple response models) or according to the type of function to estimate the probability (Linear Probability Model, Logit Model, and Probit Model): Whether the alternatives are exclusive or incorporate ordinal information (models with non-ordered data and models with ordered data) and whether the regressors refer to aspects of individuals or to alternatives in non-ordered models (multinomial and conditional models).

For the theoretical justification of the models of discrete choice based on the Theory of Random Utility, we can find three types of models of discrete choice depending on the different hypotheses that are taken for the distribution of the random term. First the linear probability model, assuming a uniform distribution, then the Probit Model, assuming a normal distribution and in third place the Logit Model, assuming a logistic distribution.

The Linear Probability Model is the linear fit regression model that is applied to a binary dependent variable. This model is estimated by ordinary minimums to the square and is easy to estimate and interpret. The estimated parameters measure the predicted change in probability of success, vs. a unit increment of X<sup>i</sup> . Although there are problems in the estimation of the regression model when the endogenous variable is binary. The specific problems encountered with this model are heteroskedasticity of the perturbation term, the predicted probabilities are inconsistent, since it cannot be guaranteed that they are bounded between zero and one, the non-normality of the perturbation and the coefficient of Determination is not appropriate. Due to these problems, what is interesting is a model that reproduces properly the behavior of a probability function and the alternatives we find the Logit model and the Probit model, similar numerically. When the normal distribution is used as a probability function, the so-called Probit model is obtained, while the use of the logistic distribution provides the Logit model.

The main advantage of the Probit model is its ability to capture all correlations between alternatives, but its main problem is the complexity of its formulation, so there are very few applications that have been developed (Daganzo, 1979) and more than three alternatives are used for the calculation Simulation procedures (Train, 2003).

The logistic regression model, also called Logit models, is much more popular thanks to its analytical flexibility. Thus, the possible hypothesis for deriving some Logit models from the operation of individual choice is analyzed here: the most general case of interest, the Logit Multinomial model (McFadden, 1974), in its particular case, beyond binary situations; The Logit Hierarchical or Nested model (Williams, 1977) and the Logit Mixed model (McFadden and Train, 2000).

### MODELING THE DEMAND FOR SERVICES: APPLICATION TO THE CHOICE OF TRANSPORT SERVICES

Transport contributes significantly to the economic development and allows the market to function in a global way. It should be noted that most modes of transport do not affect society only in a positive ways, there are negative side effects such as congestion, noise and air pollution.

The promotion of intermodality using airplanes and trains is intended to solve some negative effects of transport, such as the impact of congestion on the environment, economy, safety and passengers (Eurocontrol, 2004b). For example, regarding to the access to the airport with the promotion of intermodality, beneficial effects are expected on the economy, especially in the regional economy, and on the environment are expected (Watkiss et al., 2001; Li and Hensher, 2012).

The estimation and internalization of external costs of transport have been important issues for research in the transport sector in the European Union. A Concern for the environment has grown over the last decade and the European Commission has raised the issue of internalization in several of its directives, such as the 1995 Green Paper (European Commission, 1995), the White Paper on 2001 (European Commission, 2001), and the Mid-Term Review in 2006 (European Commission, 2006). The latter two, underline the need for equitable and efficient pricing in terms of external costs. On the other hand, although the impact on the environment or congestion are indisputable, in the literature, there are some reports that try to quantify them, some of them are: (Intergovernmental Panel on Climate Change (IPCC), 1999; Aviation Environment Federation (AEF), 2000; INFRAS IWW, 2000, 2004; Whitelegg and Williams, 2001; IATA, 2003; Steer Davies Gleave, 2006; CE Delft, 2008) and (CE Delft, 2011).

In general, all of these investigations seek to establish pricing systems that capture the external costs associated with transport, so as to reduce negative impacts, in order to improve the efficiency of the transport system, to ensure fairness between modes and improve safety and sustainability.

In recent years, considerable progress has been made in the construction of transport demand models based on the theoretical principles of choice. Within random utility theory, it can be shown that the structure of the models depends on the perceived similarity between discrete choice alternatives. Moreover, this aspect can be interpreted mathematically in terms of the correlation between the components of random utility functions. The use of Logit-type disaggregated demand models allows us to determine the likelihood of choosing one way or another depending on the previously defined attributes, as well as quantifying the willingness to pay of these subjects based on the different characteristics previously given.

To do this, we must first determine which are the parameters that measure the utility of the services or that define the attributes of the service (Martín et al., 2011; Navarette and Ortuzar, 2013; Chowdhury et al., 2015; Schakenbos et al., 2016; Ettema et al., 2017). In order to select the mode of transport, the factors influencing the choice of mode can be classified into three groups (Eurocontrol, 2004a) by the characteristics of the mode of transport (vehicle availability and/or ownership, driving license, home structure, income), the trip characteristics (the mode of choice is heavily influenced by the purpose of the trip, time of travel), and the characteristics of transport infrastructures and services (relative travel time, relative monetary cost, comfort, and convenience, reliability and regularity, protection, and safety).

#### Data Sources

The information necessary for the estimation of these models is obtained through surveys of the mobility of the different individuals at a given moment (Martínez and Muro, 2011). Two types of surveys can be differentiated into revealed preferences and declared preferences.

Revealed preferences (RP) show data on the current behavior of individuals and give us information about their travel decisions. These provide us with descriptive information about the characteristics of the traveler on a particular route. Until the mid-1980s this type of data was the most used to model the transport demand. The main drawbacks to the use of this data in the modeling are the costs of the sample and its limited capacity to understand the behavior of the traveler, the observations of the actual elections may not have sufficient variability for the construction of good models. Because of this it is difficult to be able to detect the relative importance of certain factors on behavior; (de Dios Ortúza and Willumsen, 2001; Espino, 2003; Espino et al., 2006). For the Toledo-Madrid case, information is obtained on the type of mobility of travelers in the current situation with a socio-economic profile and mobility parameters (Martínez and Muro, 2014).

The stated preferences (SP) give information on the behavior of the individual before certain hypothetical situations raised by the researcher (Navarette and Ortuzar, 2013). Unlike RP data, which give information about the trips that an individual habitually makes, they inform about the trips that the individual would realize in certain conditions. The SPs started in the field of market studies and began to be applied to the field of transport demand analysis in the late 1970s. The possibility of designing SP experiments allows, theoretically, to solve the problems they present The RP (de Dios Ortúza and Willumsen, 2001). One of the problems presented by these models is that the analyst cannot assure that the individual performs what he has answered in an SP survey, so it is important to construct realistic alternatives and design comprehensible exercises to present the individual.

In an SP exercise, three elements can be mainly distinguished (Espino, 2003). The first are the situation in which the individual finds himself to declare his preferences which can be a real situation (a journey that is carried out at that moment) or hypothetical (a journey that would take place in the future given a series of conditions), and constitutes the context of the decision. Second, the alternatives that are usually hypothetical are selected, although some of them might exist today, they are presented in the exercise as a function of a set of attributes. Third, it defines how individuals can state their preferences. The most frequent techniques are the hierarchy of the answer, the punctuation of the alternatives and the choice between alternatives.

In the case of Toledo-Madrid, the different alternatives that the user has to carry out this journey are traveling by public transport, either by High Speed Train (HST) or by regular bus (BUS), or by private transport, Private vehicle (CAR).

The basic methodology for the presentation of the options of the declared preferences (SP) is the realization of an efficient design of the scenarios. Specifically, the declared preference questionnaire is designed following a time and price reduction scheme, in three blocks or scenarios, to determine the variation of the individuals' choice. The SP surveys are structured with two approaches: a first comparative approach between the bus and the HST and a second approach comparing the car and the HST. The scenarios are conformed with the variables of trip price, travel time, comfort and cost of parking, this last variable referred only to the case of the car.

The SP survey proposes a simulation exercise is exposed of alternative scenarios that layout the mobility in the future. For the elaboration of these scenarios the combination of price and travel time variables are used. Prior to the establishment of the options, a characterization of the services and infrastructures currently available for each mode of transport is made. This current quote will serve as a basis for calculating the different alternatives for the arrival to Madrid Airport from Toledo and choose the best.

Thus, the methodology followed was to calculate the price and average travel time of each mode of transport from the origin, which in this case is Toledo, to the destination in the current situation and compare it with the different hypothetical scenarios. Therefore, the scenarios that are presented in the SP questionnaires are calculated based on the average prices and times calculated in that route, and compared in each pair of transport modes (Muro, 2012).

Consequently, three route scenarios are established:


The chosen bus and bus (CAR) optimum alternatives are fixed and compared in blocks of three, according to the three scenarios raised for the HST. The prices and travel times for the mode of transport bus and private vehicle are constant for all scenarios and are based on current values. The prices of the bus are calculated according to ticket prices and the prices of the private vehicle are calculated by applying a cost per kilometer journey that incorporates the maintenance of the vehicle, insurance, etc., adding to the cost of the toll. This price does not include parking costs since it is presented as a different variable measured in euros per day of parking.

In summary, for the calculation of these values of time and price, the following hypotheses have been taken into account:


From the econometric point of view, the main difference of both types of data (RP and SP) are the types of error that each present. The RP data suffer from errors in the measurement of independent variables and those of SP present errors in the dependent variable, motivated by the doubt about whether the individual will actually perform what he is declaring. The joint estimation of both types of data is based on the different errors mentioned and can be specified if we consider error terms with different variance. Ben-Akiva and Morikawa formulate mixed data estimates that combine revealed preferences and stated preferences with the aim of avoiding their disadvantages and taking advantage of different sources of information (Ben-Akiva and Morikawa, 1990).

#### Specification and Estimation

In order to explain the mathematical approach of the models of discrete choice applied to the case of transport choice, one can start with the assumption based on the choice of the individual between two available alternatives (binary response models), and then extend this assumption to the j alternatives available (multiple response models). The dichotomous response models have two categories and these usually indicate that an event has occurred, that some feature is present or that an option is chosen.

In the modeling of the choice of transport, the objective will be to analyze the choice of the different alternatives that the user has to make this journey, and these are for example to travel by public transport, either in High Speed Train (HST) or by regular bus (BUS), or by private transport, in a private vehicle (CAR).

As this study focuses on the probability of choosing a mode of transport, the available alternatives will be the high-speed train against that the bus and the car. This allows specifying and estimating binary logistic regression models that will serve as the basis for a first analysis of the goodness of the database. Thus, two models are specified that will have a functional form of binomial logistic regression model, whose alternatives will be HST-BUS and HST-CAR, respectively.

The functional form of the binomial logistic regression models with the HST-BUS / HST-CAR alternatives will be represented by:

$$P(a) = P(U\_a \ge U\_r) = \frac{e^{\alpha V\_a}}{e^{\alpha V\_r} + e^{\alpha V\_a}}$$

Where U is the total utilities of the high-speed train (a) and bus/car (r), respectively.

If the deterministic utility component, V, is assumed to be linear with regard to the parameters, as in most disaggregated model approaches, then:

$$P(a) = \frac{e^{\alpha \bar{\beta} \cdot \bar{\chi}\_a}}{e^{\alpha \bar{\beta} \cdot \bar{\chi}\_r} + e^{\alpha \bar{\beta} \cdot \bar{\chi}\_a}}$$

where:

β¯, is the vector of coefficients.

X¯ , is the vector of independent variables.

So if we have two alternatives, for example the binomial Logit model will be represented

$$Prob(Y\_i = 1) = \frac{1}{1 + e^{-(\alpha + \beta\_k X\_{ki})}} = \frac{e^{\alpha + \beta\_k X\_{ki}}}{1 + e^{\alpha + \beta\_k X\_{ki}}}$$

$$Prob(Y\_i = 1) = \frac{1}{1 + e^{-\alpha - \beta\_1 X\_{1i} - \beta\_2 X\_{2i}}} = \frac{e^{\alpha + \beta\_1 X\_{1i} + \beta\_2 X\_{2i}}}{1 + e^{\alpha + \beta\_1 X\_{1i} + \beta\_2 X\_{2i}}}$$

If we suppose that α = 1, since in the case linear utilities with regard to the parameters, the parameter α cannot be distinguished from the general scale of the β.

$$P\left(a\right) = \frac{e^{\bar{\beta}\overline{X\_a}}}{e^{\bar{\beta}\overline{X\_r}} + e^{\bar{\beta}\overline{X\_a}}}$$

Taking the logarithm on both sides, from the equation above, we obtain a formulation that is suitable for applying linear regression techniques:

$$\ln \frac{P(a)}{1 - P(a)} = \overline{\beta} \left( \overline{X\_a} - \overline{X\_r} \right).$$

Since the probability of choosing the alternative r is given by: P (r) = 1 − P (a)

The final functional form of the model with an additive linear utility function is as follows:

$$\ln \frac{P(a)}{P(r)} = \overline{\beta} \left( \overline{X\_a} - \overline{X\_r} \right)$$

The main explanatory variables that are considered in the transport demand analysis are, differential between the attributes and the socioeconomic variables (de Dios Ortúza and Willumsen, 2001; Cokasova, 2003; Espino, 2003; Eurocontrol, 2005a; Steer Davies Gleave, 2006; Muro, 2012) which are:

#### (1) Attribute variables:


The price variable price measured in euros will be the price paid for each of the means of transport used as public transport, and for the private vehicle, the cost of gasoline based on the kilometers traveled, including the cost of the amortization of the vehicle (insurance, maintenance, etc.). The travel cost variable represent the price paid, in euros, for the route according to the form of transport chosen. In addition to the private car, the price of the car is different, so the total cost of the car will be the sum of the cost of the trip plus the cost of parking.


The total travel time is one of the most important factors that the traveler takes into account for choosing a mode of transport (Steer Davies Gleave, 2006). It is therefore a major factor to increase the demand, as travel time is reduced (IATA, 2003). The variable of travel time measured in minutes can be represented as the total time of the trip or by the time dedicated to each of the stages of the journey, such as waiting time, transfer, etc. The variable time considered in this study is the total time spent on the trip done by the chosen mode of transport.


This variable measures the level of service and is represented by dummy variables, which include values whose coding contains the different levels of service perceived by the individual. In this case, comfort tries to measure how the individual perceives the transshipments that are presented as hypothetical scenarios of the future services of the HST to Madrid airport (for the rest of the alternatives they would be the current travel options). Comfort is the dummy variable defined as the number of transfers chosen for the alternative train. So that in the first scenario, with no transfers, the comfort is 1; in the second scenario, with one trans for, the comfort is 2; and the third scenario, with two transfers the comfort is 3. If you do not choose the HST option the comfort level will be 0.

#### (2) Socioeconomic variables:


The variable sex is included as a dichotomous dummy variable, to analyze the influence of them in the decisions



This variable is expressed in four ranges: the value 1 being from 18 to 33 years; 2 from 34 to 49 years; 3 from 50 to 64 years, and 4 over 65 years. So we are in the specification of three dummy variables that represent value one when it meets that range and zero otherwise.

The utility specification is expressed as a Logit Binomial model, to measure the choice of mode of transport between the HST/BUS (MNL1) alternatives, in the Toledo-Madrid corridor (Madrid Airport):

Utility HST1: UHST1 = ASCHST1 + θPPHST1 + θTTHST1

Utility BUS: UBUS = ASCbus + θPPbus + θTTbus + PθCj C Being,

ASCHST<sup>1</sup> and ASCbus, the specific constants of each alternative, in which ASCHST<sup>1</sup> is fixed.

θP, θT, and θCj are the parameters associated to each of the explanatory variables. In this specification, the two most important variables (cost, θ<sup>P</sup> and time, θT), where C, refers to all socioeconomic variables and characteristics of travel such as income, sex, age, motive travel, and comfort.

Cj, is a generic variable that represents each and every one of the variables not explicitly collected but mentioned above, where j refers to S, sex; ING to income; and COM, comfort depending on the transfers.

PHST, Pbus, is the price of alternatives HST1 and Bus, respectively.

THST, Tbus, is the travel time of the alternatives HST1 and Bus, respectively.

The specification of the utilities of the two alternatives of the HST-CAR Model (MNL2) will be:

Utility HST2: UHST2 = ASCHST2 + θPPHST2 + θTTHST2

P Utility CAR: UCAR = ASCcar +θPPcar +θTTcar +θPA PAcar + θCjC<sup>j</sup>

Being, ASCHST2; ASCcar, the specific constants of each alternative.

θP; θT; θPA, and θCj are the parameters associated to each of the explanatory variables.

Cj, is a generic variable that represents each and every one of the variables not explicitly collected but mentioned previously.

PHST, Pcar, is the price of alternatives HST2 and CAR, respectively.

THST, Tcar, is the travel time of alternatives HST2 and CAR, respectively.

The results of the binomial model estimates are shown in **Tables 1, 2** (Bierlaire, 2003) 4 .

The MNL1 models **Table 1** are the result of the estimation of the data of the alternatives HST1 and BUS. The MNL1.1, MNL1.2, MNL1.3, MNL1.4, MNL1.5, MNL1.6, and MNL1.7 models are variants of specified models with different explanatory variables. In general, these variables are statistically significant, as it shows in **Table 1**, and with correct signs (negative values for its inverse relationship).

The results of the first binomial models show low determination coefficients, which are a reflection of the inadequate fit of the Logit Binomial model at around 1% in the case of the HST-BUS alternatives (MNL1.1: 0.057; MNL1.2: 0.080; MNL1.6: 0.81). This can be caused by the lower amount of people who chose the BUS option vs. the HST option.

The MNL2 models **Table 2** are the results of the estimation of the data of the alternatives HST1 and BUS. The MNL1.1, MNL1.2, MNL1.3, MNL1.4, MNL1.5, MNL1.6, and MNL1.7 models are variants of specified models with different explanatory variables. In general, these variables are statistically significant, as it is shown in **Table 2**, and with correct signs (negative values for its inverse relationship), except in the MNL2.2 where we can find the travel cost with positive values. For this reason this model is nullified and in the final models only the cost of the total trip is included.

The MNL2 models (**Table 2**) are the results of the estimation with the data of the alternatives HST2 and CAR. The MNL2.1, MNL2.2, MNL2.3, MNL2.4, MNL2.5, MNL2.6, and MNL2.7 models are the same as the previous variants of the original. These models show determination coefficients around 53% (MNL2.1: 0.522, MNL2.3: 0.522; MNL2.7: 0.531).

From the results of the estimated models we proceed to calculate the payment arrangements of the users. These payment arrangements are calculated as the quotient of the parameters estimated for time and price, so they are interpreted as the willingness to pay or waiting to save in travel time (Espino et al., 2006).

The provisions for the payment of the binomial HST-BUS, fluctuate between 0.397 and 0.431, which means that individuals would be willing to pay from 23 to 25 € to save an hour of travel.

In the case of Logit Binomial HST-CAR models, **Table 2**, the results are very high being 600 € per hour what is saved in travel time. This may be motivated by individuals who answered in a lexicographic form, in which regardless the options, the users have chosen car, although this was but in all the options and by the difficulty of capturing the usual users of the car.

Next, if we assume that there are three available alternatives of choice for the individual in this path, a model is proposed with three alternatives, which are HST, bus and car, whose functional form will be that of a multinomial logistic regression model.

The specification is based on the approach of four alternatives, two from each of the sub-databases:

Utility HST1: UHST1 = ASCHST1 + θPPHST1 + θTTHST1 Utility BUS: UBUS = ASCBUS + θPPBUS + θTTBUS

Utility HST2: UHST = ASCHST2 + θPPHST2 + θTTHST2 Utility CAR: UCAR = ASCCAR + θPPCAR + θTTCAR + θPA PACAR + PθCjC<sup>j</sup>

The joint treatment of both databases requires, in the specification of the multinomial model, a specific treatment and equal to the problematic applicable to Mixed data (RP and SP).

Thus, the Logit Multinomial (MNL3) model with the HST-BUS-CAR alternatives will have the form:

The probability of choosing each of the alternatives will be:

$$Prob(Y\_i = 1) = \frac{1}{1 + e^{\alpha\_2 + \beta\_{12}X\_{1i} + \beta\_{22}X\_{2i}} + e^{\alpha\_3 + \beta\_{13}X\_{1i} + \beta\_{23}X\_{2i}}}$$

$$Prob(Y\_i = 2) = \frac{e^{\alpha\_2 + \beta\_{12}X\_{1i} + \beta\_{22}X\_{2i}}}{1 + e^{\alpha\_2 + \beta\_{12}X\_{1i} + \beta\_{22}X\_{2i}} + e^{\alpha\_3 + \beta\_{13}X\_{1i} + \beta\_{23}X\_{2i}}}$$

$$Prob(Y\_i = 3) = \frac{e^{\alpha\_3 + \beta\_{13}X\_{1i} + \beta\_{23}X\_{2i}}}{1 + e^{\alpha\_2 + \beta\_{12}X\_{1i} + \beta\_{22}X\_{2i}} + e^{\alpha\_3 + \beta\_{13}X\_{1i} + \beta\_{23}X\_{2i}}}$$

Therefore, a multinomial Logit model (MNL3) is specified, with sub-bases 1 and 2, so that a specification is made with the mixed data processing with different errors and therefore variances, so that ε is the stochastic error Of the HST-BUS data (Base 1) and η the data of the base HST-CAR (Base 2). It can be expressed as: σ 2 <sup>ε</sup> = µ 2 . σ 2 η , where µ is an unknown parameter.

So the utilities of both databases are expressed

$$\begin{aligned} U\_j^{B1} &= \, ^t V\_j^{B1} + \, \_j \boldsymbol{\varepsilon} = \, ^t \boldsymbol{\alpha} \, ^t \boldsymbol{Y}\_j^{B1} + \, \_j \boldsymbol{\varepsilon}\_j \\ \mu \boldsymbol{U}\_j^{B2} &= \, ^t \mu (\boldsymbol{V}\_j^{B2} + \boldsymbol{\eta}\_j) = \, \_t \boldsymbol{\alpha} \, (\boldsymbol{\theta} \, \boldsymbol{X}\_j^{B2} + \boldsymbol{\omega} \boldsymbol{Z}\_j^{B2} + \boldsymbol{\eta}\_j) \end{aligned}$$

where:

θ, α, and w are the parameters to be estimated.

X BI j and X B2 j are common attributes of alternative j for databases 1 and 2, respectively, while Y B1 j and Z B2 j are noncommon attributes of alternative j for each data set.

By multiplying the utility function of the Base 2 data by the unknown parameter µ, what is obtained is that the stochastic error of this type of data has the same variance as the data of Base 1.

Taking into account these utility functions we are allowed to homogenize the type of error homogenized, due to the multiplication of the SP parameters where by multiplying the utility function of the sub-base data 2 by the unknown parameter µ, we ensure that the stochastic error of this type of data has the same variance as the data of sub-base 1.

In mixed data, it is assumed that SP data should have more noise than RP data. If this is the case, the value of µ, that is known as the scale coefficient of the model, will be between 0 and 1. If the value is >1 it would indicate that the data with the highest noise level are those of RP. In our case and due to the errors detected, we assume that base 2 (HST-CAR) will have more noise (Ben-Akiva and Morikawa, 1990) 5 .

<sup>4</sup>BIOGEME. http://biogeme.epfl.ch/

<sup>5</sup>Assuming that the errors distribute as Gumbel with zero mean and different variance.

