Transitioning from interviewer-administered to self-administered survey modes: implications for the quality of political measures

Face-to-face surveys have traditionally been regarded as the “gold standard” in data collection. However, due to declining response rates and soaring costs, they have largely been replaced by self-administered methods. This trend has been further accelerated by the COVID pandemic, which has driven even major scientific surveys to adopt self-completion modes. Using evidence from the Swiss European Social Survey (ESS) and the Measurement and Observation of Social Attitudes in Switzerland (MOSAiCH)—the former intending the transition and the latter having already done it, this paper examines the implications of the transition from face-to-face interviews to self-completion questionnaires for the quality of political measures, which are the most frequently used by scholars. We draw our analyses upon telephone, web, and mail questionnaires of the LIVES-FORS mixed mode experiment and to-face interviews of the ESS 2012. The results do not show large mode differences in favor of face-to-face interviews in terms of political measures. On the contrary, telephone, web and paper modes may offer more advantages in certain situations. To better understand the implications of the intended mode transition in ESS, additional analyses were conducted with MOSAiCH, which has the same quality standards as ESS and switched from a face-to-face survey to a self-survey mode already in 2018. The findings suggest that a shift to self-administered questionnaires could provide political estimates that are of similar quality as those from face-to-face interviews.

1 Introduction

Face-to-face interviews have been considered by many to be a “gold standard” to gain respondent cooperation. Indeed, research has demonstrated that face-to-face interviews achieve higher response rates than other modes (De Leeuw, 1992; Groves and Couper, 2012; Groves et al., 2009; Holbrook et al., 2003; Villar and Fitzgerald, 2017). These surveys rely significantly on the interviewers who gather the data. Interviewers’ responsibilities include reaching out to selected households or individuals, persuading them to participate, and conducting the interviews in a standardized manner. Nowadays, interviewers are also tasked with collecting paradata through contact forms for each attempt or via an interviewer questionnaire to evaluate each interview. These duties have become more complex due to technological advancements and evolving societal factors, such as diverse household compositions, language barriers, and cultural differences. Consequently, the demands on interviewers are rising, necessitating thorough selection, training, and appropriate compensation (Olson et al., 2020).

Face-to-face interviewing is considered the highest quality survey method, but it is also the most expensive, making it rarely feasible for most scientific surveys. As experienced in other country contexts (see Dixon and Tucker, 2010; Schaeffer et al., 2010), survey researchers in Switzerland are increasingly struggling to find skilled interviewers at a reasonable cost. This challenge has been exacerbated since the COVID pandemic, as qualified interviewers appear to be shifting to better-paying, flexible jobs that emerged during this period. While cheaper than face-to-face interviewing, self-administered survey modes (i.e., mail and web questionnaires) have longtime been seen as not offering reliable alternatives for large-scale population surveys (Bowling, 2005). They were seen as unfeasible because of their length (one hour or more), and the modes were seen as producing non comparable measures (Galesic and Bosnjak, 2009; Revilla and Ochoa, 2017; Klausch et al., 2013). In addition, mail and web surveys are less suitable for individuals with literacy difficulties. However, more recent experiences show that at least in some national contexts, where literacy and internet penetration are high, long surveys are feasible and produce good quality in terms of response rates and socio-demographic composition (Luijkx et al., 2021). But research findings remain inconclusive regarding the practical implications of mode effects on measurements, for example when comparing estimates over time after changes in data collection mode (Martin and Lynn, 2011).

As response rates have declined and costs have increased, face-to-face surveys have increasingly been supplanted by self-completion methods (de Leeuw and de Heer, 2002; Dillman, 2017; Murphy et al., 2018; Villar and Fitzgerald, 2017). The COVID pandemic has further accelerated this shift, prompting even major scientific surveys to adopt self-completion modes. What are the implications of these mode changes? Can researchers confidently embrace these new data collection methods, or does the face-to-face survey remain the “gold standard”? The debate is currently lively on this topic and we aim at modestly contributing to it. This paper investigates the potential implications of transitioning from interview-administered to self-administered modes, using evidence from two major Swiss surveys, ESS and MOSAiCH—the former planning the transition and the latter having already completed it. The paper particularly focuses on the implications of changes in data collection mode for the quality of political estimates.

