Food Reputation and Food Preferences: Application of the Food Reputation Map (FRM) in Italy, USA, and China

Given the food challenges that society is facing, we draw upon recent developments in the study of how food reputation affects food preferences and food choices, providing here a starting standard point for measuring every aspect of food reputation in different cultural contexts across the world. Specifically, while previous attempts focused either on specific aspects of food or on measures of food features validated in one language only, the present research validates the Food Reputation Map (FRM) in Italian, English and Chinese over 2,250 participants worldwide. Here we successfully measure food reputation across 23 specific indicators, further grouped into six synthetic indicators of food reputation. Critically, results show that: (a) the specific measurement tool of food reputation can vary across cultural contexts, and that (b) people's reputation of food products or categories changes significantly across different cultural contexts. Therefore, in order to understand people's food preferences and consumption, it is important to take into account the repertoire of cultural differences that underlies the contexts of analysis: the three context-specific versions of the FRM presented here effectively deal with this issue and provide reliable context-specific insights on stakeholders' interests, perspectives, attitudes and behaviors related to food perceptions, assessment, and consumption, which can be effectively leveraged to foster food sustainability.


INTRODUCTION
The worldwide crisis regarding food and obesity poses a series of challenges to individuals, scholars, practitioners, policy makers, and society in general (Hawkes et al., 2015). For example, to face obesity, which is predicted to affect 51% of the population by 2030, interventions are required that can generate improvements at a systemic level (Finkelstein et al., 2012). Among other possibilities, a social-psychological approach aimed at understanding why and how consumers make certain food choices could be a promising tactic to initiate such interventions. Based on collectively shared judgements about a given entity (Moscovici, 1988), the concept of reputation may shed light on a series of socialpsychological processes that individuals use in their transactions with food. Because reputation is a concept that affects and orients knowledge, trust, attitude, and choices toward a specific object, it can be considered both a social process and the product of such a process (Bonaiuto et al., 2017). Reputation is defined as the "the distribution of opinions about a person or other entity, in a stakeholder or interest group" (Bromley, 2001, p. 154), and, critically, it can be used to refer to any entity, such as individuals (Emler, 1990), groups (Bromley, 2001), organizations (Riel and van Fombrun, 2007), and even food (Bonaiuto et al., 2012b(Bonaiuto et al., , 2017. Past research has shown that perceived characteristics of food (related to food reputation), rather than objective ones (related to information), might exert a stronger effect on consumer choices (e.g., Carfora et al., 2019). Therefore, studying food reputation can potentially provide promising insights to implement systemic interventions that would help tackling the aforementioned food and obesity crisis, among other issues. Furthermore, the conceptualization of food as a social agent implies that food should be considered as a place-specific agent: food is basic, and as such it regards biology (physiological experience), psychology (individual experience), and culture (social experience) (Rozin, 2007).
According to this logic, the general contribution of the present work is 2-fold. First, drawing upon previous results (Bonaiuto et al., 2012b(Bonaiuto et al., , 2017, it further refines the existing scale for measuring food reputation (the Food Reputation Map; Bonaiuto et al., 2017), helping the work in the understanding of food reputation. Second, in the current research FRM is validated in three languages-Italian (its original version), English, and Chinese, across their three respective cultural settingscontributing to the understanding of food as a social agent within cross-cultural lenses (Rozin, 2007). The present work could potentially inform future research on both individual and sociocultural factors that drive food consumption and consequently stimulate discussion, applications and interventions toward a more sustainable and ethical food consumption (Vermeir and Verbeke, 2006).