#### TABLE 1 | Estimated logit binomial model results MNL1 (HST1-BUS).


Value is the estimated coefficient for each of the explanatory variables included in the model. T-test: \* 90%, \*\* 95%, and \*\*\* 99% level of significance. It is the statistic that allows to contrast the null hypothesis of individual non-significance of each of the variables included in the model, (the contrast is performed through the level of significance associated with that statistic, so that if the level of significance is greater than 0.05 the variable is not statistically significant).The MNL1 Models are the results of the estimation with the data of the alternatives HST1 and BUS. The MNL1.1, MNL1.2, MNL1.3, MNL1.4, MNL1.5, MNL1.6, and MNL1.7 models are variants of specified models with different explanatory variables and the different specifications represent a robustness check.

TABLE 2 | Estimated binomial model results MNL2 (HST2-CAR).


Value is the estimated coefficient for each of the explanatory variables included in the model. T-test: \* 90%, \*\* 95% and \*\*\* 99% level of significance. It is the statistic that allows to contrast the null hypothesis of individual non-significance of each of the variables included in the model, (the contrast is performed through the level of significance associated with that statistic, so that if the level of significance is greater than 0.05 the variable is not statistically significant).The MNL2 Models are the results of the estimation with the data of the alternatives HST2 and CAR. The MNL2.1, MNL2.2, MNL2.3, MNL2.4, MNL2.5, MNL2.6, and MNL2.7 models are the same as the previous variants of the original and the different specifications represent a robustness check.

Taking into account the mixed data (MNL3.1) the specification of the model would be: UHST1 = OP <sup>∗</sup> (ASCHST1 + BETA1 <sup>∗</sup>PHST1 + BETA2 <sup>∗</sup>THST1) + (1-OP) <sup>∗</sup>µ ∗ (ASCHST2 + BETA1 <sup>∗</sup>PHST2 + BETA2 <sup>∗</sup>THST2) UBUS = OP <sup>∗</sup> (ASCBUS + BETA1 <sup>∗</sup> PBUS + BETA2 <sup>∗</sup>TBUS) + (1-OP)∗µ ∗ (ASCBUS + BETA1 <sup>∗</sup>PBUS + BETA2 <sup>∗</sup>TBUS) UHST2 = OP <sup>∗</sup> (ASCHST1 + BETA1 <sup>∗</sup>PHST1 BETA2 <sup>∗</sup>THST1) + (1-OP) <sup>∗</sup>µ ∗ (ASCHST2 + BETA1 <sup>∗</sup>PHST2 BETA2 <sup>∗</sup>THST2) UCAR =OP <sup>∗</sup> (ASCCAR + BETA1 <sup>∗</sup> CTCAR + BETA2 <sup>∗</sup>TCAR) + (1-OP) <sup>∗</sup>µ ∗ (ASCCAR + BETA1 <sup>∗</sup> PTCAR + BETA2 <sup>∗</sup>TCAR)

We proceeded to estimate multinomial models with three and four alternatives. The results of the estimation of the models are shown in **Tables 3, 4** (Bierlaire, 2003), for the multinomial simple models (MNL) with four and three alternatives, respectively.

Subsequently, models that combine the two databases have been estimated applying the specific problem of mixed data (Ben-Akiva and Morikawa, 1990), estimating Multinomial Logit models, with three and four alternatives. Once detected that the scale value of mixed data is significant, which means that there are differences in errors in the two databases, we start scaling the data, verifying that the results improve with the scaled data. From this moment on when estimating the Multinomial Logit models, the data of the sub-base 2 are scaled by this calculated coefficient in the model MNL3.1.

In **Table 3** it can be verified that the variable Theta (µ) can be verified to be statistically significant. For this reason, the database 1 must be scaled in order to be able to establish alternatives HST1 and HST2 as a single HST alternative. Thus, the three estimated alternatives models will have the scaled data. (**Tables 4, 6**).

In **Table 4** the estimated model is shown with three alternatives. From the MNL4.1 model with three alternatives without scaled data we can verify how in subsequent models with scaled data the results are improved in terms of Pseudo-R2 (ρ2). Therefore, from the scaled data the proposed specifications are


Value is the estimated coefficient for each of the explanatory variables included in the model. T-test: \* 90%, \*\* 95% and \*\*\* 99% level of significance. It is the statistic that allows to contrast the null hypothesis of individual non-significance of each of the variables included in the model, (the contrast is performed through the level of significance associated with that statistic, so that if the level of significance is greater than 0.05 the variable is not statistically significant).

estimated with three alternatives of Multinomial Logit and Mixed Logit models. From each of these results shown the provisions to be paid are calculated as the quotient of the estimated parameters for time and price.

Finally, the specification of a Logit Mixed model (ML), of fixed parameters, is carried out. The Logit Mixed model with error



Value is the estimated coefficient for each of the explanatory variables included in the model. T-test: \* 90%, \*\* 95%, and \*\*\* 99% level of significance. It is the statistic that allows to contrast the null hypothesis of individual non-significance of each of the variables included in the model, (the contrast is performed through the level of significance associated with that statistic, so that if the level of significance is greater than 0.05 the variable is not statistically significant).

TABLE 4 | Results of MNL models estimated with three alternatives and with scaled data.


Value is the estimated coefficient for each of the explanatory variables included in the model. T-test: \* 90%, \*\* 95% and \*\*\* 99% level of significance. It is the statistic that allows to contrast the null hypothesis of individual non-significance of each of the variables included in the model, (the contrast is performed through the level of significance associated with that statistic, so that if the level of significance is greater than 0.05 the variable is not statistically significant). Estimated model with three alternatives without scaled data that serves to verify how successive models with scaled data improve the results.


Value is the estimated coefficient for each of the explanatory variables included in the model. T-test: \* 90%, \*\* 95%, and \*\*\* 99% level of significance. It is the statistic that allows to contrast the null hypothesis of individual non-significance of each of the variables included in the model, (the contrast is performed through the level of significance associated with that statistic, so that if the level of significance is greater than 0.05 the variable is not statistically significant).

components is used when analyzing arbitrary substitutability or correlation patterns. In this case the utility of the alternative i is specified as:

$$U\_{iq} = \alpha' \mathfrak{x}\_{iq} + \mu'\_{q} z\_{iq} + \varepsilon\_{iq}$$

where:

$$\boldsymbol{x}\_{iq} \text{ and } \boldsymbol{z}\_{iq}, \text{ are vectors of observed variables.}$$

α, is a vector of fixed coefficients and represents the parameter vector of the explanatory variables of alternative i.

µ<sup>q</sup> and εiq, are independent, being µq, is a vector of random coefficients with mean 0 and covariance W y εiq, are random variables, iid Gumbel(0,β).

With this formulation the random term is:

$$
\eta\_{iq} = \mu'\_q z\_{iq} + \varepsilon\_{iq}
$$

It can be verified that with this formulation there is correlation between any pair of alternatives:

$$\begin{aligned} \text{cov}(\eta\_{iq}, \eta\_{iq}) &= E \left[ \mu' \_{q} z\_{iq} + \varepsilon\_{iq} \right] E \left[ \mu' \_{q} z\_{jq} + \varepsilon\_{jq} \right] \\ &= z'\_{iq} W z\_{iq} + \sigma\_{\varepsilon}^{2} I \neq 0 \text{ siz}\_{iq} \neq 0 \end{aligned}$$

The probability of choice can be obtained in a similar way to the previous case, conditioned by the value of µq. Due to the fact that there is no guarantee of being able to solve the integral, the probability is obtained by simulation.

Both models (of random parameters and error components) are equivalent when the vector parameter β breaks clown into its mean α plus the deviations µ<sup>q</sup> (β<sup>q</sup> = α + µq). Reciprocally, if ziq = xiq the error component model is equivalent to a random parameter model: and if ziq = xiq, we would obtain a random parameter model with fixed coefficients for xiq and random coefficients with zero mean for ziq.

The specification of a Logit Mixed (ML) model, with a panel effect to analyze the correlation between variables, using a fixed effects model, where four alternatives are specified first and further on three are specified.

First, a Logit Mixed model (ML5.1) is proposed to determine the panel effect, with mixed data and with the same error component for the two databases, so that the model will be:

$$\begin{aligned} \text{U}\_{\text{HST1}} &= \text{OP}^\* \left\{ \text{ASC}\_{\text{HST1}} + \text{BETA1} \, ^\* \text{P}\_{\text{HST1}} + \text{BETA2} \, ^\* \text{T}\_{\text{HST1}} \right\} \\ &+ (\text{1-OP}) \, ^\* \mu \text{\text{ (ASC}\_{\text{HST2}} + \text{BETA1} \, ^\* \text{P}\_{\text{HST2}} + \\ &\text{BETA2} \, ^\* \text{T}\_{\text{HST2}}) \\ \text{U}\_{\text{BUS}} &= \text{OP}^\* \left\{ \text{ASC}\_{\text{BUS}} + \text{BETA1} \, ^\* \right\} \end{aligned}$$

$$\begin{aligned} \text{P\_{BUS}} &+ \text{BETA2} \, ^\*\text{T}\_{\text{BUS}} \text{(}+ \text{ZERO[} \text{SIGMA1]+ }\text{)}\\ \text{(I-OP)} &\, ^\*\text{(ASC}\_{\text{BUS}} + \text{BETA1} \, ^\*\text{P\_{BUS}} + \text{BETA2} \, ^\*\text{)}\\ \text{T}\_{\text{BUS}} &+ \text{ZERO[} \, \text{SIGMA1]} \, \end{aligned}$$

$$\begin{aligned} \text{U}\_{\text{HST2}} &= \text{OP}^\* \left( \text{ASC}\_{\text{HST1}} + \text{BETA1} \, ^\* \text{P\_{HST1} \, BETA2} \, ^\* \text{T}\_{\text{HST1}} \right) \\ &+ (\text{1-OP}) \, ^\* \mu \text{ (ASC}\_{\text{HST2}} + \text{BETA1} \, ^\* \text{P\_{HST2}} \\ &+ \text{BETA2} \, ^\* \text{T}\_{\text{HST2}} \end{aligned}$$

UCAR = OP <sup>∗</sup> (ASCCAR + BETA1 <sup>∗</sup> CTCAR + BETA2 <sup>∗</sup>TCAR +ZERO[ SIGMA1] ) + (1-OP) <sup>∗</sup>µ ∗ (ASCCAR <sup>+</sup> BETA1 <sup>∗</sup> PTCAR + BETA2 <sup>∗</sup>TCAR + ZERO[ SIGMA1] )

Being: OP = 1 with values of sub-base 1 and 0 for values of sub-base 2 (OP-1 = Base 2)

And later it is specified with different errors for each base (ZERO\_SIGMA1 and ZERO\_SIGMA2)

UHST1 = OP <sup>∗</sup> (ASCHST1 + BETA1 <sup>∗</sup>PHST1 + BETA2 <sup>∗</sup>THST1) UBUS = OP <sup>∗</sup> (ASCBUS + BETA1 <sup>∗</sup> PBUS + BETA2 <sup>∗</sup>TBUS) + ZERO[ SIGMA1] ) UHST2 = (1-OP) <sup>∗</sup>µ ∗ (ASCHST2 + BETA1 <sup>∗</sup>PHST2 + BETA2 <sup>∗</sup>THST2) UCAR = (1-OP) <sup>∗</sup>µ ∗ (ASCCAR <sup>+</sup> BETA1 <sup>∗</sup> PTCAR

+ BETA2 <sup>∗</sup>TCAR + ZERO[ SIGMA2] )

The specifications of the three alternatives are performed, first without scaling the data, ML 6.2:

$$\begin{aligned} \text{U}\_{\text{HST}} &= \text{ASC}\_{\text{HST}} + \text{BETA1} \, ^\ast \text{P}\_{\text{HST}} + \text{BETA2} \, ^\ast \text{T}\_{\text{HST}}\\ \text{U}\_{\text{BUS}} &= \text{ASC}\_{\text{BUS}} + \text{BETA1} \, ^\ast \text{P}\_{\text{BUS}} + \text{BETA2} \, ^\ast \text{T}\_{\text{BUS}}\\ &+ \text{ZERO} [\text{SIGMA}]\\ \text{U}\_{\text{CAR}} &= \text{ASC}\_{\text{CAR}} + \text{BETA1} \, ^\ast \text{PT}\_{\text{CAR}} + \text{BETA2} \, ^\ast \text{T}\_{\text{CAR}}\\ &+ \text{ZERO} [\text{SIGMA}] \end{aligned}$$

And with the Base 2 data scaled, with the scale factor obtained in the MNL3.1 estimate (µ = 0.0830), the model is specified with three alternatives such that ML.6.4:

$$\begin{aligned} \text{U\_{HST}} &= \text{ASC\_{HSTE}} + \text{BETA1} \, ^\ast \text{P\_{HSTE}} + \text{BETA2} \, ^\ast \text{T\_{HSTE}}\\ \text{U\_{BUS}} &= \text{ASC\_{BUS}} + \text{BETA1} \, ^\ast \text{P\_{BUS}} + \text{BETA2} \, ^\ast \text{T\_{BUS}}\\ &+ \text{ZERO} [\text{SIGMA}]\\ \text{U\_{CAR}} &= \text{ASC\_{CARE}} + \text{BETA1} \, ^\ast \text{PT\_{CARE}} + \text{BETA2} \, ^\ast \text{T\_{CARE}}\\ &+ \text{ZERO} [\text{SIGMA}] \end{aligned}$$

**Tables 5, 6** show the results for the Logit Mixto models, with three and four alternatives as well (Bierlaire, 2003).

The results of the Multinomial Logit models with three choice alternatives (HST-CAR and BUS) improve in terms of the coefficient of determination, showing more realistic payment arrangements. Even in the case of when we include the comfort variable (measured according to the number of transfers), together with the total price of the trip (which separately includes the cost of the trip and the cost of parking for the car, is significant. As well as that obtained from models whose variables are not significant and with signs of time not expected, the results improve in a substantial way.

### CONCLUSIONS AND RECOMMENDATIONS

This article shows an application of the demand models or discreet choice of Logit type to obtain the provisions for the payment of the consumers of a transport service, specifically in the corridor Toledo-Madrid/Madrid Airport. In this case it is analyzed what will be the choice of the travelers that make this journey, considering that they have three alternatives at present, the high speed train, car and the bus.

The theoretical basis of these emodels starts from the theory of random utility specifically in the theory of the maximization of utility. This model is based on the existence of a rational consumer that represents the average behavior of the set of consumers, based on a series of parameters that measure that utility. In the case of transport, these variables are usually the travel time, the cost or price of the service and the convenience of the trip, as well as different socioeconomic variables of the individual.

The model data are obtained through a survey of declared preferences. In it an exercise of simulation of alternative scenarios is proposed that configure mobility in the future. This mobility is based on the new infrastructure of connection with the airport. So there are different service options on the route which are two transfers, one transfer or no transfer.

In general, of the results presented above, we highlight the variables of the equation and the measures of goodness of the model.

Logit models have been estimated that measure the probability of choosing the HST in front of the bus or the car independently and according to the socioeconomic characteristics of the subject and the attributes of the different modes of transport. The variables included in each of the models are: time, price, sex and income. Thus, the attributes for each of the transport modes are (1) HST, price and travel time, (2) bus, price and travel time and (3) car taking into account price, parking cost and travel time.

Specifically, the results show that the willingness to pay HST-Car binomial models is too large (600 € per hour saved on the trip), motivated by individuals who have responded in a lexicographic way and motivated by the smaller number of people who have chosen option bus in front of the HST option. The provisions for the payment of the binomial HST-Bus, are more reasonable and oscillate between 23 and 25 € per hour.

In multinomial models the provisions for payment are more realistic, except when we include the comfort variable, together with the total price of the trip, with which models are obtained whose variables are not significant and even with the signs of the time not expected, estimating The model without independent terms with which you get a result of 26.3 € per hour.

From the results of the modeling we can analyze how the consumer is willing to pay 26.3 € for each hour of travel saved which gives us information about the importance of travel time over the price paid.

In conclusion, based on this information, a sustainable transport policy focusing on HST-airplane intermodal transport is recommended in order to facilitate the reduction of travel times, by encouraging the use of efficient public transport as the high-speed train to reach the airport.

A transport policy focused on replacing private transport with public transport (HST), benefits, on the one hand, users who will have shorter travel times and in general, the transport system to reduce the congestion of access to Madrid and to the Madrid Airport (Li and Hensher, 2012).

The main variables to encourage the integration of air and high-speed modes of transport are the visibility of the offer, integrated management of the reservation and boarding passes, check-in and baggage control, full travel responsibility, the management of passenger loyalty programs, ease of connection in intermodal mode, travel time and price (Muro, 2012). As for this offer, and above all regarding the connections between Toledo and Madrid airport, a new line of direct buses to the airport has emerged (Muro and Pérez, 2016) and recently, in February 2017, the offer of combined train and flying tickets (Train and Fly)<sup>6</sup> from the city of Toledo, which indicates improvements in air-rail intermodality (air-rail intermodality).

### AUTHOR CONTRIBUTIONS

AM: modeling and coordination. IP: methodological framework. SG: introduction and conceptual framework.

#### FUNDING

This paper has been funded by Aid Research Groups at the University of Castilla-La Mancha 2017. Research Group

<sup>6</sup>http://www.renfe.com/viajeros/trenmasavion/index\_ES.html

### REFERENCES


"Observatorio de la Innovación en Distribución Comercial". This paper has been funded by The Ministry of Economy and Competitivity (Spain), Research Project with reference: ECO2014-59688-R, Programa Estatal de Investigación, Desarrollo e Innovación Orientada a los Retos de la Sociedad, Plan Estatal de Investigación Científica y Técnica y de Innovación 2013–2016. And the revision of the translation has been financed with the collaboration of the contract-program of the Faculty of Juridical and Social Sciences of Toledo.

#### ACKNOWLEDGMENTS

We are grateful to our teacher, Professor Timoteo Martínez Aguado, for the collaboration and help he has always given us and, in particular, for his contributions in this work.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Muro-Rodríguez, Perez-Jiménez and Gutiérrez-Broncano. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# New Strategies in the New Millennium: Servant Leadership As Enhancer of Service Climate and Customer Service Performance

Jorge Linuesa-Langreo, Pablo Ruiz-Palomino\* and Dioni Elche-Hortelano

Department of Business Administration Department, University of Castilla-La Mancha, Cuenca, Spain

In a world in which customers are increasingly looking for solutions to their own concerns on how to make a better globalized world, new organizational strategies are emerging to approach the customer in the current third millennium. Servant leadership, which involves putting employees' needs first and serving the broader society, is emerging as a new strategic mechanism to approach the customer in line with the new social valuesdriven Marketing 3.0 era. Yet research has ignored the role and the various mechanisms servant leadership might utilize to improve customer service performance of their service units. Spanning 185 hotels located in Spain, a sample of 247 service units –in close contact with customers– was used to investigate whether servant leadership enhances customer service performance through shaping a service climate within the service unit. Results revealed that service climate mediates the positive influence of servant leadership on customer service performance. Managers can use these findings to note the value of leading the service unit in a servant friendly direction, which is better aligned with the new aspirations of customers today.

Keywords: social-aware customers, servant leadership, service climate, customer service performance, Marketing 3.0

## INTRODUCTION

Recent marketing literature has begun to nurture from a new, very incipient perspective, the social values-driven Marketing 3.0 paradigm (Kotler et al., 2010), which proffers the idea that customers are increasingly seeking solutions to their own concerns and are interested in building a better world. Such understanding involves purchase decisions on the basis of fulfilling social and ethical values (e.g., social justice, human welfare, environmental sustainability; Shaw et al., 2005; Hollenbeck and Zinkhan, 2010). In other words, in this new millennium, driven by the Marketing 3.0. paradigm, which entails a more human-centric perspective, customers look to products and services to meet their own needs in parallel with fulfilling spiritual, social and moral values (Kotler et al., 2010). The extent to which a product or service provides freedom of choice, independence as well as benevolence, social justice, equality and environmental responsibility is becoming more and more crucial for customers when making purchase choices (Martínez-Cañas et al., 2016), especially in a developed world, where consumption appears to have become an end in itself, through which customers find a voice to promote a better society (Vrontis and Thrassou, 2007). Furthermore, customers are increasingly showing concerns about the effects of their purchase

#### Edited by:

Ana I. Jiménez-Zarco, Open University of Catalonia, Spain

#### Reviewed by:

Isabel Llodrà-Riera, Illes Balears Innovacio Tecnologia, Spain Mauro Capestro, University of Salento, Italy

> \*Correspondence: Pablo Ruiz-Palomino pablo.ruiz@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 24 March 2017 Accepted: 28 April 2017 Published: 16 May 2017

#### Citation:

Linuesa-Langreo J, Ruiz-Palomino P and Elche-Hortelano D (2017) New Strategies in the New Millennium: Servant Leadership As Enhancer of Service Climate and Customer Service Performance. Front. Psychol. 8:786. doi: 10.3389/fpsyg.2017.00786

choices not only for themselves, but also for broader society (Harrison, 2005; González and Fernández, 2016), which represents a strong embracement of transcendent motives in their actions. In other words, in addition to making purchase decisions with an eye on external benefits gained (i.e., extrinsic motivation) or the pleasure acquired from the purchase decision itself (i.e., intrinsic motivation), customers are more and more concerned about whether their purchase decisions contribute to solving the problems of someone else (i.e., transcendent motives). In effect, in the current millennium, customers are more and more worried about whether others, both known or unknown, meet human good, such as truth, beauty, work, friendship, life, and dignity, such that their purchase decisions' impact on others is carefully calculated (Martínez-Cañas et al., 2016).

In accordance with the described scenario, in which customers are demanding that businesses act in a socially responsible manner (Vrontis and Thrassou, 2007), it is of no surprise that managers are beginning to think of new customer-focused strategies to engage with the modern customer today (González and Fernández, 2016). One of these interesting strategies to appeal to the customer in the current social values-driven Marketing 3.0 era is the development of servant leadership in service units –work units which are in close contact with customers–. While leadership is deemed a central aspect to orientate service units' mission and values toward a greater involvement and co-responsibility with broader society, servant leadership is the unique leadership approach which, as captured by its name, focuses on serving others (Liden et al., 2014), including the least privileged in society (Greenleaf, 1977). Servant leadership's concerns extend beyond the organization itself to meet the well-being of followers, customers, other stakeholders, and society in a wider sense (Greenleaf, 1977; Barbuto and Wheeler, 2006). Their profound responsibility to serve others and contribute to the larger society is central in undertaking such leadership approach (Liden et al., 2014). In addition it purports to show followers how to fulfill the business mission of serving broader society, (Graham, 1991). Indeed, according to Greenleaf (1977), the first to coin the term after reading the Journey to the East by Herman Hesse, one of the central tenets of the servant leadership approach is that serving others entails encouraging others to do the same, so that they become servant leaders.