ESS is one of the rarest scientific surveys that has systematically used face-to-face interviewing in Switzerland. However, it is currently challenged by budgetary constraints and various societal changes such as a more active lifestyle of respondents leading to lower contact success rates and higher costs in interviewer-administered survey designs. At the same time, with internet penetration rates increasing rapidly and online data collection modes becoming more widely used, especially following the COVID pandemic, it was decided to switch to self-administered survey modes. Already in ESS Round 10 in 2021, some countries have moved to self-completion modes. In Switzerland, the full transition to self-completion modes (i.e., mail and web) is planned for 2027.

The emphasis of this paper on the impact of mode transition on the quality of political estimates is deliberate and well-founded. ESS is one of the main ongoing international surveys that many political scientists working with survey data resort to. Though the survey covers a wide range of substantive research topics such as ageing, health, family, work-life balance, national identity, economy and social inequalities, politics and democracy is the topic most frequently addressed by ESS-based authors.¹ In Switzerland, ESS is an alternative to the national election study, the Swiss Election Study (Selects), because it is generally less subject to the self-selection of politically interested respondents and has thus far more accurate estimates of political behavior.

The role of survey mode in measuring political attitudes and behavior is controversial. Face-to-face and telephone interviews have long been recognized as the best methods in quantitative research to produce higher response rates and high quality data that provide information about the opinions, attitudes and behaviors of the population (Dillman, 1978; Malhotra and Krosnick, 2007), but some studies have shown that telephone interviews produce lower response rates, greater social desirability response bias—the tendency of survey respondents to make a false or exaggerated self-presentation when asked about socially (un)desirable attitudes and behaviors—and lower quality data than face-to-face interviews (Holbrook et al., 2003). Over the last few decades, self-administered online surveys have been increasingly used in public opinion research because they can be conducted at relatively low costs (Dillman et al., 2014). Some studies have shown that online surveys introduce a bias as their respondents are often younger, better educated and more involved in politics (Faas and Schoen, 2006; Vehovar et al., 2002). Other studies have shown that the data quality of online surveys with respect to political variables is comparable to that of face-to-face surveys (Bytzek and Bieber, 2016) and telephone surveys (Lipps and Pekari, 2016), suggesting that online surveys are useful in political research (Bytzek and Bieber, 2016).

As ESS is one of the main sources of high-quality data for political scientists in Switzerland, the switch to self-administered modes raises questions and concerns about whether there will be a change in the quality of measures. Maintaining the quality of political measures is important because, as noted above, politics and democracy are the topics most frequently covered by ESS-based authors and these variables are subject to bias due to self-selection of politically engaged respondents and overreporting of political participation. In the present study, we summarize findings from mail, CATI, and web data of the LIVES-FORS mixed mode experiment and CAPI data of the ESS 2012. Since these data are somewhat dated and were conducted in a different survey climate than the current one, to better assess the potential consequences of the mode shift in ESS, we also present more recent results from the Measurement and Observation of Social Attitudes in Switzerland (MOSAiCH), which has the same quality standards as ESS and transitioned from a face-to-face survey to a self-administered survey mode in 2018. In sum, this paper investigates the potential effects of a survey mode switch in ESS Switzerland in 2027. The research question is whether the switch from face-to-face to self-completion mode is at risk of lowering the quality of political measures.

The paper is structured as follows. We begin by describing the datasets that we analyzed as well as our measures and methods of analysis. Next, we present comparisons of findings obtained by four different survey modes from the ESS 2012 and the LIVES-FORS mixed-mode experiment. We then present results from MOSAiCH, contrasting outcomes before and after the mode switch in 2018. Finally, we discuss the implications of the findings for future survey research practices.

2 Data and methods

2.1 Data

We used data from the ESS 2012 (Ernst Stähli et al., 2015a) and the LIVES-FORS Mixed mode experiment to (Roberts et al., 2022) to compare the four different survey modes in terms of political measures. Switzerland has taken part in ESS ever since the beginning. The target population is the permanent population aged 15 and over. Sample members were randomly selected from the register list of the Swiss Federal Statistical Office (SFSO). The sample was drawn by proportionally stratifying the seven major regions of Switzerland (NUTS2). The data was collected by the M.I.S. Trend Institute using CAPI as the mode of data collection. The total number of interviews conducted was 1,498, which corresponds to a response rate of 51.7%. One third of the respondents received a cash incentive of 10 CHF, while the remaining two thirds received 30 CHF in cash or donations. The fieldwork was conducted between September 1, 2012, and April 22, 2013. ESS 2012 covers the topics of media and social trust; politics; subjective well-being, social exclusion, religion, and national identity; and socio-demographics as core modules; personal and social well-being; understanding and evaluating democracy as rotating modules; and human values as additional modules.