Defining Food Reputation
The rationale behind food reputation lies in the conceptualization of food as a social agent, since its features integrate both individual and collective processes (Conner and Armitage, 2002). Such a wide and open perspective, appropriate to understand food reputation, defines food features along three main areas (Bonaiuto et al., 2012b): (1) features linked to food (physical-chemical features and nutritional content); (2) features linked to the environment (economical, social and cultural features); and (3) features linked to the effects on the individual (sensorial, physiological and psychological effects). As such, food reputation can be considered both the social process of creating shared meaning around the concept of food and the result of such a process (Bonaiuto et al., 2017). Therefore, food reputation is defined as the whole set of beliefs (representations, attitudes, direct and indirect knowledge, etc.) that individuals hold about food. It includes beliefs about its antecedents and consequences (i.e., its production and its effects), its present features, its overall attractiveness (based on past direct and indirect experiences), and the future expectations related to its usage and consumption (Bonaiuto et al., 2012b(Bonaiuto et al., , 2017. Based on this definition, an integrative model has been developed to operationalize food reputation's various facets. Established via a series of studies (qualitative, quantitative, and experimental) which indeed employed an international pool of experts (e.g., focus group) and participants (e.g., RCT experiment; see Bonaiuto et al., 2012aBonaiuto et al., ,b, 2017, the Food Reputation Map (henceforth, FRM) integrates the intrinsic characteristics of food, its effects on the environment, and its effects on the individual encompassing six main areas (secondorder factors), namely "Synthetic Indicators" of food reputation: Essence, Cultural Effects, Economical Effects, Environmental Effects, Physical Effects, and Psychological Effects. These areas are further articulated into 23 specific areas (first-order factors), namely "Specific Indicators" of food reputation: Composition, Genuineness, Life time, Recognition, Territorial identity, Tradition, Familiarity, Innovativeness, Context, Price, Preparation, Social and environmental responsibility, Traceability, Proximity, Safety, Ability to satisfy, Digestibility, Lightness, Organoleptic perception, Personal memories, Psycho-physical well-being, Conviviality, Group belongingness.
On the basis of a series of different studies encompassing different methodological approaches-qualitative and quantitative, correlational and experimental (Bonaiuto et al., 2012a(Bonaiuto et al., ,b, 2017, these areas represent all the known possible features of food reputation.

Food Reputation in Different Cultures
It is no secret that food and its processes vary dramatically across cultures (Rozin, 2007). The simple fact that cultural anthropology is the central discipline in the field of food and culture (Counihan and Van Esterik, 2013) exemplifies the relevance and complexity of this topic. Furthermore, considering that in the past 20 years, an enormous amount of research, scientific publications, books, websites, policies, and applied interventions have been dedicated to the social and cultural aspects of food and its consumption (Counihan and Van Esterik, 2013), it appears to be a matter of fact that any consideration related to the concept and processes of reputation cannot overlook the cultural differences that would affect it. Similarly, because of food's social agent nature (Bonaiuto et al., 2017), research on food, the individuals choosing it, food choice environments and food related processes cannot overlook the changing processes that food continuously undergoes (Devine, 2005).
Based on this reasoning, a fundamental question therefore inquires whether food reputation, as defined by the FRM model, would be different across different cultural settings. To test this, and to further validate the FRM model internationally, a series of three studies in three different cultural settings have been conducted.

The Present Research
Drawing upon the FRM model (Bonaiuto et al., 2017), the present research validates the FRM questionnaire in three different cultural contexts. The resulting three validated versions of the FRM, namely FRM-ITA, FRM-ENG, and FRM-CHI (in Italian, English, and Chinese, respectively) are reported in Appendix 1. Three studies are presented. They were conducted between 2013 and 2016, in Italy first, and then in the U.S. and in China. In the three studies, the FRM has been tested with reference to the same three different goods-vegetables, peeled tomatoes and citrus fruit-chosen because they represented the three major food categories within the Italian economy (Castiglione et al., 2007;Zaccarini Bonelli, 2012) and to keep a comparable set of products.

Aims and Hypotheses
The aim of the three reported studies is to validate the measurement model for each language-specific version of the FRM initially presented by Bonaiuto et al. (2017), through a Confirmatory Factor Analysis performed via Structural Equation Modeling (SEM). Analyzing the factorial structure of the FRM using a CFA allows for testing both convergent and discriminant construct validity (Bagozzi et al., 1991;Corral-Verdugo, 2002). According to the presupposition that individuals' perceptions of food are culturally defined (Conner and Armitage, 2002;Rozin, 2007), we assume that the specific facets that operationalize food reputation (i.e., the most appropriate item-markers measuring the Specific Indicators) can differ across different cultures.