It is of no surprise then that this leadership approach results in an appropriate strategy to improve customer service (Brownell, 2010; Wu et al., 2013). When servant leadership is present in service units, employees of such work units are more likely to provide genuine care to customers and, in turn, authentic, high-quality customer service (Brownell, 2010). This process is possible because it is highly characteristic of servant leaders to fuel a cycle of service within their service units; servant leaders are role-models of servant behavior which, in turn, is mirrored by followers (Hunter et al., 2013). Therefore, under the influence of servant leaders a service climate, i.e., workers' perception that internal practices, procedures, and behaviors support the provision of quality service– is likely to emerge. Furthermore, by shaping service climates, servant leaders should contribute to the enhancement of customer service performance, i.e., the workers' proficiency in undertaking the core parts of their service role to provide high-quality customer service. This finds support in service-linkage research (Wiley, 1996; Pugh et al., 2002; Schneider et al., 2005), which is concerned with finding the links between employees and customers in service firms -where the boundaries between both agents are fairly permeable-. According to this perspective there must be specific drivers which link employees' perceptions of various inter-organizational practices to customer perceptions, and some studies have argued that service climate is the bridge, the missing link between what happens inside –procedures, practices–, and what happens outside –customers' perceptions– (Schneider et al., 2002; Hong et al., 2013). In fact, prior research has linked service climate to perceptions of service quality (Gracia et al., 2010), and service performance (Liao and Chuang, 2007).

It seems then that servant leaders in the new millennium, which depicts a scenario where customers are more concerned about how to contribute to build a better world, might play a role in improving customer service performance; however, such influence might occur by fostering and shaping service climates within their service units. To the best of our knowledge, this particular point has not yet been addressed in existing research. Therefore, this study pursues two main objectives. First, we will investigate whether servant leadership might be a suitable leadership strategy to improve customer service performance, as a measure which captures service units' emphasis on service quality. Second, and more importantly, we examine the mediating effect of service climate between servant leadership and customer service performance. These relationships will be analyzed at the work unit level via spanning service units. The work unit, rather than the individual, is the building block of organizations, today (West and Markiewicz, 2004) and permits managers to work closely with followers on a daily basis. Service units in service, tourism firms, i.e., hotels, represent suitable units of analysis for uncovering the influence of servant leadership on service climate and service customer performance.

### THEORY AND HYPOTHESES DEVELOPMENT

### The Nature of Servant Leadership

With its strong, unique focus on serving others (Wu et al., 2013; Liden et al., 2014), servant leadership offers unique aspects which can enhance the quality of service provided to customers. Greenleaf (1977, p. 27) coined the concept of servant leadership, though he failed to give a formal definition, and described the phenomenon of servant leadership as:

The servant-leader is servant first. It begins with the natural feeling that one wants to serve, to serve first. Then conscious choice brings one to aspire to lead. The difference manifests itself in the care taken by the servant-first to make sure that other people's highest priority needs are being served. The best test, and difficult to administer, is this: Do those served grow as

persons? Do they, while being served, become healthier, wiser, freer, more autonomous, more likely themselves to become servants? And, what is the effect on the least privileged in society? Will they benefit or at least not be further deprived?

Such description clarifies two core aspects of the servant leadership strategy. First, servant leadership extend their service approach to the various stakeholders, including employees, customers, and society in general (Graham, 1991; Wu et al., 2013); these leaders even raise strong concerns focused on improving the well-being of the least privileged in society. As such, because the servant leader's area of concern extends beyond the business organization and includes the broader social environment (Brownell, 2010), these leaders demonstrate a high level of social responsibility for the well-being of society in general (Reinke, 2004). Second, servant leadership prioritizes the fulfillment of others' needs above their own personal needs (Greenleaf, 1977), and inspire followers to develop intelligently, be creative and self-manage in order to serve others (Liden et al., 2008). As a result, because servant leaders practice this "service" mindset in all aspects of their lives (Liden et al., 2008), the principle of serving others above serving oneself is unlikely not to radiate out toward followers' mindsets, attitudes, and behaviors. In effect, servant leaders encourage their followers to develop servant behaviors that will benefit all stakeholders (Liden et al., 2008; Sendjaya et al., 2008), thus ensuring that the strategies and decisions they opt for will offer a positive legacy to society (Barbuto and Wheeler, 2006). Because servant leaders are constantly searching out benefits to society (Searle and Barbuto, 2011), they are likely to inspire related servant attitudes in their followers to do their best to the benefit of all stakeholders, including customers. Thus, by developing a deep level of identification with the behavior carried out by their servant leaders (Zhang et al., 2012), employees should take personal responsibility for providing assistance and worth to customers, and thus provide services which are authentic and of superior quality (Brownell, 2010).

### Servant Leaders and Customer Service Performance

One of the key characteristics that sets service companies apart from those which produce goods is the simultaneous nature of service production and consumption, which, in many cases, results in consumer participation in the co-creation of the service (Bowen and Schneider, 1988). As such, the experience provided to the consumer upon receiving services is as important as, or even more important than, the product offered to the consumer (Bowen and Waldman, 1999). This has led to a paradigm shift for service companies when defining (customer service) performance, moving away from a focus on behavior evaluation to achieve organizational objectives (Campbell et al., 1993), toward a greater focus on behavior developed by the worker himself/herself, which is geared toward serving and helping the customer with the goal of providing high-quality service (Liao and Chuang, 2004).

This change of perspective means that the worker maintains direct and ongoing contact with the customer, thus increasing uncertainty as to how to interact with the customer, as in many cases the customer demands an immediate solution (Skaggs and Galli-Debicella, 2012). In these circumstances, the development of leaders who offer their workers both guidance and common sense, and who consistently cover all of his workers' needs, is vital for improving workers' performance (McGrath, 1962; Morgeson et al., 2010). As such, given that servant leadership develops social responsibility when serving both workers and customers (Mahembe and Engelbrecht, 2014), servant leadership in managerial roles appears to be a key element to generate a higher level of customer service performance (Chen et al., 2015). In effect, servant leaders possess impressive conceptual abilities for offering workers direction, support and clarity in solving day-to-day problems (van Dierendonck, 2011). These aspects help workers cultivate a precise understanding of their changing environment and develop individual and group skills (Hu and Liden, 2011). Some of these skills are, for example, more creative performance (Neubert et al., 2008), independence and self-confidence (Liden et al., 2008), which facilitates behavior which is spontaneous and useful in meeting customer demands without needing any type of supervision (Chen et al., 2015).

The positive relationship between servant leadership and customer service performance can be explained using the social exchange theory (SET, Blau, 1964; Cropanzano and Mitchell, 2005). SET theory indicates that social relationships are based on norms of reciprocity (Gouldner, 1960), where people look to maintain psychological balance in their social interactions, returning "favors" to those who have demonstrated proactive and positive tendencies toward them. Accordingly, when servant leaders in service units put their workers first, and display a service attitude and sincere concern for covering workers' needs of personal and professional growth (Greenleaf, 1977), they create a psychological imbalance in workers' relationships with these leaders. In a bid to benefit these leaders, workers might engage in service behaviors, directed to benefit the service unit, by superbly attending, for example, to customers' needs. These behaviors, might also be impregnated with servant leaders' strong emphasis in meeting the well-being of others, including customers and broader society, as these workers should become servants to an even greater extent according to Greenleaf (1977). As such, the display of such behaviors in encounters with customers should result in high-quality services, particularly in the present environment, in which customers are more and more needful of signals showing that the products and services they consume contribute to building a better society. Various studies support this relationship. While Netemeyer et al. (2005) recognize the better the leader–worker relationship, the stronger the impact is on the relationships maintained between workers and customers, other studies confirm a positive relationship between servant leaders and workers, which, in turn, translates into better worker– customer relationships and higher quality of service provided to customers (Jaramillo et al., 2009; Chen et al., 2015; Ling et al., 2016). Formally:

H1: Servant leadership is directly, positively related to customer service performance.

### Servant Leaders and Service Climate

Leadership is one of the most important factors in the process of climatic formation (Kozlowski and Doherty, 1989). Supervisors who model their leadership approach to all of their workers represent significant influences in forming the climate and providing it with content (Mayer et al., 2007). As such, servant leaders, by maintaining a service attitude oriented to meet both workers and customers' needs (Wu et al., 2013), are active agents in forming a service climate within the work units to which these leaders belong (Walumbwa et al., 2010). In effect, service climate, defined as "employee perceptions of the practices, procedures, and behaviors that get rewarded, supported, and expected with regard to customer service and customer service quality" (Schneider et al., 1998, p. 151), is likely to emerge in service units which are led by servant leaders, as explained by social learning theory (SLT, Bandura, 1977, 1986).

Social learning theory contends that individuals learn the appropriate behavior by observing and emulating values, attitudes and behaviors of attractive, credible role models (Bandura, 1977, 1986). Managers become these influential referents (Mayer et al., 2012) because they embody proximity, frequent social interaction, and formal authority (Merton, 1957; Sutherland and Cressey, 1970), which makes it easier for them to garner attention and convey attractive information. As such, workers, through observing supervisors' behaviors in a day-today work setting, are likely to engage in imitative behaviors (Hall and Lord, 1995), which can be further intensified whenever workers perceive their leaders to be in possession of qualities they consider to be attractive (Neubert et al., 2008; Mayer et al., 2012).

Servant leaders capture such attractiveness as they offer guidance and direction to workers and, by being humble, loving, empathetic, and servant (Sun, 2013), manifests sincere concern for satisfying the needs of both workers and customers (Wu et al., 2013). Imitative behavioral trends are highly probable in workers who are led by servant leaders, who will manifest behaviors that are similar to those of their leaders (Hunter et al., 2013), behaviors which are oriented to serve broader society by developing people committed, in turn, to serve society (Sims, 2005). This service-oriented behavior, developed by workers, results in a phenomenon of contagion (Bono and Ilies, 2006) among all members of the service unit which inspires a continuing service cycle (Hunter et al., 2013), a service-oriented culture (Liden et al., 2014) and, overall, a greater service climate (Walumbwa et al., 2010). The result is the creation of a work environment where members of the service unit share social behavioral norms, aimed first and foremost at offering high-quality service (Schneider et al., 2002). Accordingly, we propose:

H2: Servant leadership is directly, positively related to service climate.

### Service Climate and Customer Service Performance

Organizational theorists emphasize the importance of the organizational climate in determining employees' attitudes and behaviors (Schneider et al., 2013). In accordance with the social information processing theory (SIP, Salancik and Pfeffer, 1978), employees collect the various messages released by their work units and utilize this information for decision making issues. Employees usually tend to adapt their feelings, attitudes, and behaviors according to what they perceive in their immediate work environment (Van Dyne et al., 1995). Therefore, the organizational climate, understood as the shared perceptions regarding policies, procedures, and practices (Schneider, 1990) which signal how things ought to be done and what behaviors are proper in the work environment (Liao and Chuang, 2007), represents an important influence on employees' performance (Schneider and Barbera, 2014). In fact, depending on the specific dimension the organizational climate emphasizes, a number of studies have revealed its significant influence on specific behavioral outcomes regarding areas such as safety (i.e., Zohar and Luria, 2004), ethics (i.e., Deshpande, 1996), innovation (Anderson and West, 1998) or service (i.e., Schneider et al., 2009). This is because such specific climates represent the best source of cues to interpret events, undertake proper attitudes, and understand behavioral expectations concerning the different areas or dimensions emphasized.

In the particular case of service climate, it helps employees internalize that excellent service is expected, desired, and rewarded; it also represents a strong motivational force to deliver the best service in day-to-day activities (Liao and Chuang, 2007; Liao et al., 2009). Such perception is important in the service context, where services are produced and delivered in real time by unit employees (Ehrhart et al., 2011); in such contexts, the more customers perceive service quality is central for employees, the better their service experience (Schneider and White, 2004). This is not surprising as this specific climate emphasizes service quality to a great extent, so as a result, it should have a direct impact on service outcomes (Schneider, 1990) such as customer service performance. Indeed, employees highly engaged and sharing common perceptions about providing good quality of service to customers should perform well with customers (Salanova et al., 2005). This should occur because of social learning mechanisms (Bandura, 1977, 1986) as well as because perceptions that a high value for service is the tone, which should provide meaning to work and make employees enjoy their jobs to a greater extent (Hong et al., 2013). Earlier empirical research has consistently revealed that service climate enhances customer service performance (Borucki and Burke, 1999; Liao and Chuang, 2004, 2007; Salanova et al., 2005; Ehrhart et al., 2011; Hong et al., 2013). Thus, we predict:

H3: Service climate is directly, positively related to customer service performance.

### Servant Leaders-Customer Service Performance: The Mediation of Service Climate

According to linkage research (i.e., Wiley, 1996; Schneider et al., 1998; Wiley and Brooks, 2000; Pugh et al., 2002), there are internal elements of the work environment which can be strongly linked to critical external performance outcomes. Service climate is one of these internal elements which recent research

has identified as the bridge between the work environment as perceived by employees- and critical, external performance success factors oriented to the customer (e.g., customer service performance, Hong et al., 2013). This is important because it permits managers to apply an indirect approach, by focusing on more easily manageable internal aspects (e.g., leadership), to encourage customer service-minded behavior (Solnet, 2007), which can be conducive of better customer service perfomance.

Looking into internal elements, servant leadership implies an environmental stimulus (Hu and Liden, 2011) which is built upon service values based on genuine concern for and loving care of others (van Dierendonck, 2011; Hunter et al., 2013). Such stimulus should germinate, grow and propagate within the collective, because drawing on SLT (Bandura, 1977, 1986) workers feel attracted to imitate such attractive leaders by engaging in similar servant behaviors (Liden et al., 2014). Hence, servant leadership within service units should be associated with the shared perception that interpersonal relationships rest upon such service values, which should help shape a climate fostering helpful behavior oriented to offer high quality service (i.e., Walumbwa et al., 2010). Servant leaders foster climates which send clear messages that egotistical behavior is not tolerated (Liden et al., 2014), and service spirit is strongly encouraged (Liden et al., 2008).

Looking at external performance outcomes, various studies have revealed clear positive effects of service climate on customers (Bowen and Schneider, 2014), such as, for example, customer service (e.g., Schneider and Bowen, 1985), customer satisfaction (e.g., Schneider et al., 1996), and customer loyalty (e.g., Salanova et al., 2005). This is because in scenarios where service climate is perceived, workers share the understanding that the behavioral norms and expectations are to prioritize the needs of others (Liden et al., 2014), specifically, customers, which encourages employees' strong engagement in high-quality service behavior directed to the customer (Liao and Chuang, 2007). Such a service climate is ignited through a spillover process spreading service attitudes and behaviors which should be noted in employee–customer interactions. With customers perceiving employees to be warm, in a good mood, and willing to dedicate time to understand their needs, customers get a good feeling about the service received. Such good feelings would also be nurtured as long as these influenced servant workers show strong concern for building a better society –which is an increasing concern of customers, today (Vrontis and Thrassou, 2007)-. In other words, by observing workers who behave in this way, customers should enjoy such an awesome experience that they should feel that a high quality service has been received (Hong et al., 2013).

Overall, by combining both internal and external perspectives, we contend that because servant leaders enhance service climate within their service units (Walumbwa et al., 2010), these leaders improve the quality of service that workers offer to customers (Ling et al., 2016). In other words, by drawing on the broad existing body of linkage research (Wiley, 1996; Schneider et al., 1998; Wiley and Brooks, 2000; Pugh et al., 2002), we contend that service climate is the bridge between servant leadership and customer service performance. Thus, we propose:

#### H4: Service climate mediates the relationship between servant leadership and customer service performance.

This last hypothesis combined with the previous ones make it possible to summarize our research model as displayed in **Figure 1**.

### METHOD

#### Sample and Procedure

In order to test these relationships, we conducted surveys to gather data in Spain's hospitality industry, which is likely to attract managers who are servant leaders. Customer service of the utmost quality is key for success of companies in this sector, so servant leadership strategies could make an important difference and be common in these circles (Wu et al., 2013).

In an effort to minimize common method bias (CMB), and social desirability bias (SDB), we selected the most fitting participants for all our study constructs (Podsakoff et al., 2003). Firstly, workers were selected to measure the extent to which their supervisors can be portrayed as using a servant leadership strategy. Secondly, service climate was measured by using both workers' ratings and supervisors' ratings, as it allows us to minimize the same-source bias problem (Ostroff et al., 2002) and thus have a more objective indicator of the phenomena. Finally, customer service performance was assessed by the hotel

general manager, not by service units' supervisors who might give biases responses. Accordingly, we designed three different questionnaires for each target respondent (i.e., hotel general manager, service unit supervisor, service unit worker). We pilot tested each questionnaire with a convenience sample of 3 general managers, 10 supervisors, and 25 workers in 3 hotels, respectively, which confirmed the clarity, comprehension, readability, and suitability of the items included. Surveys' cover letters for each target respondent indicated absolute anonymity, and noted that only aggregated data would be utilized for research purposes.

Once consent was gained from the general managers at 185 hotels, each located in a distinct Spanish historical site, data from 247 service units (three members per unit at a minimum) which were in close contact with customers (e.g., reception desk, restaurant) could be gathered. Both supervisor and workers' responses were collected per each service unit; in total 840 responses were received –the response rate was high, around 77%-. As to collection of data concerning each unit's customer service performance, hotel general managers were also surveyed; in total 185 responses were obtained. Data were collected at specific sites in each establishment, which helped us ensure that each survey was paired with its corresponding service unit and hotel.

As further countermeasures for the CMB and SDB issues, the design of the survey was based on recommendations raised by Podsakoff et al. (2003) and Conway and Lance (2010). For example, the survey's cover letter highlighted the fact that there were no correct or incorrect answers, thanked participants in advance for being honest, and noted that all responses would remain anonymous. Participants did not have to share their names, job titles or employers' names in the survey. Furthermore, the cover letter clearly stated that the results were for academic purposes only, thus reducing SDB (Nancarrow et al., 2001). Lastly, when designing the survey (Podsakoff et al., 2003), we worked to ensure a psychological separation between predictors and outcome variables, to keep them from seeming related. We also used distractor elements and utilized items that were simple, focused and concise.

#### Measures

The survey was conducted in Spanish. Brislin's (1980) backtranslation procedure was conducted to our mediator and independent variable, and no meaningful differences between the two translations from and to English were noticed. An exhaustive analysis, according to MacKenzie et al.'s (2005) criteria, showed that all our measures contained highly correlated indicators; in other words, our survey included reflective measures in all cases.

#### Servant Leadership

Service unit workers used Ehrhart's (2004) reliable 14-item scale to rate servant leadership of their service unit supervisors. The scale used a seven-point response format (1 = "strongly disagree," 7 = "strongly agree"). Sample items were, "My supervisor spends the time to form quality relationships with service unit employees" and "My supervisor emphasizes the importance of giving back to the community." Because we were interested in overall patterns of servant leadership behavior within the service unit, we averaged employees' ratings within each service unit. To confirm that this aggregation of individual scores to the unit level was appropriate, we calculated the within-service unit agreement score (rwg, James et al., 1984) and two intraclass correlations: ICC(1), or the proportion of variance in ratings due to service unit membership, and ICC(2), or the reliability of service unit mean differences (Bliese, 2000). The average rwg value was 0.83, and the ICC(1) and ICC(2), were, respectively, 0.65 and 0.86, which met acceptable cutoffs (Bliese, 1998). In addition, we performed a one-way analysis of variance (ANOVA), which indicated significant differences across the service units in the average scores (F = 11.07, p < 0.01). Therefore, we consider the aggregation justified (Bliese, 1998).

#### Service Climate

All service unit members, including both workers and supervisors, completed on a seven-point response format (1 = "strongly disagree," 7 = "strongly agree"), Carrasco et al.'s (2012) 4-item service climate scale, which is an adaption to the Spanish context of the Schneider et al.'s (1998) global service climate scale. Sample items were, "Employees in our service unit have knowledge of the job and the skills to deliver superior quality work and service," and "The overall quality of service provided by our service unit to customers is excellent." The ANOVA indicated significant differences (F = 5.48, p < 0.01), the median rwg value was 0.83, and the ICC(1) and ICC(2) score were 0.27 and 0.56, respectively; so within-unit agreement and between-unit differentiation supported the aggregation of respondents' scores to the service unit level.

#### Customer Service Performance

Hotel general managers were asked to rate the performance of service units surveyed compared to the average of other work units in the hotel. We decided to ask hotel general managers instead of service units' supervisors because general managers should offer more accurate, far less biased responses. Specifically, hotel managers had to respond on a 7-point response format ("very poor," 1, to "excellent," 5) to four items adapted from Oh et al. (2004) to measure the extent to which the service unit in question provided high quality service to the customer. Sample items include "service unit's quality of work" and "service unit's overall performance."

#### Control Variables

In the present study, we introduced two control variables. We included service unit education to control for confounding effects because the educational level has been recognized in the past to influence helping behavior (Van Dyne and LePine, 1998) and successful service performance (Davidoff, 1994). Service unit education was measured by the average education level of participants in each service unit surveyed; individual responses on a six-point format response (1 = primary studies; 2 = secondary studies; 3 = lower level professional education; 4 = higher level professional education; 5 = bachelor degree; 6 = postgraduate degree) were averaged within each service unit. Also, we controlled for service unit size effects because it can affect group dynamics (e.g., interpersonal contacts, synergies)

and performance, positively (Brewer and Kramer, 1986; Smith et al., 1994). We measured service unit size by the number of workers who participated in each service unit we surveyed.

### Data Analysis

We utilized partial least squares via Smart PLS 3.2.6 (Ringle et al., 2015) to test our hypotheses. Such an impressive and potent statistical procedure (Chin et al., 2003) makes the causal analyses of complex situations possible (Henseler et al., 2009). As a structural equation modeling (SEM) approach, it is also suitable for testing mediation hypotheses (James et al., 2006). Also, PLS does not make it necessary to demand assumptions concerning the distribution of the variables (Henseler et al., 2009). As recommended, our PLS analysis used 5,000 subsamples to generate standard errors and bootstrap t-statistics with n−1 degrees of freedom (where n is the number of subsamples) to evaluate statistical significance of path coefficients (Henseler et al., 2009).

We tested the hypotheses at the unit level of analysis, using a sample of 247 service units (n = 247). To run our SEM model at the unit level, our servant leadership variable was formed by ratings provided by service unit workers which were averaged to yield service unit-level scores. Likewise, our service climate variable was constructed using ratings provided by the supervisor and workers of each service unit, that were averaged together to obtain unit-level scores. Finally, our criterion variable, customer service performance, was created based on ratings provided by the general manager of the hotel to which each service unit belonged to.

#### Measurement Model

By following the recommendations of Conway and Lance (2010), we present information related to good reliability and validity for our reflective measures as an additional test to show that CMB is not an issue in our study. **Table 1** shows evidence of individual and construct reliability, and convergent validity, while **Tables 2**, **3** offer good discriminant validity for all our measures. In addition, **Table 2** shows the correlations across the variables.

As seen in **Table 1**, all the individual items of the reflective variables, whose standardized loadings are far above the threshold of 0.70, are reliable (Henseler et al., 2009). In addition, the Cronbach's alpha values and composite reliability values both point to good reliability and internal consistency for all our reflective constructs, with values that are above the desired threshold of 0.80, as required for basic research (Nunnally, 1978, Table 1). The convergent validity condition was also met, because the average variance extracted (AVE) values related to each reflective construct were far above 0.50 (Henseler et al., 2009, Table 1). Lastly, we looked at the divergent validity of our reflective measures using a variety of methods. On the construct level, the criterion of Fornell and Larcker (1981) was achieved to our satisfaction, given that the AVE for each construct was greater than the variance shared by each construct with the other latent variables (**Table 2**) (Henseler et al., 2009). In addition, the heterotrait–monotrait (HTMT) criterion backed this point up as the HTMT values among our study variables were all far lower than even the most conservative 0.85 cut-off (**Table 2**), hence verifying discriminant validity for each pair of constructs (Henseler et al., 2015). This issue was also upheld when the HTMT inference criterion was utilized, which tests the null hypothesis (H0: HTMT ≥ 1) against the alternative hypothesis (H1: HTMT < 1), concluding that HTMT values among our study variables are markedly different from 1, given that confidence intervals did not include this value (Henseler et al., 2015, Table 2). On an item level, we could also affirm that our reflective constructs were different, since the cross-loading matrix showed that all items loaded on their intended constructs more than on any other construct (**Table 3**) (Henseler et al., 2015). Overall, the discriminant validity of our study variables can be considered acceptable.