In 2012, around the similar time as ESS 2012, the Swiss Centre of Expertise in Life Course Research (LIVES) and the Swiss Centre of Expertise in the Social Sciences (FORS) designed an experiment to investigate which survey modes work best in the Swiss context in order to maximize the quality of future quantitative research and find the best combination of modes with respect to response rates, biases, sample, budget and timing (Roberts et al., 2016). In line with the survey design of ESS, the target population for the LIVES-FORS mixed-mode experiment was also permanent residences aged 15 and over. However, the scale of the study was limited to the French-speaking regions of Switzerland (i.e., Suisse romande) due to budget constraints. Sample members living in the French-speaking Switzerland were randomly drawn from the register list of the Swiss Federal Statistical Office (SFSO). The data of the experiment was collected by the M.I.S. Trend Institute between November 22, 2012, and April 15, 2013. The study covers the subjects of social trust, politics and participation, perception of social conflicts; satisfaction with life, health, happiness, personal and social well-being, social contacts and support, self-deception, life events; socio-demographics, religion, nationality, language, discrimination, education, work, household, income; and survey climate and social desirability.

Prior to the first contact, sample members were randomly assigned to three experimental groups. One group was invited to participate in a telephone survey, another group in a paper survey and the third group in a web survey. However, in subsequent contacts, respondents who did not respond in the assigned mode were also given the option to participate in another mode. Sample members selected for the CATI group could participate via CATI, paper, or self-administered paper non-response questionnaire. Sample members selected for the paper group could only participate via paper or self-administered paper non-response questionnaire. Sample members selected for the web group could participate in many different ways; web, paper, CATI/CAPI or self-administered paper non-response questionnaire. The total number of interviews conducted in the CATI group was 421, of which 364 were conducted through CATI, corresponding to a response rate of 60.7%. The total number of interviews conducted in the paper group was 351 with a response rate of 65.4%. The total number of interviews in the web group was 714, of which 457 were conducted by a web questionnaire, amounting to a response rate of 44.5%. The high response rate for CATI and low response rate for web may be due to the survey climate in Switzerland at the time. From the early 1980s until the mid-2010s, high-quality surveys in Switzerland were conducted by telephone and people were therefore used to this (Ernst Stähli, 2012). Although the level of internet penetration was high, web surveys were not very common at the time, which could be one explanation for the lower response rate on the web.

2.2 Measures and the analytical approach

In our analysis, we used political measures sourced from both the ESS 2012 dataset and the LIVES-FORS mixed-mode experiment data. These measures include interest in politics, participation in political activities such as signing petitions and boycotting certain products within the last 12 months. However, it is worth noting a discrepancy in the measurement of voting behavior between the two datasets. In the ESS 2012, respondents were asked if they had voted in the last national elections, whereas in the LIVES-FORS data, they were queried about voting in any election or popular vote within the past 12 months. Due to this disparity, making direct comparisons between the measures is not feasible. Therefore, in our analysis of voting behavior, we focused solely on comparing different mode groups within the LIVES-FORS mixed-mode experiment data.

As our study focuses on assessing changes in the quality of political measures, including voting behavior, including voting behavior, we filtered our dataset before conducting analyses. We achieved this by selecting subsets from both the ESS 2012 and the LIVES-FORS mixed-mode experiment data, based on citizenship and age, thus concentrating our analyses on Swiss citizens aged 18 years and older. Subsequently, we refined our sample within the LIVES-FORS mixed-mode experiment data to encompass solely respondents whose assigned survey mode aligned with their actual completed survey mode. It is important to note a slight selection bias in the collected data in the LIVES-FORS mixed-mode experiment, stemming from differing probabilities of selection between individuals with and without phone numbers. To address this, the dataset incorporates a set of weights to balance the size difference between these groups. In our analysis, we applied design weights specific to the paper and web groups.