Operative Hypotheses
To verify such assumption, the operative hypotheses of Study 1, Study 2 and Study 3 concerned the adequacy of: H1. The model fit indexes for each Synthetic Indicator of Food Reputation, which were modeled on the basis of both the initial theoretical constructs that generated the Specific Indicators of Food Reputation (Bonaiuto et al., 2012b(Bonaiuto et al., , 2017 and the results of the principal component analyses in present and previous data (Bonaiuto et al., 2017); H2. The correlations among Specific Indicators of Food Reputation (i.e., latent variables), in terms of size/statistical significance. H3. The lambda coefficients connecting items (observed variables) and Specific Indicators of Food Reputation (latent variables), in terms of size/statistical significance.

Data Analysis
Reverse items were recoded such that higher scores always mean a positive reputation content. Then, data were analyzed through Confirmatory Factor Analysis via Structural Equation Modeling, to provide evidence for convergent and discriminant construct validity (Fornara et al., 2010). All analyses were conducted using the software STATA-14.
Across the three studies, the structural models performed to conduct the various CFAs validate the FRM-ITA, FRM-ENG, and FRM-CHI for each of the six Synthetic Indicators of food reputation. Following the approach of Hu and Bentler (1999) and Schreiber et al. (2006), three indexes to examine the model's goodness of fit are used here: the RMSEA, the SRMR, and the CFI, respectively with cut off values of 0.08, 0.08 (0.06 for close fit), and 0.95. Following previous research (Fornara et al., 2010), the RMSEA value was prioritized when deciding whether to accept the model. For each Synthetic Indicator of food reputation, a step-by-step iterative procedure was followed to modify the initial solution, including all items loading only on the expected factor (according to results of Bonaiuto et al., 2017). Both conceptual criteria (i.e., the retained sets of items reflected high content validity; see Fornara et al., 2010) and statistical criteria (i.e., statistical confirmation was provided by modification indexes analysis; Chou and Bentler, 1990) led to the emergent factorial solutions presented in the results section of each study. To ensure that the SEMs were identified, a constraint was added to the first indicator for each latent variable.

Aims and Hypotheses
Study 1 applies the aims and hypotheses reported in section Aims and Hypotheses in the Italian context and therefore concerns the validation of the Italian version of the FRM, henceforth FRM-ITA.

Participants and Procedure
The paper-and-pencil survey was administrated to a total of about 1,500 participants, from June to December 2013. Using a stratified sampling procedure, participants were recruited in public areas across Italy where they were individually asked to fill-in a 10-15-min survey about food for research purposes. Randomly, each participant was assigned to one of the three possible conditions (either vegetables, peeled tomatoes, or citrus fruit, as target food object). After a preliminary data screening (incomplete survey, response set, missing data), a finalized sample of N = 1,337 was used for data analysis. The finalized sample was evenly distributed across conditions (vegetables: 32.7%; peeled tomatoes: 34.3%; citrus fruit: 33%) and composed of: women 60.4%; Italians from North (26.4%), Center (31.4%), South (34.2%) and Major Islands (9.8%) of Italy; average age: 38.3 years (SD: 14.6); lower education (9.8%), high school (53.9%), university degree (31.2%), post-graduate education (2.2%). Importantly, participants' Body Mass Index (underweight: 4.5%; healthy weight: 66.7%; overweight: 24.4%; obese: 4.4%) was similar to a comparable sample of Italians (ISTAT, 2019), meaning that our sample is not biased about a food-related relevant index.

Measures
The FRM-ITA survey consists of 102 items measured on seven-point Likert-type scales (from "Completely disagree" to "Completely agree"); it was administrated in Italian. More specifically, the FRM-ITA tool includes 10 General Food Reputation items (Bonaiuto et al., 2017), the first of which focuses on the general reputation of the evaluated product, and the remaining on product (the target object) and process (how it is created) reputation. The FRM-ITA also includes 92 items devoted to measure the 23 Specific Indicators (originally containing 4items each) of food reputation on separate scales. In the food reputation section, participants rated each item according to the following instruction (where, for each given questionnaire, X was substituted by the specific food item label to be assessed with FRM-ITA): "For each of the following statements, please indicate to what extent the reported characteristic describes X. It is enough to express your own opinion, on the basis of what you know about X, or according to whatever you have read, seen and heard about it." This instruction aims to trigger the reputation framework judgment of a target entity based on an individual's both direct and indirect experience of it. The final section of the self-report questionnaire included six items that assessed gender, age, education, area of origin in Italy, height, and weight.