### Hypotheses Testing

The variance associated with our control variables was practically non-existent as **Table 4** and **Figure 2** reveals. Only service unit level of education was significantly, positively related to service unit's customer service performance (β = 0.10, p < 0.10), thus suggesting the importance of considering this aspect when configuring the workforce of service units. **Table 4** and **Figure 2** contain findings concerning our hypotheses, as well. Contrary to our expectations concerning H1, servant leadership was not directly related to customer service performance (β = −0.03, not significant, **Table 4** and **Figure 2**), so our H1 could not be supported. However, we found support for H2 and H3, because servant leadership related directly, positively to service climate (β = 0.65, p < 0.001, **Table 4** and **Figure 2**) and service climate was directly, positively related to customer service performance (β = 0.33, p < 0.001, **Table 4** and **Figure 2**). Thus, while servant leadership was not found to influence customer service performance directly, our findings reveal that servant leadership is an important antecedent of service climate, which, in turn, impacts customer service performance, in clear support of H2 and H3, respectively.

To test H4, regarding the indirect effects of servant leadership on customer service performance, we adopted Preacher and Hayes's (2004) approach. In a bootstrap test with 5,000 subsamples (Hayes, 2009; Preacher and Hayes, 2004), the indirect effect was significant (b = 0.20, p < 0.01), and zero was absent from the 99% bias-corrected and accelerated bootstrap confidence intervals (CI lower level = 0.11; CI upper level = 0.31). The evidence of significance of this indirect effect suggests that mediation exists (Preacher and Hayes, 2004) and provides the empirical basis to analyze the mediation effect (MacKinnon et al., 2002). For the mediation test, we used Tippins and Sohi's (2003) four-criterion procedure –which includes Baron and Kenny's (1986) criteria– but applies to SEM better because it compares an unmediated model with a mediated model to find significant differences (**Figure 2**) and test if these four statistical conditions are met. The first criterion was met because the mediated model accounted for more variance in consumer service performance than the unmediated model (**Table 5** and **Figure 2**). Also, in line with H2, servant leadership related positively, directly to service climate, which offered support for the second requirement for

#### TABLE 1 | Item loadings, construct reliability and convergent validity.


All t-values of the individual loadings are significant at p < 0.001 or better. SL = servant leadership; SC = service climate; CSP = customer service performance; λ = item loading; α = cronbach's alpha; ρ = composite reliability; VIF = variance inflation factor. AVE = average variance extracted.


SUS 8.53 4.92 0.07 0.14 0.07 0.03 n.a.

Bold values on the diagonal are square roots of AVE (variance shared between the constructs and their measures). Off-diagonal elements below the diagonal are correlations among the constructs, where correlations between 0.12 and 0.16 are significant at p < 0.05 (two-tailed), and correlations above 0.16 are significant at p < 0.01 (two-tailed). Because the square root of each reflective construct's AVE is higher than its correlation with another construct, discriminant validity is established in light of the Fornell-Larcker criterion (Henseler et al., 2009). Off-diagonal elements above the diagonal are the heterotrait–monotrait ratio of correlations (HTMT), italicized values are the confidence intervals; because the HTMT value is always below 0.85, and bias and corrected confidence intervals at the 99% level of significance do not include 1, discriminant validity is supported (Henseler et al., 2015). SL, servant leadership; SC, service climate; CSP, customer service performance; SUE, service unit education; SUS, service unit size; SD, standard deviation.

#### TABLE 3 | Cross-loadings matrix for reflective constructs.


Bold figures indicate that each item loaded highest on its associated construct, so all these constructs are conceptually distinct (Henseler et al., 2009). CSP, Customer.



For testing independent variables' effects: ∗∗∗p < 0.001 (one-tailed test): t(4999) = 3.092, ns: not significant.For testing control variables' effects: †p < 0.10 (two-tailed test): t(4,999) = 1.645, ns: not significant. <sup>a</sup>Based on a bootstrap test with 5,000 re-samples, the indirect effect b = 0.20 is significant at p < 0.01. The bias-corrected and accelerated 95% confidence interval (CI) does not include the zero (CI lower level = 0.11; CI upper level = 0.31). Effect sizes of R<sup>2</sup> ≥ 0.01, ≥ 0.09, and ≥0.25 are small, medium, and large, respectively (Cohen, 1988). CSP, Customer Service Performance.

mediation. Likewise, our results confirmed the third condition, because service climate had a significant, positive, direct effect on customer service performance, which was also medium in size (R <sup>2</sup> = 0.11; **Table 4**). Finally, according to the fourth condition, there was a significant positive effect of servant leadership on customer service performance in a model in which the mediator was excluded (βUnmediated Model = 0.19, p < 0.01), but dropped to null when the mediator was added, implying full mediation (βMediated Model = −0.03, n.s.) (See **Figure 2**). In summary, although this mediation effect was small in size (f <sup>2</sup> ≥ 0.02; **Table 5**) (Cohen, 1988), our results reveal that service climate mediates the relationship between servant leadership and

customer service performance, prominently. Thus, the positive impact of servant leadership on customer service performance is not direct, but indirect, through enhancing service climate, in full support of H4.

#### DISCUSSION AND CONCLUSION

#### Theoretical Contributions

In the service industry, the quality of employee–customer interactions is deemed a critical aspect to gain excellent customer service performance. Such encounters often represent

TABLE 5 | Mediation effect size of service climate.


f <sup>2</sup> = (R<sup>2</sup> included – R<sup>2</sup> excluded)/(1–R<sup>2</sup> included); effect sizes of f<sup>2</sup> ≥ 0.02, ≥0.15, and ≥0.35 are small, medium, and large, respectively (Cohen, 1988).

the only contact customers have with the organization, so managers should manage these interactions properly (Solnet, 2007). However, although traditional control mechanisms can help manage such interactions (i.e., reward and punishment systems, incentives), it is not enough to control every conscious or unconscious word, gesture and attitude employees may show. With this study we provide new strategies far less based on control mechanisms as helpful in approaching customers properly, especially in the current times in which customers are increasingly social-aware. Specifically, we contributed to showing how the generation and development of servant leadership behaviors in managerial roles within service units leads to higher customer service performance, if only because servant leadership enhances service climate. It is important because it sends the clear message to managers that what they do in their day-to-day worklife, including attitudes, gestures, words, behaviors, matters to the point of enhancing service climate, and, in turn, customer service performance of their service units.

In this study we focused on the mediating role of service climate in this relationship as a broad body of prior research, known as linkage research (i.e., Wiley, 1996; Schneider et al., 1998; Wiley and Brooks, 2000; Pugh et al., 2002), has highlighted its significant role as the bridge which links employee perceptions of internal factors with important external criterion measures of outcomes such as quality of employee–customers interactions. The results confirmed the key role of service climate in linking servant leadership to customer–servicer performance. As we expected, service climate mediated the relationship between servant leadership of managers and the service unit's customer service performance and did it in a complete way, as the direct effect of servant leadership disappeared when service climate was included. Such a finding is in line with and qualifies prior research (i.e., Salanova et al., 2005) as it puts on the table that one specific leadership strategy, i.e., servant leadership, which is gaining increased attention over the years (van Dierendonck, 2011), is powerful in enhancing service climate, through which customer service performance can be

ultimately improved. Servant leadership is unique in capturing servanthood, genuine concern in the growth of others, including workers, customers and the least privileged in society; this leadership approach shows care for social order as well as compassion and justice (Sims, 2005), which fits the rationales behind the values-driven Marketing 3.0 paradigm (Kotler et al., 2010). Hence, our study contributes to existing literature by highlighting the critical personal aspects (i.e., servanthood) managers should exhibit in their attitudes, values and behaviors, to shape a service climate which truly has an outstanding impact on customer service performance, especially in this new millennium in which customers are increasingly social aware.

### Practical Implications

The results of the present study allow us to suggest several implications from a practical, managerial perspective. For example, managers can use the knowledge in this research to note how important the leadership strategy they show is, and which aspects should be emphasized to improve customer firm performance. Specifically, managers should exhibit genuine servant behaviors, including genuine interest in serving workers, customers and broader society. To this end, the strategic plan managers can implement is twofold. First, training programs focused on coaching managers in the area of the servant leadership philosophy could be useful, even though the interest in developing a servant leadership approach should come from the inner self (Ling et al., 2016). Indeed, servant attitudes and behaviors can be learnt, as well (Brownell, 2010), so by implementing training programs which enable more empathetic disposition and stronger concern about needs of others to be shown, managers could learn to develop servant leadership. Second, human resource managers should emphasize servant leadership traits when hiring new managers. Using personality tests involving specific items which evaluate personal aspects such as honesty, servanthood, stewardship or empathetic orientation (Hunter et al., 2013) could help find the right candidate for the position of manager.

In addition, managers should not ignore the important role of shaping a service climate within their service units to gain excellent customer service performance. Servant leadership, and its focus on serving others over and above oneself is principal in shaping such a service climate, and can involve, in turn, the design and implementation of a number of processes as described next. For example, servant managers should make sure that a human resources practices system oriented to both support workers in their day-to-day interactions with customers and provide these contact workers with the relevant knowledge and skills to succeed (by offering a high quality service) is properly implemented. Also, this system should serve: (a) to send clear information concerning standards of customer service to be provided, (b) to educate employees about how to perform in employee–customers encounters, properly, and (c) to design two-way communication channels which make managers realize problems and needs of employees in their daily tasks, and personal interactions with customers. Overall, managers should devote time and energy to serve contact workers, including providing due resources to approach customers, properly, so workers can share the idea that all the functioning of their specific service units focus on service quality, and thus emphasize a strong service climate.

### Limitations and Further Research

Our findings must be considered in light of some limitations. Some stem from our research design. One limitation is, for example, that because our investigation was designed in a crosssectional manner, we cannot offer strong causal inferences, so future research should include longitudinal designs to address our causality inferences more precisely. Also, our study was conducted in the customer service–oriented hospitality industry of historical sites situated in a specific cultural context (i.e., Spain); hence, future studies interested in generalizing our findings to other industries and cultural contexts should design cross-cultural studies spanning various, distinct service industries. Furthermore, although we collected our data by three different ways (employees, supervisors, hotel general managers), which improves data reliability, and minimizes CMB to a great extent (Podsakoff et al., 2003), future research could include customers to evaluate the dependent variable in our investigation (customer service performance), as well as aspects such as customer satisfaction and quality of service. In this connection, future research interested in advancing our findings could also ask customers about their system of values, by utilizing scales such as the Rokeach's (1973) or Schwartz's (1994) values surveys; this could help test if the servant leadership-customer service performance is contingent upon customers who are more or less socially aware.

Another important limitation is that we examined service climate as a relevant mediating variable between servant leadership and customer service performance, but other mechanisms might explain this relationship, as well. For example, employee service unit identification has been recognized as having an important role in gaining good employee–customer interactions and customer satisfaction (Solnet, 2007). Chen et al. (2015), find that this variable and other social identity factors (i.e., service unit self-efficacy) might have to do something in this relationship. Future research could evaluate the mediating role of such social identity variables in our relationship, and test whether service climate increases customer service performance via igniting higher service unit self-efficacy and employee identification. Also, a multilevel analysis which evaluates, within our research model, the influential role of individual-level variables that are often enhanced by servant leadership (e.g., service attitude, altruistic behavior) represents an appealing area for future research.

Finally, we examined the influential role of servant leadership of supervisors within service units. This choice was made on the basis that this is the person with whom workers spend more time and interact most, which allowed us to investigate the effects of servant leadership on service climate and customer service performance, more accurately. However, some other studies have also demonstrated positive effects of general managers' servant leadership on valuable organizational

outcomes (Peterson et al., 2012; Huang et al., 2016). Thus, an interesting area of future research is to examine the trickle-down effect of servant leadership within the organization, and test the combined positive effects of servant leadership in the various hierarchical levels on both service climate and customer service performance, and at either the service or organization unit level.

In short, our investigation provides key insights about new strategies, i.e., servant leadership, to gain customer service performance in a new era in which customers are more concerned about building a better society. This research also reveals the mechanisms, i.e., service climate, by which servant leaders boost such customer service performance in service units, and provides a map for avenues of appealing, ongoing research.

#### REFERENCES


### AUTHOR CONTRIBUTIONS

JL-L participated in the data collection. JL-L, PR-P, and DE-H worked on the theoretical framework, methodology, results analysis and discussion in equal measure.

### FUNDING

This work was funded by the Spanish Ministry of Economy and Competitiveness and FEDER Funds (project eco2016-75781-p). Plan Estatal de Investigación Científica y Técnica y de Innovación 2016–2020.



variable effects. Psychol. Methods 7, 83–104. doi: 10.1037//1082-989x. 7.1.83


Nunnally, J. (1978). Psychometric Theory. New York, NY: McGraw-Hill.


mediation of service climate. J. Appl. Psychol. 90, 1217–1227. doi: 10.1037/0021- 9010.90.6.1217


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Linuesa-Langreo, Ruiz-Palomino and Elche-Hortelano. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Behavior of Internal Customer in Family Business: Strategies and Actions for Improving Their Satisfaction

Santiago Gutiérrez-Broncano\*, Pedro Jiménez-Estévez and María del Carmen Zabala-Baños

Business Administration, University of Castilla-La Mancha, Talavera de la Reina, Spain

#### Edited by:

Ana I. Jiménez-Zarco, Open University of Catalonia, Spain

#### Reviewed by:

Carme Moreno, Stanford University, United States Marta Viu Roig, Open University of Catalonia, Spain Antoni Olive-Tomas, Universitat Ramon Llull, Spain

> \*Correspondence: Santiago Gutiérrez-Broncano santiago.gutierrez@uclm.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 29 March 2017 Accepted: 11 July 2017 Published: 25 July 2017

#### Citation:

Gutiérrez-Broncano S, Jiménez-Estévez P and Zabala-Baños MC (2017) Behavior of Internal Customer in Family Business: Strategies and Actions for Improving Their Satisfaction. Front. Psychol. 8:1266. doi: 10.3389/fpsyg.2017.01266 Determining the relevant aspects of family businesses (FBs) that make them increasingly competitive is the main objective of researchers in this field. Despite this, there is little empirical literature on the behavior of the internal customer (IC) in FBs or how businesses increase their satisfaction. Basing our work on psychological theories and with both quantitative and qualitative information from 31 semi-structured interviews, this work establishes certain characteristics of the ICs of the FB and proposes a series of guidelines for increasing their satisfaction, thereby facilitating the continuity of this type of business. FBs that are able to understand that the motivation of their ICs is more important than other qualities, and that this requires a more comprehensive management will be able to get sustainable competitive advantages in the future.

Keywords: family business, internal customer, socio-emotional wealth, human resource management, human organizations

### INTRODUCTION

Research in the area of family businesses (FBs) has grown exponentially in recent years. This is partly due to recognition of their economic impact, generation of employment, and contribution to economic development (Colli, 2003; Zahra, 2005; Zahra et al., 2008; Dawson, 2012), but also due to increased interest of researchers in this area, mainly due to the special characteristics of this type of business and also the results that they achieve (Sharma et al., 2007; Stewart and Miner, 2011). Recent evidence suggests that FBs significantly outperform non-FB (NFBs) in sales and assets (McConaughy et al., 2001; Anderson and Reeb, 2003; Anderson et al., 2003; Weber et al., 2003), maintaining a higher revenue for the first generation (Weber et al., 2003) and possibly firmer longevity (De Geus, 1997; Mackie, 2001).

Scholars have tried to explain this advantage with several reasons that Dawson (2012) groups in three levels: group level (social relations), firm level (familiness), and individual level; but no conclusive result has been obtained. This study is focused on the individual level, as it is where we find the least literature, more specifically viewing the members of FBs from a marketing perspective, as internal customers (ICs).

Today, businesses must get their competitive advantage through the effective use of all resources, especially those that are more difficult to imitate, to achieve more sustainable distinguishing advantages. Although the literature has paid close attention to external customers, it has not done

the same toward ICs (Stanley and Wisner, 1998). Many writers place both types of customer on the same level (i.e., Stanley and Wisner, 1998; Chang and Huang, 2010; Jun and Cai, 2010) due to the degree of difficulty involved in both their ability to imitate and their effective management.

Internal marketing (IM) seeks to attract, develop, demonstrate and retain qualified employees (Berry and Parasuraman, 1991) and these employees are seen as ICs (Foreman and Money, 1995). It has traditionally been understood as a management philosophy focused on effective internal relationships between individuals at all levels of the organization (Keleman and Papasolomou-Doukakis, 2004). However, today the theory of IM shows us how to make employees more satisfied, motivated and prepared to act better toward external customers (Joseph, 1996). To do so, the company has different tools that it must promote, such as effective internal communication, necessary training and development, encouragement of work teams, development of a well-defined organizational structure, and recognition and empowerment of employees (Bennett and Barkensjo, 2005). Among the literature on Strategic Human Resource Management, we also find empirical evidence of the impact that management oriented toward IC satisfaction has on the creation of competitive advantage and improved performance of the organization (Ullah and Yasmin, 2013). Little is known with regard to the process through which this value is created (Wright et al., 2003) and knowledge of the way in which FB achieves this objective is practically non-existent (Reid et al., 2002).

Traditionally, the owners of FBs are often criticized for employing less-skilled members of the family to carry out high responsibility functions (Chrisman et al., 2004), but what is true is that the IC has characteristics that make them especially unique, belonging to two systems at the same time: that of the business and that of the family. This leads to behavior which in certain cases favors the achievement of advantages for the company, but in others may be harmful (Nicholson, 2005b). Analyzing what is the most effective behavior of the IC in FBs and determining what set of strategies and actions carried out within businesses which help to increase the level of IC satisfaction is the main objective of this research work. There is empirical evidence of how the satisfaction and commitment of employees serves to increase effort and show more energy and enthusiasm from employees at work, thus improving performance and customer service (Alfes et al., 2013; Menguc et al., 2013; Karapete and Demir, 2014; Paek et al., 2015). Achieving IC satisfaction, we increase both motivation and business performance.

This orientation toward IM includes three essential aspects (Ruizalba et al., 2015): collecting information, internal communication, and responding to detected needs. Following these steps, this work begins firstly with the compilation of information on the characteristics of the IC in the FB, allowing their behavior to be known and better predicted, and secondly analyzing the main strategies that FBs carry out both for conveying and for implementing different actions that will help to achieve greater satisfaction of ICs. Finally, research findings, discussion, academic implications, business practices, limitations, and future research lines are also presented.

### LITERATURE REVIEW

### Importance of Internal Customer in Family Business

To be able to understand the importance of the IC within the FB we must firstly know what has been traditionally understood by FB. Scholars have understood FB as a combination of two systems: family and business, and the overlap between them (Habbershon et al., 2003). The definition has always been related with regular interaction between members of the family and the control of ownership, and this generates unique characteristics, which may in some cases generate competitive advantages, or conversely cause risks and vulnerabilities to the business and family (Nicholson, 2005a).

The definition of FB has focused on some combination of the four components of the family's involvement in the business, which are ownership, governance, management and transgenerational succession (Chua et al., 1999). If we search for a theoretical definition, we can understand FB as the family's influence over the strategic management of a firm (Davis and Tagiuri, 1989), or as the intention of the family to retain control (Litz, 1995), or potentially as the unique, inseparable, synergetic resources, and capabilities arising from family involvement and interactions (Habbershon et al., 2003).

The family's involvement favors the notion that the FB often possesses unique characteristics and sources of competitive advantage compared to the non-FB. Among its sources, Zahra et al. (2008) include reduced agency costs through owner control, longer investment time horizons, increased commitment to intergenerational wealth, stronger investment time horizons, lowered transaction costs due to a higher level of trust, and a less formal or more flexible organizational structure (Geeraerts, 1984). All of these resources and capabilities related with family involvement and interactions are referred to as "familiness." Familiness is defined as the idiosyncratic set of resources and capabilities resulting from the interaction between two systems: family and business (Habbershon et al., 2003; Zahra, 2005).

Familiness is believed to be of such importance due to it reflecting the positive aspects of overlapping subsystems of family and business (Habbershon et al., 2003), and it is often used as a unique element that can differentiate family and NFBs (Pearson et al., 2008). Nevertheless, these interactions between family and business do not always result in a positive contribution to the FB. Recent research has shown that family influences may have negative effects on the FB as well (i.e., Kellermanns et al., 2012).

When we focus on analyzing the characteristics of the IC, we realize that this refers to the human capital of the FB and find that members of the family have a greater degree of commitment and cooperation compared to non-family employees (Barnett and Kellermanns, 2006). This human capital is a vital resource for the strategy of organizations (Colbert, 2004), but we cannot forget that the knowledge, skills and abilities of individuals are not enough if we want to create value for an organization (Wright et al., 1994; Colvin and Boswell, 2007).

Following resource based-view (Barney, 1991), human capital is the most valuable and difficult to imitate resource, as it is the result of a complex social structure established over time,

essentially in FBs. Human capital has included knowledge, skills and abilities, but recently, Hoy and Sharma (2009) have included an intellectual and psychological dimension. In the context of FBs, human capital includes factors such as commitment and emotions as well as integrity, compassion and forgiveness of family members (Puhakka, 2002; Dawson, 2012), with all these factors being referred to as "family human capital."

For this reason, Kidwell et al. (2012) consider increased attention to employees or ICs in FBs to be necessary. They are important not only for the fast growth of FB but because FBs have two additional characteristics, which affect how they manage their ICs, the presence of family and non-family members, as well as the family influence (Botero and Litchfield, 2013).

To better understand the characteristics of the IC in the FB, Pieper (2010) proposes making greater use of psychological theories, which will improve understanding of the subject. Concepts such as motivation, power and authority, obedience, groupthink, group cohesion, leadership and commitment must be taken into account upon implementing a strategy adapted to FB to improve IC satisfaction.

### Characteristics and Behavior of Internal Customer in Family Business

Traditionally, research carried out on FB has been undertaken with a smaller theoretical basis, mainly focused on agency theory and resource-based view (Chrisman et al., 2010). In whole, agency theory has been somewhat criticized, as it does not exhaustively consider cooperative behavior between the members of the family (Eddleston and Kellermanns, 2007). However, as we previously stated, in recent years, authors have appeared who defend the need to apply psychological concepts to understand certain essential aspects in FBs (i.e., Björnberg and Nicholson, 2007; Pieper, 2010). These theories incorporate aspects such as motivations, emotions, knowledge, personal bonds, expectations and family relationships, which are priorities in the FB (Schulze et al., 2001) and which have not been deeply analyzed. Individual psychology focuses its interest on perceptions, emotions and motivations, while social psychology helps to understand how all of these variables are influenced by the presence of other people (Pieper, 2010).