3 Results

Figure 1 presents the distribution of respondents across four distinct survey modes concerning measures of political interest, voting in an election or a popular vote in the last 12 months, boycotting certain products in the last 12 months, and signing a petition in the last 12 months. Regarding political interest, the proportion of politically interested respondents is higher in the CAPI mode compared to the CATI, web, and paper modes, respectively. However, the differences between various survey modes do not exhibit statistical significance. The results indicate that over half of each sample demonstrates an interest in politics. With respect to voting behavior, a substantial portion of the sample across three survey modes reports participation in elections or popular votes. The CATI group exhibits the highest proportion of participation compared to the web and paper groups, respectively. In terms of boycotting behavior, it is evident that more than half of the sample across all survey mode groups did not boycott certain products in the last 12 months. Notably, the proportion of respondents engaging in boycotting is highest among the web group, followed by the CAPI, paper and CATI groups, respectively. The difference between the CATI group and the other modes is statistically significant, p < 0.05. Finally, regarding petition signing, the behavior is most prevalent among respondents in the CAPI and web groups, followed by the paper and CATI groups. The CATI and paper groups significantly include fewer respondents exhibiting petition signing behavior compared to the CAPI and web groups. The differences among different survey modes, whether significant or insignificant, may stem from different factors such as survey mode characteristics, sample composition and response bias. Different survey modes appeal to distinct demographic cohorts. The composition of respondents within each survey mode could differ based on demographic factors, such as age, education, or socioeconomic status. For instance, CATI surveys often draw higher participation from older individuals who are more at ease with conventional communication methods, whereas web surveys tend to attract younger, digitally adept demographics. Consequently, the demographic makeup of the sample may differ depending on the chosen survey mode and this may influence respondents’ likelihood of engaging in certain behaviors like political interest, voting, boycotting, or petition signing.

Figure 1

Figure 1. Political measures by survey mode. Sources: LIVES-FORS Mixed-mode experiment data (CATI, Web, Paper), ESS 2012 (CAPI); design-weighted results.

Table 1 presents the sample composition across four distinct survey modes categorized by various sociodemographic characteristics, including age, gender, education, income, migration background, and urbanicity. Expectedly, there are variations in the age distribution across survey modes compared to CAPI, with web and paper modes showing the most pronounced differences. Self-administered survey mode groups include more younger adults and less older adults compared to interviewer-administered modes. With respect to the gender distribution across different survey modes, the paper mode has the highest percentage of female respondents, followed by CATI, web, and CAPI. Overall, while there are slight variations, the differences in gender distribution between CAPI and the other survey modes are relatively small. As for the education distribution across survey modes, web mode has a significantly lower proportion of respondents with compulsory education and a slightly higher proportion with secondary education compared to CAPI. Variations in the income distribution across survey modes are rather minor. The comparison between four survey modes in terms of migration background reveals some slight differences. Paper mode has a slightly higher proportion of natives and slightly lower proportions of Swiss nationals originally from neighboring countries and Swiss nationals from non-neighboring countries. Web mode has lower proportions of natives and higher proportions of Swiss nationals from neighboring and non-neighboring countries compared to CAPI. Finally, regarding urbanicity, web mode has a lower proportion of respondents from rural municipalities/isolated towns compared to CAPI, a similar proportion from other municipalities, and a slightly lower proportion from centers. Overall, CATI and paper modes exhibit minor differences in urbanicity distribution compared to CAPI, while web mode shows more pronounced differences, particularly in rural municipalities/isolated towns and centers.

Table 1

Table 1. Sample composition by survey mode, n—1,684.

Following the differences in sample composition across various survey modes, we proceeded to perform a set of binary logistic regression analyses. These analyses accounted for respondents’ sociodemographic characteristics while assessing political interest, voting, boycotting, and petition signing. As indicated in Table 2, when accounting for sample composition, there are minimal disparities across survey modes, with the exception of CATI, which notably includes a higher proportion of respondents less inclined to boycott or sign petitions. This observation might stem from the likelihood that CATI respondents are more engaged in conventional communication channels, and they may exhibit a greater inclination toward conventional forms of political participation, such as voting, as opposed to non-conventional forms like boycotting and petition signing (Martin and Lynn, 2011). The findings suggest that individual factors such as age, gender, education level, migration background, and urbanicity play significant roles in shaping political interest and behavior. Consequently, the impact of survey mode appears to be indirect, as varying modes attract distinct respondent demographics with differing tendencies toward political interest and behavior. Ultimately, it is the individual characteristics that emerge as the primary influencing factors.