Results and Discussion
Following the FRM model, results of the first-order CFAs show the best measurement models (item-markers) for each Specific Indicator of food reputation: fit indices for each model measured by the FRM-ITA (H1), covariances among Specific Indicators for each Synthetic Indicator of food reputation (H2), and lambda coefficients for the retained items (H3) are reported in Tables 1,  4

Cultural Effects
For the scale measuring Cultural Effects, the model includes four correlated Specific Indicators of Food Reputation: Territorial Identity (three items), Tradition (three items), Familiarity (three items), and Innovativeness (four items).

Economic Effects
For the scale measuring Economic Effects, the model includes three correlated Specific Indicators of food reputation: Context (three items), Price (three items), and Preparation (four items). The factors Context and Price were not significantly correlated.

Environmental Effects
For the scale measuring Environmental Effects, the model includes four correlated Specific Indicators of food reputation: Social and Environmental Responsibility (three items), Traceability (three items), Proximity (four items), and Safety (three items). The Social and Environmental Responsibility and the Proximity factors were not significantly correlated.

Physiological Effects
For the scale measuring Physiological Effects, the model includes three correlated Specific Indicators of food reputation: Ability to Satisfy (three items), Digestibility (four items), and Lightness (three items).

Psychological Effects
For the scale measuring Psychological Effects, the model includes five correlated Specific Indicators of food reputation: Organoleptic Perception (four items), Personal Memories (three items), Psycho-physical Well-being (four items), Conviviality (three items), and Group Belongingness (four items). The following correlations among factors were not statistically significant: Personal Memories-Conviviality, Psycho-physical Well-being-Conviviality, and Conviviality-Group Belongingness.

Conclusion
Study 1 results, based on an extensive correlational survey conducted across Italy, overall confirm the measurement model theorized by the original FRM (Bonaiuto et al., 2017): the tested models produced good fit indexes (H1); first order factors (i.e., the Specific Indicators of Food Reputation) were all correlated to each other (H2) within the following Synthetic Indicators of Food Reputation: Cultural Effects and Physiological Effects; and, a total of 76 items (from the initial total of 92 items) were retained (H3). In the finalized FRM-ITA, one item was removed from each of the following Specific Indicators of Food Reputation (the relevant overarching group's Synthetic Indicator of Food reputation is indicated in brackets): Composition, Genuineness, Life time, Recognition (Essence); Territorial identity, Tradition, Familiarity (Cultural Effects); Context, Price (Economic Effects); Social and Environmental responsibility, Traceability, Safety (Environmental Effects); Ability to satisfy, Lightness (Physiological Effects); Personal memories, Conviviality (Psychological Effects). The finalized FRM-ITA is given in Appendix 1. In conclusion, in Study 1 the FRM-ITA results successfully achieved. Hence, the English version of the instrument is targeted.

Aims and Hypotheses
Study 2 applies the aims and hypotheses reported in section Aims and Hypotheses in the U.S. context, and therefore is concerned with the validation of the English version of the FRM, henceforth FRM-ENG. It originates from the FRM-ITA, which was translated and then back-translated by a team of English and Italian native speaker scholars (including some of the authors).

Measures
The FRM-ENG survey consists of the same 102 items used in the FRM-ITA (see section Method for details). The latter has been translated in English and back-translated in Italian. All items were measured on seven-point Likert-type items (from "Completely disagree" to "Completely agree"). The final section of the questionnaire included eight items measuring gender, age, education, marital status, employment, nationality, height and weight.

Results and Discussion
Following the FRM model, results of the first-order CFAs show the best measurement models (item-markers) for each Specific Indicator of food reputation: fit indices for each model measured by the FRM-ENG (H1), covariances among Specific Indicators for each Synthetic Indicator of food reputation (H2), and lambda coefficients for the retained items (H3) are reported in Tables 2, 10-15, 22, respectively.