When we focus on IC of FB, we must be aware that they have very specific characteristics, essentially being members of the same family. These individuals are interested not only in aspects related with their obligations and duties to the company and derived from their job, but also show interest in aspects such as, for example, loyalty, reputation, security, love and affection, cooperation, etc. (Rothausen, 1999). Recently, the theory of socioemotional wealth (SEW) has moved in this direction. It is established that FBs have a preference for objectives which are not exclusively economic, but focused on the maintenance of family influence and control, the identification of family members with the company, the strengthening of social bonds, and even the tendency to increase the emotional attachment and affection of the family toward the business through dynastic succession (Berrone et al., 2012), or even family harmony (Zellweger and Astrachan, 2008).

A study carried out by Björnberg and Nicholson (2007) shows us the existence of both positive and negative emotions between members of the FB (Schulze et al., 2001) and expands its study to analyze the different motivations existing in them.

This leads us to suggest that FB is in many ways more consistent with human design than other NFBs. Nicholson (2008, p. 106) even goes so far as to define it as "the primary economic unit in the history of our species." The extreme overlap of family and business allows the IC's need for dependence and acquisition of status in the same measure, mainly due to the possibility of intergenerational transfer, the maximum responsibility generated with the economic unit of the family and the commitment of all members, whether family members or not. This has a direct effect on the IC, as they feel free to express their emotions, trust in their lasting bonds, their personal and organizational objectives being closely aligned with ownership, thus achieving greater satisfaction.

Nicholson (2008) made one of the main studies of the few existing on the characteristics of the members in FB. Which based on evolutionary psychology establishes different characteristics, which in certain cases differs from existing literature on FBs.

There are various types of motivation, which the IC generally has. For example, those who are intrinsically motivated mainly seek their own interests, enjoyment or satisfaction, while those who are extrinsically motivated do so in pursuit of a certain goal, which is usually economic (Abelson et al., 2004). This extrinsic motivation may generate cooperative behaviors, but in a short time period. However, in FB there is a strong element of altruism (Karra et al., 2006). This altruistic behavior is not only between members of FB, but also toward the exterior of the business, affecting the communication, participation and between satisfaction of IC (Van den Berghe and Carchon, 2003). Aspects such as owning their own business and the possibility of transferring it to the next generation, feeling proud of their surname and working as a family have been found to be reasons as or more important than financial reward (Astrachan and Jaskiewicz, 2008).

Furthermore, due to the interaction between the two systems, the members of FBs develop a tacit knowledge and greater social intelligence (Sirmon and Hitt, 2003). This stock of social capital includes personal contact in the business, networks, the ability to remember faces, develop empathy, and analyze the main motivations of others (Lee et al., 2003). All of this enables them to perceive betrayal or the breaking of agreements more reliably, and even detect opportunistic behavior (free rider).

Compared with what is traditionally thought of teams work, in FBs there is a greater diversity based on this affirmation of what Nicholson (2008) refers to as the "genetic lottery," which provides each member of the family with unique attributes and different potential (Ilies et al., 2006). This makes teams in FBs more heterogeneous than the majority of teams formed in NFBs, which use standardized criteria when selecting members.

With regard to management style, Nicholson (2008) differentiates between leadership gets by inheritance compared

with leadership by selection. Their behavior differs, inheritance leaders know and accept their limitations better, leading them the propose new forms of leadership (i.e., co-leadership) and to see themselves as servants of the business and responsible for the family legacy, and serving to attract the loyalty of others. Meanwhile, selected leaders perceive greater capability than the rest of their colleagues.

The existence of blood relationships between the different members of the FB has been criticized by the literature due to the risk of falling into nepotism, this being understood as the tendency to favor members of the family over nonmembers (Kruger, 2003; Neyer and Lang, 2003). However, this characteristic may also be seen to generate advantages when cooperating and generating confidence between the different members of the FB (Arregle et al., 2007). This situation favors the creation of large clans that share common interests and creates an ideal environment for generating unity, collaboration and cooperation between its members.

Conflicts have also been seen by literature on FB as a great threat (Zellweger and Astrachan, 2008). However, we must differentiate between types of conflict. For example, certain levels of conflict (slight or moderate) in tasks or processes incentivise more innovative behavior and improve results (Pieper, 2010). Conversely, in FB, the conflicts of greater risk are those, which are interpersonal and inter-group, generated for different reasons, which lead to inefficient work (Egea Romero, 2014). There are different types of interpersonal conflicts in FBs: between parents and children, between siblings, between parents, and between cousins. The conflict between parents and children arises because while one group protects its genetic investment, the other pursues its own autonomy. When there is conflict between siblings it is due to them competing for what they consider limited resources, whether economic or emotional (Hertwig et al., 2002). However, the conflict that is the greatest danger in the FB is between parents, due to its danger of divorce, and because in the majority of cases it puts the survival of the business in serious danger (Davis and Harveston, 2001). Despite all these potential conflicts, when they are overcome, it leads to faster decision-making, more effective communication and greater loyalty and commitment of members (Tagiuri and Davis, 1996).

Finally, with regard to more informal and more irrational behavior that characterizes FBs, this is a more natural human behavior. Aguado (2005) describes emotion as being older than cognition, and the organism trusting in it more than in rationality. Emotions represent our first contact with the reality that we perceive, determining the direction of our responses, and it is later that we give a cognitive explanation to these emotional sensations. Therefore, the human mind does not operate like a computer, and is instead imprecise, illogical, selective and intuitive. It is organizational rationality that attempts to correct is through systems and routines. This is the main reason why we find structures that are more organic in FBs, for integrating both systems in a more natural way (Denison et al., 2004). Harris and Reid (2008) found that FBs use four times less work commissions and formal meetings compared with indirect communication with their employees than non-FBs.

## Strategies and Actions for the Improvement of Internal Customer Satisfaction

Research directly related with strategies and actions for the improvement of IC satisfaction is scarce and fragmented. Furthermore, the lacking empirical evidence is inconclusive and even has contradictory results. For this reason, and based on literature on strategic human resource management and on SEW theory, we establish a series of actions of FB that may improve the satisfaction of the IC.

In the previous section, it has been demonstrated how IC in FB has different characteristics and behaviors from other types of customers. When business shows an interest in the satisfaction of the IC, actions must be established that considers these characteristics, as it is a member of the business that also belongs to the family, and this make the process more complex, requiring it to be dealt with in a comprehensive and holistic way.

Strategic management of human resources establishes the main objective of attracting, developing and retaining the best human capital for organizations. This is achieved through practices such as recruitment, selection, training and development, performance evaluation, remuneration system, participation and communication (Wright et al., 2003; Shih et al., 2010). This set of practices seeks performance through the participation and satisfaction of the IC (Guthrie et al., 2002), and has been referred to in several ways: high performance work system (Huselid, 1995), high involvement work system (Edwards and Wright, 2001) and high commitment work system (Arthur, 1992). Its main characteristic is that, among other objectives, greater satisfaction and commitment is obtained among ICs (Ostroff and Bowen, 2000).

Firstly, the strategies and actions for attracting and selecting employees has sought alignment between the employee and the job to fill (Gomez-Mejia et al., 2012). However, in the FB it seems more logical, given that it has a double objective (economic and emotional), to think that it seeks a greater alignment with the organization, rewarding the coinciding of values of the individual and the organization. In this way, the IC will find greater satisfaction not only in what they do (task), but in how they do it (process) and with who they do it (team). It is necessary to introduce tools that make the IC enjoy their work and feel emotionally attached to it Paek et al. (2015).

In the same way, actions focused on training are oriented to the employee learning specific skills and improving their performance. While development intends for the employee to acquire a large number of personal skills and competencies, which in the majority of cases are acquired through different experiences that they receive from different jobs through which they pass (Gomez-Mejia et al., 2012). FBs consider that IC satisfaction is greater when the employee achieves their own professional development, and less so due to alignment with the job (Cruz et al., 2011). Following the model of job characteristics developed by Oldman and Hackman (2010), which is widely accepted in the field of organizational psychology, it is established that skill variety, task identity, task significance, autonomy and feedback are the essential factors for achieving greater IC

satisfaction, better performance, greater commitment to the FB, and thus obtaining what is referred to as "psychological ownership" (Oldman and Hackman, 2010).

Additionally, the evaluation of performance and the remuneration system are very useful actions for aligning the objectives of each individual with the objectives of the organization (Subramony, 2009). However, FB does not specify as many measures for achieving this alignment, as both executives, employees, and all members of the business in general seek the best interests of shareholders, who in the majority of cases are themselves or a close relative. The payment of bonuses in NFBs is up to 10% greater than in FBs (Anderson and Reeb, 2003). In this way, it is logical to think that FBs do not also require an evaluation and remuneration system as sophisticated as NFBs to achieve the same result (Cruz et al., 2011).

Finally, all organizations specify an effective flow of communication, as this is crucial for the proper operation of any kind of business (Riggio and Lee, 2007) and FBs are not an exception. Furthermore, effective communication does not only contribute to improved productivity, but also leads to greater satisfaction, motivation and increased morale of IC in FB, and thereby their level of commitment to the organization. For FBs, the use of formal communication channels is less frequent than NFBs (Pittino and Visitin, 2013).

After reviewing the literature, we presented the following research questions. How does the IC behave in the FB? Which are the main strategies and actions undertaken by the FB with respect to its ICs? And finally, do different perceptions regarding distinct agents implied in the study and management of the FB exist?

### METHODOLOGY

#### Design

The research was carried out through semi-structured, in-depth interviews, as these are highly effective for understanding what people think or believe about certain aspects (Krueger and Casey, 2014). In this way, the objective of identifying the main characteristics and behavior of IC in FBs, and the strategies and actions that successful FBs undertake to get greater satisfaction was achieved. Afterward, information was analyzed through two procedures, depending on the data obtained. One was focused on the quantitative analysis through the frequency analysis and contingency tables, and the other was qualitative analysis of the interview information (Braun and Clarke, 2006). One of the recommendations of the literature when this methodology is used is to select candidates who are truly qualified in the area of study, and to recruit them until a point of saturation is reached at which no new information appears (Krueger and Casey, 2014). The use of this methodology could also be justified because it is a little known subject as established Morse and Richards (2002). Otherwise, Yin (1994) recommends the use of this methodology to analyze a situation that has been studied infrequently and that is unique, and we can learn something new and important. By that, he establishes a protocol to justify the validity and reliability of the qualitative analysis that we have followed in this research.

#### Participants

Purposive sampling was undertaken for the selection of participants. The most important factor was selecting specific participants who were the most prepared to provide the researcher with the best understanding of the issue being analyzed and to respond to the object of this research in the best way possible. Ethical approval was not required for this study in accordance with the national and institutional guidelines. Firstly, experts were selected from outside of FB such as professors, researchers and consultants of this type of business, all belonging to the FB department of the Scientific Association of Economics and Business Management (ACEDE)<sup>1</sup> . Secondly, managers and employees who were or were not family members, belonging to different FBs were sought. The recruitment was stopped when a point of saturation was reached in the data collected. Ultimately, the sample consisted of 31 participants, of whom 14 were professors, researchers or consultants of FBs, 6 were family member managers, and 11 non-family member employees of FBs. The sample thereby contained interviewees who were wellqualified, although from different backgrounds and with different experiences, so that the sample was sufficiently diverse. Of the 31 participants, 17 were men and 14 were women, with ages spread in the following way: 9.7% under 30; 29% between 31 and 40; 45.2% between 41 and 50; 12.9% between 51 and 60%; and 3.2% over 60 years old.

Different works which use a similar methodology use similar or even smaller samples (Galea et al., 2014; Lyddy and Good, 2016; Mehra et al., 2016).

#### Data Collection

The data was collected between November 2016 and January 2017, and all participants were informed of the confidentiality of information and the purpose of the work, all-agreeing to participate in it. When possible, interviews were conducted in person, but due to logistical issues, some were conducted by telephone. The 31 participants were guided with semi-structured interviews, with a list of topics that served as a guide, following the instructions set out in the literature (CBO, 2004). The list of topics includes general aspects aimed toward the IC, influential psychological factors, actions undertaken in recruitment and selection of employees, training and development, evaluation and remuneration systems, and participation and communication of the IC. The guide of the interview includes both open and closed questions and a panel of experts with different experiences reviewed it before being carried out. One of them was a researcher and another was a senior manager of a FB. In this way, we contrasted the validity of the interviews (Yin, 1994).

The interviews were carried out in Spanish, as all were intended to be carried out in a common language. Furthermore, both the interviewer and the interviewee are highly qualified and this facilitated understanding of thoughts and ideas. A special emphasis was placed on there being no right or wrong answers

<sup>1</sup>ACEDE is an association of university professors interested in improving scientific research on the main issues arising from business administration and disseminating this knowledge between both students and managers. It was created in Spain in 1990. More information can be found on its website, at http://www. acede.org/.

and that the interviewer equally valued all opinions. 73% of the interviews lasted between 20 and 30 min, and the remaining 27% lasted longer than half an hour.

Once the interview was completed, the researchers analyzed whether new information had been collected or whether the point of saturation had been reached, thereby determining when to proceed to analysis of the data. Once the interviews were transcribed, the information from the closed questions was statistically analyzed, while the information from the open questions was encoded and the details of each category found was analyzed. Two researchers undertook this encoding separately, and the different classifications were analyzed until a consensus was reached. Only in the case of not reaching consensus did a third party intervene to decide the most appropriate category in which to include it. In this phase, special emphasis was placed on the ability to analyze the questions and the specific meaning of the words and phrases that seemed most significant (Corbin and Strauss, 2008). Although some experts (Morse and Richards, 2002) argue that a single researcher conducting all the coding is both sufficient and preferred, this could be particularly true in studies where being embedded in ongoing relationships with research participants but not in this case.

### RESULTS

### Quantitative Analysis

To analyze the information collected from the interview through the closed questions, a frequency analysis was carried out, which was very useful for describing the opinions with regard to the issues raised.

For the case of characteristics that shown by ICs of FBs, based on the opinion of the participants in in-depth interviews, the following data was found:


– Finally, relationships with greater problems may arise in FBs if they are conflicts between siblings (61%), between cousins (41.9%), between parents and children (25.8%) and finally between married couples (21%).

When we analyze the strategies and actions undertaken by FBs to increase the satisfaction of their ICs, we find the following results:


As we can observe from the information obtained from the interviews, there is not a consensus on the main characteristics of the IC or on what are the best mechanisms for generating greater satisfaction among them. For this reason, we have proceeded to analyze this information through contingency tables in order to determine whether there is a relationship between these opinions and the characteristics of the participants. Specifically, we have analyzed the contingency tables of each one of the questions analyzed, and the type, gender and age of the participant. This type of statistical analysis does not provide us with either the magnitude or direction of the association between the variables, therefore in the case of finding association between them, we have proceeded to analyze the contingency coefficient to discover these dimensions.

With regard to the age and gender of the participants, we must state that we found no relationship between their responses, but the same does not occur when we analyze the relationship with the participant type variable. In this case, the information

TABLE 1 | Association between variables of the study with regard to the characteristics of the IC.


<sup>∗</sup>EE, external experts (professors, researchers, and consultants); NFE, non-family employees; FM, family member managers.

obtained after contingency analysis is that there is a statistically significant association between the type of participant variable and the majority of variables studied as IC characteristics, as we can observe in **Table 1**.

As shown in **Table 1**, there is not a consensus on the characteristics of the IC in the FB. Depending on the group asked, we find different profiles.

EE group (78.6%) see nepotism as a great risk to the FB, and although they consider the predominant motivation among ICs to be intrinsic motivation (57.1%), they understand that they are less skilled in terms of social intelligence (64.3%) and that the roles that women play is of equal importance as in non-FBs (57.1%).

FM group evaluates nepotism similarly as a risk factor (83.3%) and intrinsic motivation as being predominant (57.1%); however, they do not find a difference in the roles of women in FB (50%) or in the degree of social intelligence of IC (50%).

With regard to NFE group, we can observe different perceptions. This group considers the main motivation to be transcendent (72.7%), the roles of women to be lesser than in non-FBs (90%), the IC having greater social intelligence (72.7%) and not perceiving nepotism as a risk (72.7%).

In the same way, we find an association between the participant type variable and some of the variables analyzed on strategies and actions carried out by the FB to improve IC satisfaction, as we can observe in **Table 2**.

As seen in **Table 2**, the strategies and actions that participants have established as those, which improve the satisfaction of ICs in FBs also differ between the groups from which we have obtained the information.

EE group believes that the selection of staff must seek alignment between the individual and the organization (61.5%), salaries must be below those of the market (77.8%), remuneration should incorporate non-monetary aspects (100%), evaluation systems should be focused more on qualitative and nonformalized criteria (85.7%) and more informal communication channels should be used (64.2%). Meanwhile, the FM and NFE groups have a different perception, more oriented toward the professionalization of the FB, seeking greater alignment between the individual and the job (100% FM and 81.8% NFE), standardizing and measuring selection criteria (100% FM and 90.9% NFE), salaries above those of the market (100% FM and 100% NFE), with remuneration that incorporates non-monetary aspects (66.7% FM and 50% NFE), the use of quantitative and formalized evaluation criteria (66.7% FM and 66.7% NFE) and improvement of formal communication channels (50% FM and 81.8% NFE).

#### Qualitative Analysis

As previously explained, the interviews were transcribed for subsequent study. From the encoding of the interviews carried out, we obtained a first result in which we checked that the information obtained based on the participants differing from each other, as occurred in the quantitative analysis. Below (see **Table 3**), we show the encoded results and select narrative bullet points, derived from the constant comparison between the texts, as suggested by Suddaby (2006). All qualitative information collected was analyzed following the theory approach (Corbin and Strauss, 1990) in which the main aspects are classified (Riessman, 1993).

**Table 3** shows some similar issues to those obtained in the quantitative analysis as EE group establishes trends related with the alignment of objectives, the use of more informal mechanisms and the achievement of gains which are not strictly economic as main differentiating factors of FBs which must be incentivised to achieve greater IC satisfaction. However, it is clear that the main aspect that FM group highlight is the orientation toward professionalism as a way to improve the cooperative behavior, motivation and satisfaction of the IC. Finally, NFE group shows a clear orientation toward the IC, focused on personalized relationships, communication, participation in decision making and commitment.

#### Research Findings

A number of conclusions have been reached. First, we find a greater motivational quality among ICs of FBs. Although literature establishes intrinsic motivation as predominant in FBs (Tyler and De Cremer, 2006), according to non-family members employees, FBs show more importance on the transcendental motivation based on altruistic behavior. This motivational quality


TABLE 2 | Association between variables of the study with regard to strategies and actions the improve IC satisfaction.

<sup>∗</sup>EE, external experts (professors, researchers, and consultants); NFE, non-family employees; FM, family member managers.

is shown in more personal and integrated relationship among members of FB, relatives or not. However, we were not able to reach consensus on qualities such as greater social intelligence, diversity of work teams, or the role of woman in these companies, which evolutionary theory assume (Nicholson, 2008).

According to nepotism, that literature has always seen it as a great risk to FB (Dyer, 2006), the FM group also considered it as a risk. In contrast, the NFE group do not perceive it as such. A possible explanation is that although they are aware that it may exist, they do not perceive it as an element that may have negative consequences either for themselves or for the survival of the business. Conversely, the fact that business remains within the family is a guarantee of continuity, due to the responsibility and effort that the managers show with the main economic support of the family, both in the present and the future.

Some similar occurs about internal conflicts among family members. Conflicts among siblings have been as the greatest risk to FB and it decreases when conflict is among cousins or among parents and children.

Another great aspect to consider is the disparity found both in quantitative and qualitative analysis regarding strategies and actions to be undertaken by the FB to allow the satisfaction of the IC, and thereby its success. First, the EE group focuses more on the development of more informal aspects related with the management of emotions and trying to extend qualitative and personal type aspects. Some studies establish that emotions can limit the firm in its ability to adapt the certain business demand (Vandekerkhof et al., 2015). By that, we put emphasis on develop emotions theory, if we want preserve of socio-emotional wealth and help FB to keep family control without it do not affect negatively the decision-making process.

Second, the FM group focuses on professionalizing the business and try to continue the same development that non-FBs. Sánchez-Marín et al. (2017) established that more formalized human resource practices, increase the financial performance of the company. Despite the fact that FB put more emphasis on nonformalized human resource practices adapted to non-economic goals related to the welfare of the family and employees.

And finally, the NFE group is the only one, which establishes that orientation toward the IC must be the essential pillar for achieving these objectives. This reinforced the idea that the impact of some practices on IC satisfaction create comparative advantage (Ullah and Yasmin, 2013). This support the basic premise that if employees in FB receive greater support from FB, they are like to reciprocate with greater affective commitment and greater efforts to contribute to achieving organizational performance (Huang, 2013).

From an integrated vision, we can establish the essential criteria that the FB must maintain a balance between these three dimensions. Professionalizing the business is essential, but FB must do so by integrating qualitative and emotional aspects into its management, with special orientation toward to IC. In this way, creating more personable organizations, which do not lose their orientation toward the market, but which are able to make decisions in which certain short-term economic objectives are renounced in order to strengthen the future competitive advantage of the business.

#### TABLE 3 | Information from the qualitative analysis.


<sup>∗</sup>EE, external experts (professors, researchers, and consultants); NFE, non-family employees; FM, family member managers.

#### DISCUSSION, ACADEMIC IMPLICATIONS, LIMITATIONS, AND FUTURE RESEARCH LINES

Although we found much literature in the field of FB, few studies empirically analyze its issues and focus on the IC. The initial research was based on psychological theories, which have been applied to FBs in recent years (Nicholson, 2008; Pieper, 2010). It contributes to focus on the research of FB from others points of view. Even then, psychology has a lot to offer to better understanding FBs and their particular behavior (Pieper, 2010).

Scholar have not paid enough attention on IC, despite the fact that individual human capital in FB is often inferior to that non-FB (Dawson, 2012) and FBs have difficulties managing their human resource, especially when it concerns a family members (King et al., 2001).

This paper is a response to the need for further research on this line to increase the knowledge about IC behavior, highlighting the strategies and actions that FBs implement to improve their satisfaction.

To that end, we build on a combination of evolutionary psychology (Björnberg and Nicholson, 2007) and SEW perspective (Berrone et al., 2012) to explain why FBs need use different practices to increase IC satisfaction. And we collected data from 31 participants with different backgrounds and experiences (experts, family business managers, and nonfamily members employees) to try to understand some important gaps and contradictions that there are still in literature on FB (Sánchez-Marín et al., 2017).

This paper has also academic implications for IM literature contributing new evidence to the importance academic debate about the effectivity of some practices in FB in terms of IC satisfaction. Furthermore, this study offers several contributions from the viewpoint of business practices. First, the paper emphasize that IC orientation gets a balance between economic goals and non-economic goals like continuity and preservation of family wealth (Sánchez-Marín et al., 2017). Considering that the main goal of the society is to create competitive and sustainable companies, FB should be seen like a more flexible structure, with a whole vision of the needs of company's members and thus, develop a motivational quality about their members. It makes we think to regards as more human organizations, whenever FBs do not fall into several traps like an excessive nepotism, the lack or insufficient professionalization in the decision-making process or

### REFERENCES


the loss of employees orientation as members of the company and family at the same time.

Non-family businesses can learn about FB to develop knowledge about emotions management, because it could help to develop a better workplace and a friendly climate and it would have an impact on the productivity and innovation of the company (Bammens et al., 2015).

Finally, this study is not without limitations, which may provide fruitful lines for future research. It would be desirable to include more participants in the research, using explicit measures for each dimension such as professionalization, motivational quality and IC orientation. It would provide more information about different types of FBs and the interrelation among variables and its effect on firm performance and continuity of FB.