Essence
For the scale measuring Essence, the model includes four correlated Specific Indicators of food reputation (factors), each measured by four items: Composition, Genuineness, Life Time, and Recognition.

Cultural Effects
For the scale measuring Cultural Effects, the model includes four correlated Specific Indicators of food reputation, each measured     STUDY 3: FRM-CHI

Aims and Hypotheses
Study 2 applies the aims and hypotheses reported in section Aims and Hypotheses in the U.S. context, and therefore is concerned with the validation of the Chinese-Mandarin version of the FRM, henceforth FRM-CHI. It originates from the FRM-ENG, which was translated and then back-translated by a team of English and Chinese-Mandarin native speakers scholars (including some of the authors).

Participants and Procedure
The online survey was administrated to about 350 participants, during May-August 2015. Participants were recruited at Zhejiang University (Hangzhou, China) via email using an available mailing list of students; to further populate the sample, participants were also recruited in the streets around the University and surveys were administered either via a mobile device or filled out in paper-and-pencil form. Respondents volunteered to participate in a 10-15-min survey about food for research purposes. Similar to Study 1 and 2, each participant

Measures
The FRM-CHI survey consists of the same 102 items used in the FRM-ENG (see section Method for details). The latter was translated in Chinese and then back-translated in English by a team of experienced researchers who were native in one language and fluent at the professional level in the other one (supervised by some of the co-authors). All items were measured  Cultural Effects 5. 6. 7. 8.  on seven-point Likert-type items (from "Completely disagree" to "Completely agree").

Results and Discussion
As per Study 1-2, results of the first-order CFAs show the best measurement models (item-markers) for each Specific Indicator of food reputation: fit indices for each model measured by the FRM-CHI (H1), covariances among Specific Indicators for each Synthetic Indicator of food reputation (H2), and lambda coefficients for the retained items (H3) are reported in Tables 3,  16

Cultural Effects
For the scale measuring Cultural Effects, the model includes four correlated Specific Indicators of food reputation: Territorial         Identity (three items), Tradition (three items), Familiarity (four items), and Innovativeness (four items).

Auxiliary Analysis
To further corroborate the importance of defining the three versions of the FRM, we conducted a post-hoc auxiliary analysis where we compared the Specific Indicators of food reputation describing the three aggregated product categories in the three countries. A series of 23 ANOVAs has been conducted to test for significant differences in each Specific Indicator of food reputation (of the three food categories aggregated) across the three cultural contexts. Specifically, in Figure 1, we show the statistical differences (one-way ANOVA and 95% confidence interval of the mean) among the three countries for each specific indicator of food reputation. Results show that some reputational features of the examined food products are indeed perceived differently across cultures (all p < 0.05): for example, Italians perceive Composition, Life Time, Familiarity, Social and Environmental Responsibility, Lightness and Psycho-Physical Well-being to be significantly lower than Americans and Chinese; Americans perceive Genuineness, Recognition, Familiarity, Price, Social and Environmental Responsibility, Traceability, Safety, Digestibility and Lightness to be significantly higher than Italians and Chinese; Chinese perceive Preparation and Ability to Satisfy to be significantly lower, while Innovativeness significantly higher than Italians and Americans. Furthermore, as an example, we suggest the visual representation of the FRM in the form of a Kiviat graph (Morris, 1974)-or radar chart (Figure 2)-to have a visual representation of the magnitude of each given product category's reputational profile.
In conclusion, although the test of such statistical differences goes beyond the aim of the present research, we believe that such representation is a useful example to highlight one of the possible applications of the FRM. Results show that people's reputation of the food categories here examined changes significantly across different cultural contexts. These results, together with the fact that consumers increasingly demand local and traditional food (Pieniak et al., 2009), substantiate the idea that in order to understand people's actions in relation to food preferences and therefore food consumption, it is important to take into account the repertoire of cultural differences that underlies the contexts of analysis. In order to do so, the culture-specific measures of food-reputation here provided can therefore be very effective and useful tools to acquire such knowledge. In turn, the knowledge of culture-specific food perception could be indeed leveraged to promote a more sustainable consumption of food.

GENERAL DISCUSSION
The three reported studies succeeded in validating the original model proposed for the FRM (Bonaiuto et al., 2017). Drawing on the idea that food can play different roles in different cultures, and therefore the concept of food reputation might change across different cultural settings, it is important to provide available tools capable of measuring food related constructs cross-culturally. The three versions of the Food Reputation Map, namely the FRM-ITA, FRM-ENG, and FRM-CHI in Italian, American English, and Mandarin Chinese, respectively, were created and validated by administration of the same FRM item sets in Italy, USA and China (Table 1). Together, the three studies represent a first attempt at creating a series of tools that could be applied in future studies for improving the understanding of individuals' food perceptions, assessments, and consumption. Overall, the three studies confirm the measurement structure of the FRM through the verification of the three operational hypotheses of the present research: in fact, model fit indexes, correlations among Specific Indicators of food reputation, and lambda coefficient-only with few exceptions in each context-were satisfactory and confirmed the original theoretical model of food reputation already presented in past research (Bonaiuto et al., 2012b(Bonaiuto et al., ,c, 2017. The Food Reputation Map theoretical model, encompassing twenty-three Specific Indicators, which can be further grouped into six Synthetic Indicators of food reputation, was replicated in the three different cultural contexts, keeping constant three target food categories. The results held across three diverse samples of Italian, American, and Chinese respondents. According to the assumptions that (a) food is a social agent (Bonaiuto et al., 2017), (b) it is fundamentally linked to specific cultural settings (Counihan and Van Esterik, 2013), and (c) is subject to continuous change (Devine, 2005), the three measures of food reputation presented here can be applied to study food reputation according to, and within, different cultural settings. The finalized sets of items measure each of the twenty-three Specific Indicators of food reputation either by a three-item or four-item marker pool: such items, in each cultural setting, apparently have a different weight in measuring the specific indicator of food reputation. Also, results show that correlations among Specific Indicators of food reputation-which, according to the FRM model cluster into a Synthetic Indicator of food reputation-can change across cultural settings. These results support the idea that food reputation can be measured by a standard set of items and can be synthesized by the same set of Indicators (Specific and Synthetic ones). This tenet does not exclude the possibility that specific target food categories can also be defined and rated differently according to different cultural settings in terms of their respective food reputation: further research should investigate which parameters of food reputation are, on the one hand, context specific and which are, on the other hand, generalizable across cultures.

Limitations and Future Directions
The new and promising tools developed in this research should be considered in light of some limitations, which can guide future research developments for understanding both generalizable and context-specific features of food reputation. First, it should be noted that the new FRM measures developed here emerged from    Thus, future research should aim at replicating the FRM model within larger, representative sample within each context, possibly including various other socio-demographic information, such as ethnicity and immigration status to allow for a deeper understanding of how reputational features of a given food could be perceived differently within specific groups. Second, the FRM tools developed in this research have referred to three different goods, namely vegetables, citrus fruit and peeled tomatoes, all very relevant goods for the Italian food market (Castiglione et al., 2007;Zaccarini Bonelli, 2012;Bonaiuto et al., 2017). However, these goods might not be central within other food markets in other cultural contexts, and therefore future research should consider how food reputation features are linked to products whose importance is more or less central within various specific cultural settings.
Also, as already pointed out in previous research (Bonaiuto et al., 2012b(Bonaiuto et al., , 2017, the FRM model originated and has been developed from a set of initial researches (carried out with both qualitative and quantitative methodology) based in Italy (Bonaiuto et al., 2012a(Bonaiuto et al., ,b,c, 2017. Such a feature on the one hand is a strength considering the variety of approaches and methods used in its development; on the other hand, it could potentially represent a limiting factor in terms of cultural diversity. In fact, it is well-acknowledged that food is a fundamental aspect in the Italian culture (Parasecoli, 2004), and therefore it might be the case that some outcomes which emerged in the Italian sample could be culture-specific rather than cultural universals. Thus, a test of the FRM in other cultures is needed to generalize the validity of the FRM across different cultural and linguistic contexts, possibly considering classes of products, which are very relevant for those specific cultures, to assure the best benchmarking approach. Overall, the present research is a first attempt to set a standard measure, which can be used to assess this issue; however, future research should investigate whether other fundamental tenets of food reputation can arise in different contexts.
Furthermore, concerning factorial the structure of the FRM model, the second-order factors of food reputation (namely, the Synthetic Indicators) have not been discussed here-a goal that would have been out of the scope of the present manuscript. Rather, here we test which items of each Specific Indicator of food reputation (first-order factor) are indeed the most appropriate markers to measure the intended Specific Indicator within each culture. Future research should therefore confirm the secondorder overall structure of the model in different cultural contexts. In addition, concerning the comparison between cultures, we assumed and demonstrated that the best item-markers for each Specific Indicator of food reputation can vary cross-culturally. However, one un-answered question is whether or not (and, if yes, which) facets of food reputation could be universally relevant: starting from the present results, future research should therefore test the multi-group invariance of the FRM model across different cultural groups.