In short, this study finds that an integrated vision and a balance among different objectives are needed in FB. According to experts, family managers and non-family employees, a balance among management emotions, professionalization and IC orientation are needed to reach the continuity of FB.

### AUTHOR CONTRIBUTIONS

SG-B and PJ-E played an important role in the design of the study. PJ-E and MZ-B took charge of carrying out the in-depth interviews and the processing of data. SG-B and MZ-B analyzed and interpreted the data. All authors reviewed the article and gave the manuscript their final approval.


review of 25 influential articles. Fam. Bus. Rev. 23, 9–26. doi: 10.1177/ 0894486509357920



**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer MR and the handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.

Copyright © 2017 Gutiérrez-Broncano, Jiménez-Estévez and Zabala-Baños. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Consumers' Loyalty Related to Labor Inclusion of People with Disabilities

#### Marta González\* and José Luis Fernández

Chair in Economic and Business Ethics, Comillas Pontifical University, Madrid, Spain

Purpose: the purpose of this paper is to show that reporting the corporate commitment to labor exclusion of people with disability correlates with the increase of consumer loyalty.

Methodology: It is a theoretical revision that will relate consumer loyalty to three main topics: disability and labor exclusion, responsible consumerism toward disability, and corporate communication to increase loyalty of those consumers that are concerned about this problem.

#### Findings:


Originality/value: This paper focuses on a topic usually neglected by companies and even by literature. However, the fact that more and more companies are paying attention to this problem allows us to think that we are facing a social change that will challenge companies.

Keywords: responsible consumer, loyalty, disability, functional diversity, labor inclusion, certifications

## INTRODUCTION

This paper focuses on how companies inform society about the commitment they hold with the labor exclusion of people with disabilities. We sustain the hypothesis that this information will increase the loyalty of those responsible consumers who take upon themselves to check if the products they acquire take into account the hopes and wishes of this group of people regarding their labor inclusion.

The methodology that we will use to prove the consistence of our statement will be a theoretical revision. It will link the concept of consumer loyalty to three topics: disability and labor exclusion, responsible consumerism toward disability, and the forms used by the companies to increase loyalty of those consumers that are concerned about this problem. Throughout the

#### Edited by:

Maria Pilar Martinez-Ruiz, University of Castilla-La Mancha, Spain

#### Reviewed by:

Cristina Olarte-Pascual, Universidad de La Rioja, Spain Manuela Saco, University CEU San Pablo, Spain

#### \*Correspondence:

Marta González marta.gonzalez@ubu.es

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 30 March 2016 Accepted: 30 May 2016 Published: 21 June 2016

#### Citation:

González M and Fernández JL (2016) Consumers' Loyalty Related to Labor Inclusion of People with Disabilities. Front. Psychol. 7:885. doi: 10.3389/fpsyg.2016.00885

literature it will be shown that more and more companies are reporting about their commitment with this problem, even the exclusion suffered by the collective is not widely known. It seems that they are sending a message for those consumers that could be interested in this challenge. Different methods of doing it will be exposed. As a line of research, the kind of information and how to provide it to satisfy this social concern remains to be determined.

This topic is important considering that 15% of population has a disability (WHO, 2011) and that this percentage will increase due to chronic diseases that will be overcome due to the increasing age of the population (Angeloni, 2013). On the other hand, the Council for Disability Awareness (CDA, 2013) highlights that 25% of young people that are 20 years old will suffer a disability before finishing their professional career. Moreover, we can't forget that behind a person with a disability there is almost always a family who directly or indirectly suffer or share the hardships of their relative with a disability (Angeloni, 2013; Adecco, 2014).

We can also say that we are facing a phenomenon that deserves attention for demographic reasons. Thus, it goes further than a marginal character of disability, but it concerns everyone. It may seem like an uncomfortable truth, but sooner or later everybody, will face some type of disability. Despite the fact that it is not widely known, nor envisioned in the long term (Coe and Belbase, 2015), physical and mental vulnerability are conditions that everyone experiences in the course of their own existence (Reynolds, 2008). With this in mind, the way that companies inform society about how they respond to the concerns about disability becomes crucial when talking about loyalty of potential costumers (Valor et al., 2012).

The concept of customer loyalty has received considerable attention in the marketing literature. In this sense, several researchers explain loyalty purely from the behavioral point of view (Jaiswal and Niraj, 2011), whilst some argue that an attitudinal perspective is more reflective of customer loyalty (Jacoby and Chestnut, 1978; Flint et al., 2011). Regarding the loyalty to a brand related to the labor inclusion of people with disabilities, our point of view embraces an integrated theory which suggests that customer loyalty is a combination of both behavioral and attitudinal loyalty (Dick and Basu, 1994; Oliver, 1999). This way of thinking is consistent with the main goal of this paper, which aims to link labor inclusion of people with disabilities, responsible consumerism, and loyalty to a brand.

Limited studies examine consumer's attitudes toward hiring people with disabilities. However, a national survey in USA found that 92% of consumers felt more favorable toward companies that hired them because it was assumed that these firms would care about their workers. Another positive aspect that was pointed out was that around 83% of the participants felt that companies did not take advantage of their workers with disabilities, nor did people with disabilities create problems in the workforce. Almost all of those surveyed (96%) shared the belief that companies who hire people with disabilities help those individuals to lead more productive lives. Furthermore, by including people with disabilities in their workforce, the participants also viewed companies that helped their employees as having a better understanding of people with disabilities. Finally, over one third of the survey group agreed that they would prefer to give their business to companies that hire people with disabilities (Siperstein et al., 2006). The results of this study show that the public view of hiring a person with a disability is considered as the socially responsible thing to do, as well as a gainful business practice. Thus, given these data, companies would have to find the right way to communicate this practice to the consumer because studies suggest that when the actions held by the companies are perceived as socially responsible they will influence costumer's purchase intention (Ajzen and Fishbein, 1980; Kuo and Kalargyrou, 2014).

Even the good perspectives and results given by the findings shown by the findings above, the labor exclusion of people with disabilities appears to continue over time. We can point out that in Spain a survey conducted by the NSI from 2009 to 2012 (NSI, 2013) showed that the employment of people with disabilities decreased gradually over that period, from 28.3 to 24.5%, in men, and from 23.7 to 22.5% in women. Unemployment, on the other hand, rose to a large extent, exceeding 6.2% in that of the population without disabilities. Unfortunately, this situation seems to be similar everywhere in Europe (ESS, 2010).

Several studies suggest that disability is one of the main causes of poverty all over the world (Laparra et al., 2007; Loeb et al., 2008; Braithwaite and Mont, 2009; Mitra et al., 2011; Martínez, 2013). This is due not only to the objective limitations of their state of health, but also to the fact that they tend to hold the worst positions when they can find a job (Malo and Muñoz-Bullón, 2006; Jiménez-Lara and Huete García, 2011). Additionally, they usually have a lesser income than the average population (Humer et al., 2007; Jones et al., 2007; Brown and Emery, 2008), and must meet the additional costs that their disability entails: care workers, removal of architectural obstacles in their homes, and medicines, among others (Durán, 2002; Parckar, 2008). For that matter, the arrival of disability in the home involves personal costs, particularly for women (García et al., 2004; Torns and Recio, 2012), who on many occasions are forced to give up their careers to look after the person directly affected by the problem (Malo, 2004; CERMI, 2012; Martínez, 2013; Adecco, 2014).

The controversial point of this paper is that despite the high percentage of population affected by some type of disability, the labor exclusion experienced by them, and its bad consequences, is still unknown by the large part of society and companies. It is widely accepted that economic help and assistance is the best solution to the problem (Peloza and Shang, 2011; Browne and Nuttal, 2013; González, 2015). Even so, we insist on the importance of communicating to society, in a measurable and assessable way, that labor inclusion of people with disabilities is a main challenge for the company.

Nevertheless, neglecting their labor inclusion has for years been legitimized by citizens who understood the labor exclusion of people with disabilities as a part of their condition of being "incapable" to work (Berger and Luckmann, 1971; Albrecht and Levy, 1981; Bickenbach, 1993); even the collective itself understood it, given their limitations to carry out "normal" activity (Seligman, 1972; Jost et al., 2003).

It is true that the concept of disability has changed over the years. It has been linked to several paradigms that had

considered it as a problem to be eradicated to accepting it as a part of society that needs to be healed and supported to become "normal" (Wolfensberger, 1975; Casado, 1991; Aguado, 1993). Under this way of thinking people with disabilities were accepted in the workplace, but their performance had to be equivalent to the other workers in the company (González, 2015). It appears, however, that way of understanding reality is starting to change, and more people are increasingly demanding companies to adapt the working environment to guarantee that people with disabilities can reach the labor market in equal conditions as others (Asís and Barranco, 2010).

We point out that this claim is made under a new paradigm of disability, in which the dignity of people and their moral autonomy are underlined. From this approach, disability is to be called "functional diversity", describing a vital situation in which people function in a different way to the majority of others (Romañach and Palacios, 2008; Ferreira, 2010). In this sense, the negative medical connotations that have historically accompanied this collective, under the clinical model, have disapeared (Romañach and Lobato, 2005; Palacios and Romañach, 2006).

This new way of facing disability was reflected in the International Convention on the Rights of People with Disabilities (2006), which focuses on equal opportunities, equal rights, and the dignity of people. The main effect of this conceptual change has been to consider those with disabilities as people with rights and not objects of charitable policies. It also entails understanding that the social disadvantage is a clear example of discrimination and breaches of human rights (Asís and Barranco, 2010).

However, as the employment results show, it appears that companies are not reflecting this new discourse. Companies, in general, are attempting to help people with disabilities with causerelated marketing and philanthropy actions (Min-Young et al., 2009; Peloza and Shang, 2011; Vauclair and Fischer, 2011). These kinds of actions are usually seen by society as a good thing (González, 2015), and therefore they are legitimized (Berger and Luckmann, 1971). But, if we bear in mind the new demands of the collective, it seems clear that the answer put forward by "benevolent companies" (Zamagni, 2012) is insufficient. Going further on this topic, we would dare to say that this way of marketing could be considered as harmful for society at large, because it leads to harmful consumption decisions (Smith et al., 2010). In other words, whenever a consumption behavior is influenced by this kind of publicity, the consumer is accepting benevolence and is neglecting the possibility of including people with disabilities into the value chain of the company.

Considering what we have exposed above, labor exclusion is a significant problem that is gradually coming to light and one which companies should respond to if they want to speak of responsible management (Freeman and Reed, 1983; Freeman, 1984; Carroll, 1991; Mitchell et al., 1997; Freeman and Philips, 2002; Young, 2003; Freeman et al., 2007).

Thus, there is a new character among the consumers that will pay special attention to the way the companies deal with this problem. In this paper we call this character a "consumer socially responsible toward disability", and we will link his behavior with Corporate Social Responsibility. In this sense, we consider responsible management as a preceding that includes, in a strategic way, the expectations and concerns of those persons that are around the company and are affected by it (Freeman, 1984, 2012; Carroll, 1999).

To be perceived as responsible toward the inclusion of people with disabilities, companies develop different actions to show society their policies in favor of these people. Some organizations show this through social events where their task is recognized and others show their commitment with a seal obtained by a certification process. In this paper we will focus on two kinds of certificates that are currently used to report business activity in the area of disability. The first one is already extended in more than thirteen countries and would label the firm as "Certified Professional in Disability Management" (CPDM). They work following the pattern established by the Integrated Disability Management (IDM). Its main goal is to avoid that disability blocks the professional career because of its arrival. The second one has been present in Spain since 2013. It is called the Bequal certificate and it analyzes the inclusion of disability following the criteria established by the Convention on the Rights of People with Disabilities (UNO, 2006). It can be said that these two certificates are quite different in their philosophy and about what they want to measure, but both of them gather most of the concerns of society regarding disability. These three models are not exclusive, as they could be used in a complementary way.

### SOCIALLY RESPONSIBLE CONSUMERS TOWARD DISABILITY

Socially responsible consumers (SRCs) are defined as citizens who are not just interested in satisfying their own needs, but who integrate their concerns for the environment and social causes into their purchase decisions (Arredondo et al., 2011), in a regular way (Stolle and Hooghe, 2004), and as a part of their personal project (Little, 1993). This kind of purchasing constitutes and expresses their identity (Newholm and Shaw, 2007), and their response to social or environmental issues does not come exclusively from cause-related marketing campaigns launched by companies (Roberts, 1996; Carrington et al., 2010; Smith et al., 2010). Consequently, a SRCs can be defined as the one that considers his acts of consumption as a chance to preserve the environment and the quality of life in a local context. In this sense, the SRC refuse those products that are dangerous for health, those that are packed in a non-ecological way, or might be harmful because of the material employed to produce it. At the same time, this consumer takes into consideration the responsible behavior of the companies and those products offered by fair trade. It can be said that price is not the only factor considered when the product is being purchased anymore (Akehurst et al., 2012).

It has been proved that the information given by the company, the Corporate Social Responsibility strategy, and the purchase intention are all related (Lee and Shin, 2010). The positive influence of Corporate Social Responsibility on consumer behavior has been shown by literature even when the

price of the product increase (Mohr and Webb, 2005; Alvarado and Schlesinger, 2008; National Geographic and GlobeScan, 2012). For this to be so, there is a relevant condition: the more knowledge and trust the consumer has about Corporate Social Responsibility, the better the response is (Tian et al., 2011; Kozar and Hiller Connell, 2013). In addition, if the beneficiaries of this socially responsible product are people, instead of animals or environment, the willingness to pay for it will increase (Tully and Winer, 2014). Something that has been pointed out is that women are more susceptible to support Corporate Social Responsibility than men (Arredondo et al., 2011) and are more likely to be influenced in their purchasing decision after a marketing related campaign (Hyllegard et al., 2010). But be that as it may, quality of the product is the main aspect considered by all consumers, including those who are called socially responsible (Feldman and Reficco, 2015).

In any case, the process to become a responsible consumer seems to follow a pattern strongly influenced by life experience. The process of becoming a responsible consumer is fostered during the person's childhood due to family values (Valor et al., 2012). The social environment will help to build their personal project. In addition to that mentioned so far, the degree of consumer knowledge of the problem they are trying to alleviate, and the perception of self-efficacy when engaging in a responsible purchase has to be considered (Gupta and Ogden, 2009; Valor et al., 2012). It is also important to bear in mind age, gender (Arredondo et al., 2011), socio-economic status (Fraj and Martínez, 2003), pertaining to a group involved in the defense or protection of social causes, and personal values (Fraj et al., 2004; Valor, 2008; Gupta and Ogden, 2009; Valor et al., 2012).

It appears that the character of responsible consumers has acquired great importance in recent years, and, as we anticipated, companies have to adjust to these new demands and find a way of making society aware of it if they truly want to influence the behavior of consumers (Camacho et al., 2013) and be sustainable (Moneva and Ortás, 2010; Fraile and Fradejas, 2012; Orozco and Ferré, 2013; Retolaza et al., 2014).

Not only do these new consumers question the way companies work, but they can also boycott their products or services (Paeck and Nelson, 2009) if they believe that they are not the fruits of responsible management (Freeman, 1984; McAlister and Ferrell, 2002; Porter and Kramer, 2006). Consequently, current companies must be prepared to offer information that differentiates them from their competitors. Such information must express not only the quality of the product or service, but also the values of the company and its management of its chain of worth (Auger et al., 2008; Russell and Russell, 2010; Shandwick, 2011).

Nevertheless, to be a responsible consumer is not an easy task. There are obvious barriers against putting the intention of responsible purchasing into practice, namely, the lack of time of consumers or the difference in price regarding other products (Beckmann, 2007), the difficulty in finding appropriate information on the chain of worth of companies that offer the service, and the lack of these types of products in normal consumer points of purchase (Valor, 2008; Carrington et al., 2010; Arredondo et al., 2011).

In line with, among other things, responding to this type of consumer, Corporate Social Responsibility has been increasingly emerging (Garriga and Melé, 2004; Fernández, 2007). This business model will seek to include social concerns into its management method (Freeman, 1984), without neglecting financial profit (Freeman, 1984; Kelly and White, 2009; Canals, 2010; Krauss and Brtitzelmaier, 2012).

Some institutions have developed verifiable management models, of voluntary application, to facilitate dialog with different affected public groups and to make society aware of the management of companies. Significant examples of these verifications are the Global Reporting Initiative (GRI) guidelines, the AA1000 model, the SA8000, the SGE21 standards, and the IS0 26000 recommendations and guidelines. Some of those are even certifiable through independent agencies, and, although all are voluntary, provide useful information when interpreting and comparing management quality and the companies performance (Andreu and Fernández-Fernández, 2011). It is noteworthy to underline that implementing socially responsible policies generate client loyalty, differentiates them from the competition, and attracts talent, investment and "responsible" consumers, and improves the working environment (Bhattacharya and Sen, 2004; Bhattacharya et al., 2009).

However, these tools have overlooked the existence in companies of people with disabilities. Evidence of this is that none of the standardized instruments for business auditing in the responsible management area has specific indicators on the inclusion of disabled people in companies. It is true that sustainability reports include sections on the Corporate Social Action (Moneva and Ortás, 2010), in which diverse activities such as sponsorship or voluntary actions in favor of people with disabilities are described. In this sense, nothing registers markers to measure how equal access to employment is guaranteed (Asís et al., 2007). In fact, it was not until 2013 that these parameters were included (GRI-ONCE, 2013). The involvement of people with disabilities in adding parameters that could report objectively the commitment of the companies can be considered as an indicator that they don't want help or charity but work.

In order to give a response to this new social demand, in the next section we present two main approaches to certify that disability is being taken into account within the companies.

### REPORTING CORPORATE COMMITMENT WITH DISABILITY TO IMPROVE CONSUMER LOYALTY

As we have already explained, the loyalty of the costumers comes from considering that the company is running the business in a responsible way. In other words, the costumer will be prone to maintain a purchasing pattern whenever he/she will feel that the firm has a behavior connected with his/her concerns. In this sense, companies will have to make a special effort to let society know how management is.

As Oliver proposed (1999), this information given by the company will mean the first step to build loyalty. This author defined a pattern that starts by the cognitive comprehension of the

information given by the brand. If it is accepted by the consumer, and brings a pleasurable fulfillment, it will become an affective loyalty showed in a positive attitude toward the brand. It goes without saying that if people perceive and understand that the brand is taking their needs and concerns into consideration, they will have a positive attitude toward the organization and its products and services. The next phase of loyalty development is the conative stage, where the consumer feels committed to purchase.

Following this idea, for some time now, several companies look for the acknowledgment of society of the role they are playing toward the exclusion of people with disabilities (Valor, 2004). To get it, they hold parties and events where they can assure their good practices are visible. Prizes and awards are given receiving media support. The main idea is to be perceived as an ethical brand, as a company that doesn't harm, but instead promotes it showing its honesty, integrity, diversity, responsibility, quality, respect, and accountability (Fan, 2005).

Despite being an obvious method of showing the reality of people with disabilities and a useful way of demonstrate to society that ways to eradicate exclusion exist, there is no doubt that the evaluation carried out of the proposals is neither official, nor has institutional recognition, let alone a standardized report, or control procedure. The visibility and impact of these types of actions appears somewhat limited, and, as it has been said, harmful for people with disabilities.

In other words, analyzing the value chain is needed to get to know how deep the commitment is that the company holds with people with disabilities and to endorse it.

To give verifiable information about how the company is managing diversity, and specifically disability, companies are looking to be certified. These certifications can focus on different aspects. As it has already been said, the CPDM seal will take into consideration how managers deal with disability in the company to avoid or decrease the risks and consequences of a disease that could turn into disability. It is reasonable to think that this approach will be accepted by a large group of population that considers that disability as a problem that has to be eliminated or healed to become "normal". But as it has been explained, the new claims of people with functional diversity consider that what causes the disability are the barriers that the person finds to be included. In this sense, the Bequal Certificate could generate more loyalty among the citizens that are trying to make society understand that diversity has to be respected.

Thus, the designation of "Certified Professional in Disability Management" (CPDM), offered by IEA (Insurance Educational Association) and DMEC<sup>1</sup> is given to companies that can demonstrate knowledge on the following areas: disability and work interruption case management; workplace intervention for disability prevention; program development, management, and evaluation; and employment leaves and benefits administration<sup>2</sup> . The CPDM designation is an internationally recognized certification accredited by the International Disability Management Standards Council (IDMSC) within 13 countries, namely, Australia, Austria, Belgium, France, Luxembourg, the Netherlands, Canada, Germany, Hong Kong, Ireland, New Zealand, Switzerland, and UK<sup>3</sup> . It can be used as a seal to improve communication and marketing to win new customers, new suppliers, and new partners. In fact, the growing public attention to the issues accentuates the level of ethical maturity of all stakeholders, which will reward companies that invest in health and safety (Angeloni, 2013).

In this sense the management of the company will follow an IDM that will pay special attention to any kind of disability that might overcome the workers of the company: specific health risks (e.g., physical inactivity, poor nutrition, tobacco use, stress, depression), conditions (e.g., obesity, muscle-skeletal disorders, mental health), and diseases (e.g., heart disease and stroke, high blood pressure, diabetes, high cholesterol, cancer, arthritis) can be addressed (Rieth et al., 1995; Calkins et al., 2000; Shrey et al., 2006; Rosenthal et al., 2007; Habeck et al., 2010; Angeloni, 2013).

In turn, the Bequal certificate would recognize those companies where managers could demonstrate that policies to avoid discrimination and that guarantee equal opportunities across all the areas of the are being implemented. Same criteria should be followed for suppliers, and has to be included in the implementation of occupational risk prevention. Special attention has to be paid to the existing limits for people with a different way of functioning: mobility barriers, communication and information barriers, or access to the product or service given by the company, and negative attitudes (Asís et al., 2007). If this analysis is not carried out, the information taken about how the company includes disability may overlook the heterogeneity of people with disabilities (Díaz-Velázquez, 2010; Ferreira, 2010).

It is worthy to mention that the Bequal certificate has been created by the most important organizations that represent people with disabilities in Spain<sup>4</sup> . It can be said that with this certificate they are giving voice and solution to most of their concerns, according to their personal experience (Valor et al., 2012), and that they are proposing an alternative themselves. Their participation in those decisions that affect their quality of life becomes vital to be supported by consumers (Stanaland et al., 2011).

Even though its implementation and actual scope can seem to be short-sighted, it is being implementing not only in national companies, but in several multinational companies such as: Repsol, Acciona or Gas Natural Fenosa, all of them having strong programs of Corporate Social Responsibility.

The certification granted by Bequal focuses on aspects that have not been considered by others standards to date:


<sup>1</sup>www.ieatraining.com

<sup>2</sup>www.dmec.org

<sup>3</sup>www.idmsc.org

<sup>4</sup>http://www.bequal.es/sello\_bequal.html


Thanks to these seals, companies can give reliable data that will offer understandable information about their way of running the business and their relationship with people with disabilities. In short, both certificates will report to the company cognitive and pragmatic legitimacy (Suchman, 1995). In this way, society will understand how management is held, how and why things are done. It will facilitate that the purchasing intention becomes a real fact in a sustainable way (Fombrun, 1996; Deephouse, 2000; Basdeo et al., 2006; Helm, 2007; Schwaiger et al., 2009). The more information the company facilitates, the more reliable it will be for the community (Berger and Luckmann, 1971; Suchman, 1995; Carreras et al., 2013).