Practical Implications
In spite of such limitations, the present tools set an effective and useful standard of measures, which can be implemented in various practical activities. A series of possible applications can be considered in light of these new tools, which are highlighted here in view of future developments.
At the consumer level, one possible application lies in the opportunity to gather new knowledge on different reputational features, or perceived features that can be linked to reputation (Péneau et al., 2006), which may affect consumer choices in different contexts (Bonaiuto et al., 2012b). This goal could be achieved, for example, by investigating a given product's reputation in two different cultural contexts to understand its culture-specific reputational features' strengths and weaknesses. Such a strategy could be used to address various issues. For example, it could help individuals' decision-making on how to self-regulate eating behaviors (Johnson et al., 2012); or it could shed light on how ethnic identity, socialization and other culture-specific behaviors affect food consumption (Xu et al., 2004); or, in more general terms, it could deepen our understanding of cultural specific influences of food reputation on the attitude-behavior consistency (Crano and Prislin, 2011) related to food choices.
From a marketing perspective, because in both physical and online markets peer-to-peer knowledge represents a fundamental asset for consumers and businesses to derive information related to reputation and trust (Ert et al., 2016), the knowledge and management of a given food product's reputational features can obviously be an important asset. Drawing upon the evidence that different food reputational features have different impacts on consumers' food choices (Bonaiuto et al., 2012b,c), food reputation management could be a very effective strategy for various stakeholders (e.g., businesses or consumers) to gain strategic advantage from their competitors. On the one hand, businesses could, for example, grow their own reputation (Riel and van Fombrun, 2007) by taking advantage of their products' best reputational features, or they could improve their own products' reputation by investing in specific weak reputational features to be addressed. On the other hand, specific clusters of consumers, such as athletes or clinical patients, could acquire knowledge about specific reputational features of a given product and then use this knowledge to their advantage (for example, for improving performance or for sticking to prescribed nutritional programs; e.g., Johnson et al., 2012).
At the broader perspective, policy makers, opinion-leaders and various institutional stakeholders can potentially use the present tools to promote well-being at the community level. In fact, the various reputational features of the food reputation model, confirmed here, could be studied to serve purposes related to, among others, community-based health interventions (Schulz et al., 2005;Ball et al., 2010;Brand et al., 2014), environmental sustainability (Tilman and Clark, 2014), and price control (UN World Food Programme, 2012). Indeed, one of the major strengths of our research lies in the fact that by considering all possible facets of food reputation (defined by the FRM), and by developing a culture-specific instrument measuring such facets (e.g., FRM-ENG), it would be possible to understand whether a specific reputational feature (e.g., Tradition) could be leveraged to promote for example health or sustainable consumption at the community level in a given culture. We could draw upon the example of the Specific Indicator "Tradition": although it has been often argued that there is no real cuisine tradition in the U.S. (e.g., Mintz, 2002), results of our Auxiliary analysis show no significant differences in Tradition across the three different cultures. This specific result could indicate that, despite that perhaps Italy and China have (at least historically) greater food traditions than the U.S., this is not reflected in the perceptions of individuals. Findings like this can inform marketers, practitioners, and policymakers alike to engage in more informed actions and solutions toward healthier, more sustainable and better informed food-related decisions.
Within this approach, the study of food reputational features could indeed be leveraged to promote food sustainability in a variety of ways. For example, provided that information about sustainability are communicated to consumers, consumption behaviors might have a major role in bringing about more sustainable food production (Grunert, 2011). Past research has already shown that some perceived features of food, such as food safety, environmental concern, nutritive value, taste, freshness and appearance-that are, arguably, very much comparable to the reputational features by the FRM, might influence organic food consumer preferences (Shafie and Rennie, 2012). However, how individuals can be encouraged to cut unsustainable consumption behavior (e.g., excessive meat consumption) has been underexplored, and more in-depth studies on the factors that could increase people's willingness to engage in a more sustainable food consumption are much needed (Hartmann and Siegrist, 2017). We argue that, ideally, the FRM model could be used to develop international evidence-based knowledge, which in turn could inform and support international exchange of information and effective policy design on drivers of sustainable consumer behavior and evaluations across countries worldwide (McGeevor, 2009).