Advantages to the company that will be rewarded by these seals will come from the fact of using a tool where the main point is to give an answer to expectations and concerns of people with a disability, or with a potential disability, and their families. In this sense, the behavior of the company is perceived as authentic and aligned with the values of people that could be affected by this problem (Becker-Olsen et al., 2006; Varman and Belk, 2009; Wagner et al., 2009; Ruiz de Maya et al., 2015).

As a result of this genuine collaboration, literature in marketing points out the following benefits for the company: greater sales volume, enhanced operating efficiencies, positive word of mouth, reduced marketing expenses, and enhanced consumer loyalty (Reichheld and Sasser, 1990). In addition, it has been found that consumer participation has a positive effect on consumer behavior (Dellande et al., 2004; Chan et al., 2010) which turns into brand loyalty (Bagozzi and Dholakia, 2006), commitment to the brand (Casaló et al., 2007), quality perceptions (Dabholkar, 1996), trust (Ouschan et al., 2006), and affective commitment to the product (Atakan et al., 2014).

We are aware about the main barrier of implementing these tools. It is related to the instrumental vision of Corporate Social Responsibility (Jensen, 2001), geared toward acquiring a brand image and obtaining a competitive advantage (Yelkikalan and Köse, 2012). Thus, the involvement of interested groups seems to be linked more to capital contribution than to achieving value through generating intangible good (Retolaza et al., 2014), ignoring a humanist approach aimed at people, their needs, and particularities (Melé, 2013).

That is why initiatives that try to show the commitment of the company with any social cause are frequently questioned. They are seen as manipulating actions of marketing, and the connection with and consistency of organizational actions are called into question (Ballabriga, 2009; Morata, 2010; Benavides, 2013).

In this context, these certificates could be misinterpreted; they could be understood that using them is not for the Common Good (Kuhn and Deetz, 2008; Zamagni, 2012; González-Fabre, 2015), but rather obeys other interests, whether it be the fulfillment of current regulations (Fernández, 2007; González, 2015), greater economic competitiveness (Friedman, 1970; Jensen, 2001), better brand positioning (Villagra and López, 2013; Benavides, 2015), or a method of obtaining social status (Yoon et al., 2006).

In any case, the existence of these kinds of certifications are making companies more ethical, trying to give a response to the concerns of others to make a better company for people with disabilities specifically and for society at large (Freeman, 1984).

### CONCLUSION

The main goal of this paper was to show that clear corporate information about the inclusion of people with disabilities might increase costumer loyalty.

As a result of our theoretical revision, we can say that this statement is consistent with the findings that we have exposed. In this sense, several arguments have been given to support it.

First of all, we have pointed out that disability is a phenomenon that concerns the entirety of human society, in the short or long term. Assuming this reality, the impairment that faces people with disabilities becomes an unacceptable social problem that companies should take into consideration if they want to be perceived as socially responsible.

Secondly, multiple researches show that companies that are considered as responsible are rewarded with the loyalty of the consumers. If those consumers are concerned about how the company is responding to the labor exclusion of people with disability, we have labeled them SRCs toward disability.

We have also shown that, in general terms, if the company is able to give clear information about how disability is treated, those responsible consumers will reinforce the loyalty toward the brand.

Even though we have underlined that the social exclusion is not really known by the society at large, companies are more and more aware about this question and they use different ways to show society their commitment to the problem. They may organize events to get direct acknowledgment from society because of their social actions in favor of people with disabilities, or they may implement programs to get an official certificate.

If we take into consideration that nowadays the collective is asking for being included in the companies and not for charitable actions, it is thinkable that they would consider that these certificates, CPDM as much as Bequal, more aligned to their own values than a social event that is hardly valuable.

Even though each certificate tends to evaluate different aspects the inclusion of disability in the company, both certificates together would reach the society as a whole. CPDM reminds that any disease can turn into disability and Bequal guarantees equal rights for all the citizens when going to work. Thus, any consumer properly informed about these questions may become a SRC toward disability and remain loyal to that brand or product. Even

if we see this problem improbable to happen in our life, it can appear in any moment, and we all know it.

### LIMITATIONS AND FUTURE RESEARCH

Despite our enthusiasm and conviction that certifying the inclusion of people with disability can decisively contribute to improving the quality of life of people with disability. Therefore, to society at large, we are conscious that there is still a long way to go. Society has to become aware of the problem, and to accept that it can happen to anyone. So, awareness-raising education programs are needed, and they could be led by companies to change values among the citizens (Camacho et al., 2013).

We have also witnessed that there is little literature focused on disability and consumer loyalty, as we have pointed out in our revision. This may be due to the invisibility of the problem and the permanent exclusion of society that people with disability have been suffering for years. The lack of research in this area can mean that society is not interested in it. We suggest that it is a good time to change this situation through scientific studies.

### REFERENCES


Some questions are still pending from our theoretical revision:


### MANAGEMENT MODELS WEBSITES


### AUTHOR CONTRIBUTIONS

MG and JLF have worked together on the state of art of this review. They have also worked together on the design of the certification tool that we introduce in this article.


2001 PALS. Department of Economics, University of Calgary. Available at: http://econ.ucalgary.ca/sites/econ.ucalgary.ca/files/publications/PALSworking paper2008.pdf


Jones, M. K., Latreille, P. L., and Sloane, P. J. (2007). Disability and work: a review of the British evidence. Estudios de Economía Aplicada, 25, 473–498.


Martínez, B. (2013). Pobreza, Discapacidad y Derechos Humanos. Madrid: Cinfa.



responsible consumption as a self-determined personal project. Hum. Ecol. Rev. 19, 159–174.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 González and Fernández. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Dispositional Employability and Online Training Purchase. Evidence from Employees' Behavior in Spain

Joan Torrent-Sellens <sup>1</sup> \*, Pilar Ficapal-Cusí <sup>1</sup> and Joan Boada-Grau<sup>2</sup>

<sup>1</sup> Faculty of Economics and Business, Open University of Catalonia, Barcelona, Spain, <sup>2</sup> Department of Psychology, Rovira i Virgili University, Tarragona, Spain

This article explores the relationship between dispositional employability and online training purchase. Through a sample of 883 employees working for enterprises in Spain, and a using principal component analysis and binomial logit probabilistic models, the research revealed two main results. First, it was found that dispositional employability is characterized by five factors: "openness to changes at work," "career motivation and work resilience," "work and career proactivity," "optimism and engagement at work," and "work identity." Second, the research also found a double causality in the relationship analysis between dispositional employability and online training purchase. However, this causality is not direct. In explaining dispositional employability, certain motivations and types of behavior of employees participating in online training are significant. In particular, greater sensitivity toward career-related personal empowerment, a greater predisposition toward developing new experiences at work, and a greater awareness of the fact that positive job outcomes are related to preparation conscientiousness. In explaining online training purchase, employees who are more motivated and who better identify with their jobs are more likely to pay. Moreover, employees who spend more time on training and have less contact with new trends in their jobs, find it hard to keep calm in difficult situations, and have a greater predisposition toward effort, and preference for novelty, variety and challenges at work are more likely to purchase online training.

Keywords: dispositional employability, online training, purchase, skills, career motivation, work resilience, work identity, Spain

### INTRODUCTION

In recent years, employability has become a growing field of economic, psychological, and social research (Thijssen et al., 2008; Smith, 2010; Hogan et al., 2013). It is acknowledged that employability is the ability to get and keep a job in a formal organization (Baruch, 2001; Harvey, 2001). The literature on employability emphasizes the important role of individual characteristics in adaptation at work. Ashford and Taylor (1990), and Ashford and Black (1996) claim that individual and psychological factors are essential components of effective adaptation at work. The premise is that employability is a synergistic collection of individual characteristics driven and directed by an individual's adaptability, career identity, and human and social capital (McArdle et al., 2007). Employability can be considered a psychosocial construct that enables career success (Fugate et al., 2004).

#### Edited by:

Monica Gomez-Suárez, Universidad Autónoma de Madrid, Spain

#### Reviewed by:

Jorge Sainz, Univeridad Rey Juan Carlos, Spain Esteban Agulló Tomás, University of Oviedo, Spain

> \*Correspondence: Joan Torrent-Sellens jtorrent@uoc.edu

#### Specialty section:

This article was submitted to Organizational Psychology, a section of the journal Frontiers in Psychology

Received: 26 April 2016 Accepted: 18 May 2016 Published: 01 June 2016

#### Citation:

Torrent-Sellens J, Ficapal-Cusí P and Boada-Grau J (2016) Dispositional Employability and Online Training Purchase. Evidence from Employees' Behavior in Spain. Front. Psychol. 7:831. doi: 10.3389/fpsyg.2016.00831

Within the psychological framework, empirical research on individual differences in career success represents the main contribution to the study of employability (Hogan et al., 2013). The literature on career success is typically organized into three combined models: views of human capital, of structure, and of social capital. According to the human capital view (Becker, 1975), organizations distribute rewards to their members according to their contributions, and employees' ability to contribute depends on having relevant skills and competing capabilities (Brown and Hesketh, 2004).

Although prior education plays a decisive role, the evidence suggests that the relationship between educational attainment and career success is only modest and acts as a "first-pass filter" (Pfeffer and Fong, 2002; Ng et al., 2005; Chamorro-Premuzic and Furnham, 2010). In this sense, and beyond the traditional relationship between cognitive ability and educational attainment, new research highlights the importance of personality characteristics and individuals' behavior (especially "effort," "ability," "curiosity," and "conscientiousness") as predictors of not only academic performance (Ng and Feldman, 2010; von Stumm et al., 2011) but also job outcomes (Roberts et al., 2007). Economic research has also obtained similar results (Lindqvist and Westman, 2011).

In contrast with the large body of research on the psychological determinants of career success, there has been little research on the determinants of employability (Baruch and Bozionelos, 2011). According to the literature, employability depends on identifiable personal characteristics that can be assessed and possibly trained (Fugate et al., 2004; Smith, 2010). Van der Heijde and Van der Heijden (2006) propose a five-scale employability model: "occupational expertise" (job competences), "anticipation" (job ambition), "personal flexibility" (high degree of openness, conscientiousness, and adjustment), "corporate sense" (disposition to behave in a social manner), and "balance" (work-life proportionality). Their best employability predictor was "corporate sense." Thijssen et al. (2008) propose an employability model that includes three dimensions: an individual's "ability to perform a job," "personal skills and learning capabilities," and "contextual factors" (organizational and social), which may influence future employment status. Finally, Wittekind et al. (2009) consider three employability dimensions: "job-related skills," "willingness to learn and develop new skills," and "knowledge of the labor market." Like career success, employability seems to be more a function of personal characteristics and behavior than of educational level.

Based on this evidence, a self-perceived employability approach (Ghoshal et al., 1999; Forstenlechner et al., 2014) has been developed in the literature. Perceived employability refers to an individual's perception of his or her possibilities of getting and keeping a job (Vanhercke et al., 2014). It is an individual's subjective evaluation of such possibilities and involves the integration of personal and structural (job, organizational, and social) characteristics (Berntson and Marklund, 2007). Consequently, perceived employability is relevant to different groups in the labor market throughout their careers. Evidence on perceived employability has been obtained from graduate students (Rothwell et al., 2009), employees (Rothwell and Arnold, 2007), and the unemployed (Wanberg et al., 2010).

Perceived employability can be applied to either current employment (internal labor markets) or future jobs (external labor markets) (Eby et al., 2003). Finally, it also concerns a focus upon both the quantity and quality of jobs (De Cuyper and De Witte, 2010). The literature shows that self-perceived employability is a robust measure across cultures (Rothwell et al., 2009) and is positively related to self-determination. Selfdetermination refers to a combination of skills, knowledge, and beliefs that enable employees to make decisions and re-evaluate past solutions, to generate new solutions if needed, and choose the best option to engage in goal-directed self-regulated behavior (Parker et al., 2010).

In the literature, two additional approaches have been taken to the explanation of perceived employability: the competencebased approach and the dispositional approach (De Cuyper et al., 2012). The competence-based approach focuses on an individual's perception of his/her abilities, capacities, and skills (specific or generic competences) that promote employment opportunities (Van der Heijde and Van der Heijden, 2006). The dispositional approach focuses on the perception of an individual's proactive attitudes toward his/her career and work in general (Fugate and Kinicki, 2008).

When comparing the three approaches, several similarities can be found. Firstly, all three approaches focus on an individual's perceptions: the labor market position in the perceived employability approach, employability abilities in the competence-based approach, and motivational attitudes in the dispositional employability approach. Also, all three approaches account for personal and structural factors, as well as their interaction, in explaining internal or external employability. Regarding the differences between the three approaches, it is important to note that while the perceived employability approach has been applied to different groups in the labor market and across career stages, the competence-based and dispositional approaches have been applied mainly to employees. Also, the perceived employability approach endorses both the quantity and quality of jobs, while this distinction is less relevant for the competence-based and dispositional approaches.

The flexibilization, segmentation, and individualization of labor markets, the advent of new work organization and human resources practices, and the global knowledge-based economy have fostered new approaches to the labor market in general (Torrent-Sellens and Ficapal-Cusí, 2009) and to employability research in particular (Brown et al., 2003). Some studies have noted the influence of the labor market's growing flexibility and the materialization of new contracts and labor relations frameworks (Esser and Olsen, 2012). Also noted in the literature is the emergence of new problems associated with structural change in employment (Green and Mostafa, 2012). From the point of view of new employability trends, recent research has also made significant progress. Specifically, it has highlighted the importance of employees' future job expectations, which are clearly linked to educational level (Gallie et al., 2012). Workers with a highly positive perception of the future of their jobs tend to have better employability and better quality jobs (Graso and Probst, 2012).

In the global knowledge economy, where technology and innovation are key to developing enterprise competitiveness, new value generation processes and sources of productivity inevitably call for trending changes in employability (Díaz-Chao et al., 2015, 2016a). In such an economy, creating and maintaining employment depends largely on employability factors, of enterprises' ability to generate jobs with trained, autonomous, committed and satisfied employees, who are able to innovate and create more added value (Díaz-Chao et al., 2016b).

Thus, within the context of changes in employability resulting from the advent of the knowledge-based economy, this article analyzes the determinants thereof. In this regard, it is important to note the following considerations. First, according to the literature on employability, our perspective goes beyond the traditional relationship between educational attainment and employability to take into account employees' motivations and behavior. Second, in accordance with the literature on knowledge-based employment, our perspective bears in mind the relevance of motivational attitudes toward new job-related trends. Third, as discussed below, we consider a specific form of training that has proved to be very important in the knowledge economy: online training. And fourth, we work with employees' motivational perceptions and with personal and structural factors. We therefore take a dispositional employability approach, in which we understand employability as "a constellation of individual differences that predispose employees to (pro)actively adapt to their work and career environments" (Fugate and Kinicki, 2008, p. 504). This perceived employability approach is reflected in five dimensions: "openness to changes at work," "work and career resilience," "work and career proactivity," "career motivation," and "work identity." Our first research hypothesis is:

H1. Employees who are more motivated to anticipate and optimize job changes, who identify with and are involved in corporate sense, who have less stress, who have self-efficacy or the ability to manage obstacles, overcome difficulties and achieve their goals, and who have hardy personality traits to commit, challenge, and control job situations are more likely to have greater dispositional employability.

As a result of the advent of the knowledge-based economy, the restructuring of business activity has also transformed the foundations of labor. The impact of knowledge-based innovation on the organization, conditions, and results of labor remains an open debate in the literature (Osterman, 2000; Neumark and Reed, 2004). On the one hand, the introduction of information and communication technologies (ICT) and massive flows of knowledge have gone hand in hand with an increase in cognitive demands, enhanced autonomy, a decline in hierarchical control, a rise in job creation and maintenance, and improvements in wages. But, on the other hand, it has led to a decline in labor based on routine manual and routine cognitive tasks, and the downskilling and destruction of employment in some segments of the population or industries (Autor et al., 2003).

In the context of labor market transformation, the debate on continuing training -particularly in online environmentsacquires special importance in the sense that continued learning processes are placed at the heart of career development and the improvement of employees' employability (De Vos et al., 2011). Although the positive relationship between educational attainment and employability in the knowledge-based economy has already been amply contrasted in the empirical literature on human capital (Heckman, 2004), in recent years an employability training divide has also been confirmed. In this regard, vocational education and training, understood as on-the-job learning activities, are mostly done by people who are already trained (Kyndt and Baert, 2013). One of the reasons that might explain this phenomenon is the lack of suitability between educational and training programs for employability and the socio-cultural complexity of jobs that are constantly changing (Billet and Choy, 2013). Also noted in the literature is the importance of certain student characteristics and types of behavior, such as their attitudes toward their careers and their self-determination, which would be clearly linked to their prior educational level (Kyndt and Baert, 2013).

Indeed, and despite their significant potential, the emergence, and consolidation of online training programs for employability has not managed to break down this educational barrier. The empirical evidence shows us, again, that online training for employability encounters problems when it comes to training those who are not trained. As is the case with non-online training, the use of learning methodologies that are not very collaborative, are undertaken outside the workplace, and are poorly adapted to the socio-cultural reality and specific characteristics of the students seems to explain the weak results of online training for employability (Inayat et al., 2013; Silva et al., 2013). In this sense, our second research hypothesis relates to the motivational and behavioral issues of employees performing online training activities:

H2. The effect of online training on dispositional employability is indirect. Employees participating in online training programs who are more sensitive to career-related personal empowerment, more predisposed to developing new experiences at work, and more aware of the fact that positive job outcomes are related to preparation conscientiousness are more likely to have greater dispositional employability.

Finally, and beyond their technological and pedagogical dimensions (Selim, 2007; Sun et al., 2008; Lee et al., 2009), employees' personal, and motivational determinants of online training enrolment and satisfaction (Rosenberg, 2001) have also been addressed in the literature. Regarding socio-demographic determinants, gender, and age differences in the effects of e-learning, including students' satisfaction and Internet selfefficacy, have been supported in prior research (Chen and Tsai, 2007). In this sense, new research has confirmed that emotional family support has both direct and indirect influences on students' perceived effects of e-learning (Chu, 2010). Tangible support significantly predicts adults' perceived effects of elearning, mediated by Internet self-efficacy. Compared to male adult learners, female adult learners rely more on tangible family support for increasing their Internet self-efficacy. The similarities between women and older adults imply that the issue is not specifically related to gender, but instead relates to the complexity of the social context of these disadvantaged learners.

Regarding the motivational and behavioral dimension, a set of determinants related to job features has also been identified in the literature. Brown (2005) shows that employees' perceptions of peer and supervisor support, job characteristics (especially workload and autonomy), and motivation to learn predict time spent on online training courses. In a comparative study, Rovai et al. (2007) provide evidence that e-learning students possess stronger intrinsic motivation than on-campus students who attend face-to-face classes on three intrinsic motivation measures: to know, to accomplish things, and to experience stimulation. In a learner loyalty approach, Chiu and Wang (2008) indicated that performance expectancy, effort expectancy, computer self-efficacy, attainment value (self-image and core personal values), utility value (current and future career goals), and intrinsic value (personally enjoyable activity) were significant predictors of students' intentions to continue using Web-based learning, while anxiety had a significant negative effect. In a perceived learning achievements approach, Paechter et al. (2010) show that students who attach a high value to specific achievements are also likely to invest more effort in learning, to apply more elaborate information processing strategies, or to dedicate more time to learning. In addition, flexibility in the choice of learning strategies and the exchange of innovative knowledge with peer students are positively related to learning achievements. Compiling this evidence, our third research hypothesis is:

H3. Employees are more likely to purchase online training if they are younger, women, more motivated, better identify with their jobs, have more time to spend on training, do not have recent contact with new trends in their jobs, find it hard to keep calm in difficult skill situations, have a greater predisposition toward effort, and preference for novelty, variety, and challenges at work.

### MATERIALS AND METHODS

#### Sample

The sample was selected by random accidental sampling (Kerlinger, 2001). The response rate was 83.7%. Of the original sample, 12.2% of the employees abstained from participating, and a further 4.1% of the questionnaires returned were rejected due to completion mistakes or omissions. After contacting the employees selected to take part in the study, the Web-based interviews were administered individually in work time with the prior consent of the enterprises' managers. The participants were given instructions to enable them to answer the questionnaires. They were also given an assurance about the confidentiality and anonymity of the data obtained. The fieldwork period ran from 15 May to 4 June 2013.

This research was mainly designed and performed as a research work in the Open University of Catalonia (UOC)—the institution where the first and second author works. In the following link you will be able to find the function and aims of the UOC ethics committee: www.uoc.edu/portal/en/recer ca-innovacio/activitat-rdi/comite-etica/funcions/index.html. No permission from a board or committee is necessary in the case of own lecturers social sciences research. In the context of social sciences (and not only in the case of this university) voluntary completion of questionnaires after the goals of the research are explained, is considered as a guarantee that individuals want to participate in the study.

The sample comprised 833 employees working for enterprises operating in Spain. Of the total, 49.8% was male and 50.2% female. The employees' mean age was 34.3 years (SD = 8.83), distributed in the following range: 18–25 years (15.5%), 26–34 years (41.1%), 35–45 years (30.1%), and 46–65 years (13.3%). The participants were highly educated. Most of them had completed at least a bachelor's degree (72.5%). The mean length of work experience was 10.4 years (SD = 8.14) in their workplace and 7.4 years (SD = 7.37) in their enterprises. The employees belonged to enterprises whose activities covered a wide sectoral range, such as financial intermediation, education and social services, health and hospitals, commerce, telecommunications, metallurgy and other similar activities, pharmaceuticals, chemicals, security, sales-oriented services, information, and communication technologies (ICT), general consultancy, hotel industry, distribution, tourism, and food.

#### Measures

We measured dispositional employability by adapting the fivefactor scale created by Fugate and Kinicki (2008). In their study, the first factor "openness to changes at work" (α = 0.76) comprises five items; the second factor "work and career resilience" (α = 0.75) comprises eight items; the third factor "work and career proactivity" (α = 0.55) comprises three items; the fourth factor "career motivation" (α = 0.68) comprises three items, and the fifth factor "work identity" (α = 0.66) comprises six items. The employees answered the questions on a five-point Likert scale, where: 1 = never/none; 2 = sometimes/little; 3 = usually/somewhat; 4 = almost always/quite; and 5 = always/a lot.

In order to obtain the predictors of dispositional employability and online training purchase, we used items for three additional and adapted scales. The first one uses a competence-based approach to employability (Van der Heijde and Van der Heijden, 2006). Five factors of employability are considered: "occupational expertise" (α = 0.90) comprises 15 items; "anticipation and optimization" (α = 0.81) comprises eight items; "personal flexibility" (α = 0.79) comprises eight items, "corporate sense" (α = 0.83) comprises seven items and "balance" (α = 0.78) comprises nine items. The employees answered the questions on a five-point Likert scale, where: 1 = never/none; 2 = sometimes/little; 3 = usually/somewhat; 4 = almost always/quite; and 5 = always/a lot.