Conclusion
From a social-psychological perspective, understanding processes driving individuals' food preferences and food consumption (i.e., food consumer behavior), food markets, and political decision-making is an important asset to be developed. The fundamental importance of understanding the cultural specificity of food reputation (Parasecoli, 2004), is reflected in the assumption that human behavior can only exists in a given place, and therefore it is both the product of, and it produces, a whole series of transactions between individuals and the specific environments where their behavior occurs (Proshansky et al., 1970;Bronfenbrenner, 1977;Bonnes and Secchiaroli, 1995;Bonnes and Bonaiuto, 2002;Devine-Wright, 2013). The global issues and challenges related to food consumption must be faced by interdisciplinary efforts and tackled by multiple perspectives (FAO, 2009). On the one hand, a stunning 113 million people across 53 countries are suffering acute hunger (Global Report on Food Crises, 2019); on the other hand, trend forecasts suggest that by 2030, 51% of the population will be obese (Finkelstein et al., 2012). In this respect, the crucial importance of food reputation lies in both its theoretical and applied implications for understanding food consumption choices and behaviors. By drawing upon the present research program, and by further developing the measures provided, the detailed and specific knowledge capital that could be derived provides the initial building blocks for a number of new possible interventions and action plans designed to tackle the current global foodrelated challenges, and potentially be leveraged to foster food sustainability. From a behavioral science perspective, whether in the realm of the consumption, production, marketing, political, or clinical intervention over food and drink matters, a "think global, act local" approach (Devine-Wright, 2013) could materially facilitate the development of international sustainable solutions to some of the global challenges related to food.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
SD and MB designed the research questions and the study. SD, MB, FF, WC, and JM supervised the data collection. SD performed the statistical analysis to discuss with MB and FF for interpretation and drafted the manuscript. WC, FF, and MB provided the feedback on the manuscript. SD, FB, UG, and IP contributed to the data collection. All authors contributed to the article and approved the submitted version.

FUNDING
This research was carried out by developing tools and procedures created within previous funding received from Nestlé Italiana S.p.A. (Grant Food Reputation Map: Sviluppo per il Progetto Axía 2013), awarded to the last author as a first development of an initial grant within the same programme. Its contents are the responsibility of this article's authors and do not necessarily represent the position of the funding body and initiative. Special thanks to Manuela Kron (Corporate Affairs Director at Nestlé Italiana S.p.A.) and Alessandra Medolago Albani (Partner and CEO at Cantiere di Comunicazione-TakeGroup) for their support, feedback, and discussion across several meetings. A further partial support for the research presented here was made possible thanks to Sapienza Università di Roma (Grant Food Reputation Map: sviluppo dello strumento, Year 2014-prot. C26A1482K4) awarded to the last author. Moreover, both incoming and outgoing staff international mobility and collaborations were funded by Sapienza Università di Roma with both China (Grant Accordi Interuniversitari di Collaborazione Culturale e Scientifica Internazionali con Zhejiang University, China, Year 2014-Prot. AI2614MWLJ) and USA (Grant Accordi Interuniversitari di Collaborazione Culturale e Scientifica Internazionali con Claremont Graduate University, USA, Year 2015-Prot. AI2615KLJA): those two mobility grants for staff-and their corresponding graduate students mobility grants which funded 3-months fellowships abroad to Italian master students gathering data in USA and in China-were awarded to the last author.