The second one is the General Self-Efficacy Scale (Baessler and Schwarcer, 1996). Self-efficacy refers to people's firm belief in their ability to appropriately manage a wide range of obstacles and adverse experiences (Bandura, 1997, 1986). This is a onedimensional construct explained through 10 items (α = 0.81), and measures employees' opinions on a Likert-scale from 1 to 10: 1 = Totally disagree and 10 = Totally agree. The third one is the Hardy Personality Scale (Moreno-Jiménez et al., 2001). The Hardy personality construct (Kobasa, 1979) refers to the personality model of an individual who, when active and committed, perceives stress stimuli as less threatening. The scale is formed by three factors: "commitment" (α = 0.81) comprises eight items; "control" (α = 0.75) comprises six items; and "challenge" (α = 0.81) comprises seven items. The employees answered the questions on a Likert-scale from one to four: 1 = Totally disagree; 2 = Somewhat disagree; 3 = Somewhat agree; 4 = Totally agree.

#### Procedure

The dispositional employability indicator was constructed by following the steps shown in the literature (Muñiz and Bartram, 2007; Muñiz et al., 2013). First, the initial items were translated from English into Spanish by research experts (university lecturers) and language experts belonging to the Language Service at the UOC in Spain. Second, a focus group was held to discuss the translated items (equivalence of meaning, for example). Third, the language experts back-translated the items into English. Fourth and lastly, the equivalence of meaning of the original and adapted versions was checked.

In order to build the exploratory factor analysis structure and the reliability coefficients of the dispositional employability measure, the FACTOR 7.2 program was used (Lorenzo and Ferrando, 2006). This program allows exploratory analysis to be performed using polychoric correlation matrices. It also allows additional analyses—unavailable in SPSS 21.0 to be performed, such as parallel analysis. SPSS 21.0 was used to evaluate the internal consistency of the dispositional employability scale (alpha coefficients). An exploratory factor analysis was therefore performed to analyze the dimensionality of the dispositional employability measure by applying the principal axis extraction and oblimin rotation methods. In addition, polychoric correlation matrices were used; these are particularly suitable for items with a Likert-type response format (Lorenzo and Ferrando, 2009). **Table 1** presents the descriptive statistics of the items obtained from the exploratory factor analysis.

After obtaining the factors of dispositional employability, we performed a probabilistic analysis of discrete choice (binomial logit). We dichotomized the factors and the employability composite indicator by their mean values. We then identified the factors explaining greater propensity toward dispositional employability. In the same way, we explored the factors explaining online training purchase. The interpretation of standardized coefficients determines the probability of independent variables explaining Spanish employees' greater dispositional employability and online training purchase.

In order to estimate the overall effect of individual variables on the explanation of dispositional employability, we applied a binomial logit model. The parameters for the binary logistic regression model are described below. Dispositional employability is the dichotomous dependent variable (DDISEMP). It takes value 1 when the composite indicator of dispositional employability (an arithmetic mean of the factors obtained in the exploratory factor analysis) are equal to or greater than the mean, and value 0 otherwise. This dichotomous indicator has a mean value of 0.5 points and a standard deviation of 0.4 points. Some 52.1% of Spanish employees' dispositional employability is higher (greater than the mean).

Regarding the independent variables, we contemplated a first group of four discrete explanatory variables related to job ambition, in the sense of anticipating and optimizing job changes. Self-initiating proactive job ambition entails employees preparing for future work changes in order to strive for the best possible job and career outcomes. The TRAINING variable relates to vocational education and training and responds to the item: "Time dedicated to improving the knowledge and skills that will be useful at work." The CORRECTING variable relates to the systematic correction of job weaknesses and responds to the item: "I try to correct my weaknesses in a systematic way." The NTRENDING variable relates to interacting with new job-related trends and responds to the item: "Last year, I interacted with the latest developments and trends in my job." The PERSEMPOWR (Personal Empowerment) variable relates to employee responsibility for assessing their job value and responds to the item: "I take responsibility for keeping my value in the job market." These four variables are taken from the employability competence-based approach from Van der Heijde and Van der Heijden (2006).

The analysis also incorporates two additional variables relating to job competences. The first relates to organizational identification, in terms of corporate sense. In new working environments, employees have to participate more as members of an integrated team, identify with corporate goals and accept collective responsibility for the decision-making process. The variable PARTICIPATION relates to employees' participation in corporate sense, and responds to the item: "In my organization, I participated in the formation of a vision with common values and objectives." The second relates to work-life balance. The STRESS variable relates to employees' stress at work and responds to the item: "I suffer job-related stress."

A second group of three discrete explanatory variables is associated with self-efficacy and work-related behavior. Selfefficacy refers to employees' ability to appropriately manage a wide range of obstacles, overcome difficulties and achieve their goals. The PERSISTENCE variable relates to the capacity for persistence and responds to the item: "I find it easy to persist in what I have set up to achieve my goals." The OVERCOME variable relates to the ability to overcome unexpected situations and responds to the item: "Thanks to my skills and resources, I can overcome unexpected situations." And, the CALM variable relates to the ability to keep calm in difficult work situations and responds to the item: "When I'm in trouble, I can stay calm because I have the necessary skills to handle difficult situations." These three variables are taken from the General Self-Efficacy Scale (Baessler and Schwarcer, 1996).

And finally, a third group of four discrete explanatory variables is associated with a hardy personality and work-related behavior. Employees with hardy personalities are characterized

#### TABLE 1 | Dispositional employability items: mean, standard deviation, Skewness, and Kurtosis.


N = 883.

by a high degree of commitment, experiencing situations as challenges rather than as threats, and tend to perceive a certain degree of internal control in various situations (Jiménez et al., 2006). The CONSCIENTIOUSNESS variable relates to preparation for positive job outcomes and responds to the item: "Things go well when you prepare thoroughly." The NEXPERIENCE variable relates to a job choice involving new experiences and responds to the item: "Even if that entails more effort, I choose the work involving a new experience for me." The INNOVATIVENESS variable relates to the preferences for innovative jobs and procedures, and responds to the item: "In my job, I preferably attract innovations and new developments in the proceedings." And, the COMMITMENT variable relates to employees' commitment and responds to the item: "My daily work satisfies me and makes me totally dedicated to it." These four variables are taken from the Hardy Personality Scale (Moreno-Jiménez et al., 2001).

**Table 2** presents the descriptive statistics of the dichotomous dispositional employability explanatory and discrete Likertpoint variables. The analysis of the correlation matrix between explanatory variables suggests the absence of multicollinearity (correlations <0.4 points).

Completing the previous analysis, and in order to estimate the overall effect of individual variables on the explanation of paid online training, we re-applied a binomial logit model. The parameters for the binary logistic regression model are described below. Paid online training is the dichotomous dependent variable (PONLINET). It takes value 1 when employees participate in paid online training activities and value 0 otherwise. This dichotomous indicator has a mean value of 0.52 points and a standard deviation of 0.5 points. Some 52.2% of Spanish employees are in paid online training.

Regarding the independent variables, we contemplated a first group of two explanatory variables related to the sociodemographic conditions. The AGE variable relates to the age of employees. The variable is discrete and takes four values: 18–25 years (value 1), 26–34 years (value 2), 35–45 years (value 3), and 46–65 years (value 4). The GENDER variable relates to the gender of employees. The variable is dichotomous and takes the value 1 for men and value 0 for women.

A second group of two explanatory variables related to dispositional employability factors. The MOTIVATION variable refers to "career motivation and work resilience" factor obtained from the exploratory factor analysis (F2). This factor specifically

#### TABLE 2 | Dichotomous dispositional employability explanatory variables: mean, standard deviation, Skewness, and Kurtosis.


N = 883.

relates to optimism and control over career opportunities, and assigning and planning career goals. The IDENTITY variable refers to the work identity factor also obtained from the exploratory factor analysis (F5). This factor specifically relates to the importance of success and external acknowledgment of work. Both factors have been dichotomized according to their mean values: value 1 when the factors are equal to or greater than the mean; and value 0 when they are less than the mean.

A third group of two discrete explanatory variables related to job ambition (i.e., anticipation and optimization of job changes) from the competence-based approach to employability (Van der Heijde and Van der Heijden, 2006). The TRAINING variable relates to vocational education and training and responds to the item: "Time dedicated to improving the knowledge and skills that will be useful at work." The NTRENDING variable relates to employees interacting with new job-related trends and responds to the item: "Last year, I interacted with the latest developments and trends in my job."

The fourth group of one discrete explanatory variable related to employees' self-efficacy or the ability to achieve their goals and overcome difficulties (Baessler and Schwarcer, 1996). The CALM variable relates to the ability to keep calm in difficult work situations and responds to the item: "When I'm in trouble, I can stay calm because I have the necessary skills to handle difficult situations."

The fifth and final group of four discrete explanatory variables is associated with the hardy personality traits of employees who, when active and committed, react better to stressful stimuli at work (Moreno-Jiménez et al., 2001). The EFFORT variable relates to job effort and responds to the item: "Things are only based on personal effort." The CHALLENGE variable relates to preferences for job challenges and responds to the item: "In my work I attract those tasks and situations involving a personal challenge." The VARIETY variable relates to employees' preferences for job variety and responds to the item: "I like the fact that there is great variety in my work." The NOVELTY variable relates to job choice involving new and different job situations, and responds to the item: "As far as possible, I seek new and different situations in my working environment."

**Table 3** presents the descriptive statistics of the dichotomous online training purchase explanatory and discrete variables. The analysis of the correlation matrix between the explanatory variables suggests the absence of multicollinearity (correlations <0.4 points).

### RESULTS

### Dispositional Employability Indicator

The results of Bartlett's sphericity test (Chi-square = 7110.9; p < 0.01) and the Kaiser-Meyer-Olkin index of sampling adequacy (KMO = 0.898) confirmed the adequacy of the data for factor analysis. The scree test, parallel analysis, and the minimum average partial test yielded a solution based on five factors (Timmerman and Lorenzo, 2011).

After establishing the most suitable factor solution, the oblimin rotation method was used to obtain a simple factor solution. This method of oblique rotation tends to yield the very simplest of solutions, even in cases where one of the items displays a complex structure. The scale was honed down from the 25 original items by removing any that had saturations lower than 0.4, or complex saturations higher than 0.4 in more than one factor. The resulting number of items with the highest saturations was 23, distributed as follows: five items for the "openness to changes at work" factor, six items for the "career motivation

#### TABLE 3 | Online training purchase explanatory variables: mean, standard deviation, Skewness, and Kurtosis.


N = 883.

and work resilience" factor, four items for the "work and career proactivity" factor, five items for the "optimism and engagement at work" factor, and threeitems for the "work identity" factor.

The saturation matrix of the factor solution obtained enabled us to identify the content of the five factors, which together explained 56.4% of cumulative variance. The correlation between the five factors was rather high; it varied between 0.31 and 0.61. The α-reliability for the combined scale was 0.75.

The first factor obtained (F1) refers to "openness to changes at work" (Cronbach's alpha = 0.72). This factor specifically relates to employees' positive feelings, openness, adaptation, management, and acceptance of changes at work. It explains 30.3% of the variance, and has a mean of 0.1 points and a standard deviation of 0.9 points. The second factor (F2) refers to "career motivation and work resilience" (Cronbach's alpha = 0.75), which explains 8.1% of the variance, and has a mean of 0.6 points and a standard deviation of 1.1 points. This factor specifically relates to optimism and control over career opportunities, and assigning and planning career goals. The third factor (F3) refers to "work and career proactivity" (Cronbach's alpha = 0.80). This factor specifically relates to employees' ability to keep abreast (including vocational education and training) of new developments in their jobs, enterprises and industries. It explains 7.2% of the variance, and has a mean of 0.7 points, and a standard deviation of 1.1 points.

The fourth factor (F4) refers to "optimism and engagement at work" (Cronbach's alpha = 0.76), which explains 6% of the variance, and has a mean of 0.1 points and a standard deviation of 0.8 points. This factor specifically relates to employees' positive thinking, attitudes and involvement in work. Finally, the fifth factor (F5) refers to "work identity" (Cronbach's alpha = 0.70). This factor specifically relates to the importance of success and external acknowledgment of work. It explains 4.8% of the variance, and has a mean of 0.9 points and a standard deviation of 1.2 points. **Table 4** shows the factor scores, reliability coefficients, variance explained, mean, standard deviation, and correlations between the five factors obtained from the exploratory factor analysis. Items in the Spanish language are also presented.

### Dispositional Employability Explanatory Factors

The probabilistic model correctly classified 81.3% of employees (79.0% of those who do not have a greater disposition toward employability and 83.5% of those who have a greater disposition toward employability). The Cox-Snell and Nagelkerke R 2 -values were 0.43 and 0.58, respectively. The improvement (from 500.5 to 722) in the likelihood function is significant. It confirms the goodness-of-fit of the predictive capacity, and the variables as a whole have an outstanding explanatory power (Chi-square Hosmer-Lemeshow test = 15.1; p = 0.047).

From the model's estimation (**Table 5**), it was found that all the included variables have significant explanatory power for Spanish employees' greater disposition toward employability (p < 0.1 in the worst case). Regarding standardized coefficients and odds ratios [Exp (β)], the variables with greater explanatory power are employees' career-related personal empowerment, participation in corporate sense, correcting job weaknesses, and commitment to employment. In this sense, it is confirmed that the predisposition toward more employability would be explained differentially by three dimensions linked to employees' behavior to anticipate and optimize their job changes or job ambition, to identify and participate in corporate sense, and by the hardy personality traits linked with the commitment to employment.

The dataset allows our sample to be segmented by employee online training purchase. In this regard, we have segmented our microdata into two samples: employees involved in paid online training activities (N = 461) and otherwise (N = 422). With the

#### TABLE 4 | Dispositional employability indicator: Factor loadings, explained variance, and correlations.


\*p < 0.01. N = 883.

#### TABLE 5 | Determinants of the dispositional employability.


\*\*\*p < 0.01; \*\*p < 0.05; \*p < 0.1; N.S., Not significant.

first sample of employees, we replicated the explanatory model of greater predisposition toward employability. The results are also shown in **Table 5**.

Regarding employees in paid online training, the probabilistic model correctly classified 82.4% of employees (79.3% of those that do not have a greater disposition toward employability and 85.2% of those that have a greater disposition toward employability). The Cox-Snell and Nagelkerke R 2 -values were 0.44 and 0.59, respectively. The improvement (from 637.5 to 869.4) in the likelihood function is significant. It confirms the goodness-of-fit of the predictive capacity, and the variables as a whole have an outstanding explanatory power (Chi-square Hosmer-Lemeshow test = 11.8; p = 0.049).

Comparing standardized coefficients and odds ratios [Exp (β)] between all employees and employees in online training, the second ones attribute greater explanatory power of dispositional employability to career-related personal empowerment, new experiences related to the job and conscientiousness in the preparation for positive job outcomes. In this regard, the effect of online training on predisposition toward employability would be indirect. Being involved in paid online training activities is not a direct explanatory factor of dispositional employability. However, certain types of behavior of employees participating in paid online training would determine a positive effect on greater dispositional employability. This is especially the case for those most sensitive to career-related personal empowerment who are more predisposed to developing new experiences at work and more aware of the fact that positive job outcomes are related to preparation conscientiousness.

### Online Training Purchase Explanatory Factors

The probabilistic model correctly classified 74.2% of employees (69.7% of those that do not purchase online training and 78.5% of those that purchase online training). The Cox-Snell and Nagelkerke R 2 -values were 0.33 and 0.37, respectively. The improvement (from 400.5 to 1105.3) in the likelihood function is significant. It confirms the goodness-of-fit of the predictive capacity, and the variables as a whole have an outstanding explanatory power (Chi-square Hosmer-Lemeshow test = 16.2; p = 0.037).

From the estimation of the model (**Table 6**), it was found that all the variables included have significant explanatory power for Spanish employees' online training purchase (p < 0.1 in the worst case). Regarding standardized coefficients and odds ratios [Exp (β)], it is possible to assess the considerations referred to below.

First, it was confirmed that those most likely to purchase online training are young employees followed by female employees. Second, the relationship between dispositional employability and online training purchase is not direct. The dichotomous composite indicator of dispositional employability is not significant in the model. However, employees who have certain dispositional employability factors tend to do more online training. These particular predispositions are related to motivation, i.e., optimism and control over career opportunities, and assigning and planning career goals; and to work identity, i.e., the importance of success and external acknowledgment of work. And third, it was also found that the employees' jobrelated behavior also determines online training purchase. This behavior is associated with employees' job ambition, self-efficacy, and hardy personality. Specifically, it was found that employees who spend more time on training and have less contact with new trends in their jobs, find it hard to keep calm in difficult situations, and have a greater predisposition toward effort, and preference for novelty, variety, and challenges at work are more likely to purchase online training.

### DISCUSSION

Employability, understood as an individual's ability to get and keep a job, is a synergistic collection of characteristics driven by employees' motivation and behavior. Although prior education and training play a decisive role as predictors of employability, the findings in the literature on career success have tempered this relationship. In this sense, new research highlights the importance of personality characteristics and individuals' behavior as determinants of not only academic performance but also job outcomes. As a result of this evidence, the literature shows that a dispositional employability approach has been developed, which focuses on the perception of individuals' motivational attitudes and behavior related to careers and work in general. Also, the flexibilization, segmentation, and individualization of labor markets, the advent of new work organization and human resources practices, and the global knowledge-based economy have fostered new approaches to employability research. Again, the literature highlights the relevance of motivational attitudes toward new job-related trends.

Online training acquires special relevance in the sense that continued learning is placed at the heart of career development and employees' improved employability. Indeed, and despite its significant potential, empirical evidence has now tempered the online training effects on employability. Little attention to students' specific characteristics and behavior, such as attitude toward careers, seems to explain the weak results of online training for employability. Finally, and beyond their technological and pedagogical dimensions, employees' personal, and motivational determinants of online training enrolment and satisfaction have also been addressed in the literature.

Through a sample of 833 employees working for enterprises in Spain, our research proposed three main objectives: to construct a dispositional employability indicator, to analyze the relationship between online training and dispositional employability, and to study the effect of dispositional employability on online training purchase.

### Dispositional Employability Evidence and Implications

Using principal component exploratory analysis, dispositional employability was found to be characterized by five factors comprising 23 items: "openness to changes at work," "career motivation and work resilience," "work and career proactivity," "optimism and engagement at work," and "work identity." The results obtained for our sample of Spanish employees are consistent with previous research (Fugate and Kinicki, 2008). The indicator presented here has practical value both for managers (especially in human resources practices) and for employees (especially for improving their career development).

Using logit binomial regression analysis, the research obtained the motivational factors explaining dispositional employability. Specifically, it is concluded that employees who are more motivated to anticipate and optimize job changes, who identify with and are involved in corporate sense, who have less stress, who have self-efficacy or the ability to manage obstacles, overcome difficulties and achieve their goals, and who have hardy personality traits to commit, challenge, and control job situations are more likely to have greater dispositional employability. This approach seems to be appropriate for detecting individual

#### TABLE 6 | Determinants of the online training purchase.


\*\*\*p < 0.01; \*\*p < 0.05; \*p < 0.1; N = 883.

differences and to tailor interventions. An assessment based on dispositions can be used to identify personal strengths and weaknesses that should be accounted for in future careers, which may help to develop an individualized coaching trajectory (Vanhercke et al., 2014). Indeed, our evidence suggests that employability motivations could be linked with employability outcomes, but this relationship must be tested in the future.

Furthermore, it seems reasonable to suggest that employees with higher dispositional employability are well suited to empowered, supportive, and developmental management and work practices, due to their proactive and active motivations (Fugate and Kinicki, 2008). In knowledge-based work environments, the importance of employees' future job expectations has been highlighted (Gallie et al., 2012; Graso and Probst, 2012) in terms of explaining employability improvements. But, the link between employee perceptions of control and coping with labor and organizational change must be analyzed in the future.

### Online Training and Dispositional Employability Evidence and Implications

Using logit binomial regression analysis, the research also obtained the relationship between dispositional employability and online training. Consistent with suggestions in the literature (Inayat et al., 2013; Kyndt and Baert, 2013; Silva et al., 2013), for the whole sample of employees we did not obtain any direct relationship. In other words, employees' involvement in online training does not explain per se motivational predisposition toward employability. To observe the effect of online training on dispositional employability, we had to analyze a specific sample of employees in online training, and address their motivations and behaviors. Our results suggest that employees participating in online training who are more sensitive to career-related personal empowerment, more predisposed to developing new experiences at work, and more aware of the fact that positive job outcomes are related to preparation conscientiousness are more likely to have greater dispositional employability.

These results are consistent with the latest evidence on workplace learning outcomes (Virtanen et al., 2014). "Active membership" (i.e., influencing the way things were done at the workplace), "invention" (students' willingness to invent new solutions at work), and "learning orientations" (students' willingness or motivation to learn new things at work) are also determinants in non-online workplace learning environments. Despite this evidence, further investigations need to be conducted in the future. Particular attention should be paid to the relationship between employee-related motivational factors (both in online and non-online training environments), learning design and job-related structural features.

Finally, and using binomial logit analysis, we obtained the socio-demographic, dispositional employability, and motivational factors explaining online training purchase. Employees are more likely to purchase online training if they are younger, women, more motivated, and better identify with their jobs, have more time to spend on training, do not have recent contact with new trends in their jobs, find it hard to keep calm in difficult situations, have a greater predisposition toward effort, and preference for novelty, variety, and challenges at work.

One of the relevant implications of this is that overall dispositional employability does not explain per se online training purchase. We only found significant effects for two specific motivational factors: the first factor relates to optimism and control over career opportunities, and assigning and planning career goals; and the second factor refers to work identity. On the other hand, job ambition (Van der Heijde and Van der Heijden, 2006), self-efficacy (Baessler and Schwarcer, 1996), and hardy personality (Moreno-Jiménez et al., 2001) employee-related motivational effects are also obtained. Self-initiating proactive job ambition entails employees' preparing for future work changes in order to strive for the best possible job and career outcomes. Self-efficacy refers to employees' ability to appropriately manage a wide range of obstacles, overcome difficulties and achieve their goals. Employees with hardy personalities are characterized by a high degree of commitment, experiencing situations as challenges rather than as threats, and tend to perceive certain internal control in various situations. Therefore, we have linked a set of employees' motivational and behavioral factors explaining online training purchase. But, as pointed out in the literature (Brown, 2005; Rovai et al., 2007; Chiu and Wang, 2008; Paechter et al., 2010), motivational factors interrelate with other personal, structural, learning, and social determinants in explaining online training purchase. In this regard, future research on interaction effects should be conducted.

Nevertheless, our contribution is relevant for online training organizations. Despite the rise in online training enrolment, competition in the supply of e-learning programs (formal and informal) is growing exponentially. E-learning organizations increasingly need to take new approaches toward their consumers, and affective loyalty plays a greater role in explaining the rise in market shares. Knowing about the motivations and

#### REFERENCES


behavior of employees (students) that buy online products is a strategic option that should be considered in markets with everincreasing competition. In the future, extension in terms of time and sample size will allow us to further our knowledge of how the motivational and behavioral effects explain online training purchase.

### AUTHOR CONTRIBUTIONS

All the authors made substantial contributions to the design, data analysis, and interpretation of the results. JT contributed to the formulation of the research questions, study design, literature review, interpretation of the results, and article drafting. He is the corresponding author and the guarantor of the article. PF participated in the formulation of the research questions, study design, literature review, data analysis, statistical modeling, interpretation of the results, and article drafting. JB participated in the data analysis and statistical modeling. All the authors have read, revised, and approved the final manuscript.

#### ACKNOWLEDGMENTS

The authors acknowledge the support of the Open University of Catalonia in the English review of the article.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Torrent-Sellens, Ficapal-Cusí and Boada-Grau. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.