Table 2

Table 2. Binary logistic regression analyses of political measures by survey mode and sociodemographic characteristics of respondents.

3.1 Lessons from MOSAiCH

In addition to the results of the LIVES-FORS experiment, the analysis of the mode change in MOSAiCH, a survey conducted by the same team with the same quality standards as ESS and which has recently undergone a mode switch, can also provide instructive lessons for understanding what a future mode transition in the ESS might or might not entail. MOSAiCH is a cross-sectional survey that investigates the Swiss population’s values and attitudes towards a range of social issues ever since 2005, though selected parts of the survey had previously been administered in Switzerland under different surveys. The content of the survey is designed in a way to enable comparisons over time and across countries and the thematic focus is the current module of the International Social Survey Programme (ISSP). Ever since 2018, the MOSAiCH questionnaire is composed of two parts: a main survey with the current module of ISSP together with selected socio-demographic questions, and one or two follow-up surveys that include Switzerland-specific questions. The survey was conducted every 2 years as a face-to-face interview until 2017. As of 2018, the mode of data collection has been changed to a self-administered survey (online/paper) and the survey has started to be administered annually (see Supplementary material for more details). On average, MOSAiCH has a lower response rate, albeit insignificantly, than ESS. MOSAiCH experiences fluctuating response rates annually, with some years exceeding 50% while others dip below 40%. Notably, response rates were lower during the mode transition in 2018 and 2019. Nonetheless, in subsequent years, rates rebounded, reaching again as high as 50%.

Figure 2 provides an overview of respondents’ political interest levels and voting behavior across successive rounds of the MOSAiCH survey spanning the period from 2005 to 2022 (Ernst Stähli et al., 2012, 2014, 2015b, 2019a, 2019b, 2020, 2021, 2022, 2023; Joye et al., 2007, 2008, 2010). The data reveals a consistent average level of political interest among participants, hovering around 60% throughout most rounds. However, an exceptional deviation from this norm is observed during the 2007 round, characterized by a significant downturn in reported political interest. This departure from the expected trend could be attributed, in part, to the specific thematic focus of the 2007 round, which centered on sports—a relatively quotidian subject matter likely to draw individuals with comparatively lower political inclinations into the respondent pool. Moreover, subsequent rounds conducted after the mode switch in MOSAiCH reveal a slight decline in the proportion of respondents indicating an interest in politics. While various confounding factors, such as the survey context in the aftermath of the COVID pandemic and the rotational topics featured in these rounds, may contribute to this modest decrease, these findings provide reassurance that transitioning from interviewer-administered to self-administered modes does not necessarily result in greater inclusion of politically interested individuals in the survey pool. In regard to voter turnout, across all rounds, self-reported voter turnout consistently exceeds 70%, whereas the actual turnout rates fall short of 50%. This results in a significant turnout gap, reaching as high as 32.7% in the 2022 round. These findings indicate a substantial overestimation of voter turnout by MOSAiCH. There exists an increase in reported turnout rates in the rounds following the mode transition, albeit not substantially.

Figure 2

Figure 2. The share of politically interested respondents and voters in MOSAiCH 2005–2022. The question of political interest was not asked in the rounds of 2018 and 2019, n = 596–3,476. The questions on voter turnout in MOSAiCH 2013 and 2015 refer to the 2011 federal elections; the questions in MOSAiCH 2017, 2018 and 2019 refer to the 2015 federal elections; and the questions in MOSAiCH 2020, 2021 and 2022 refer to the 2019 federal elections, where the actual turnout rates are given in the graph, n = 621–3,455.

4 Discussion

Interviewer-administered face-to-face and telephone surveys have long been considered the best methods for collecting data, but have faced budgetary challenges and societal changes over time. In recent decades, online surveys have become more widely used as internet penetration has increased. The COVID-19 pandemic has triggered the further expansion of the use of self-administered survey modes and the transition of major surveys from interviewer-administered to self-administered survey modes. ESS has been one of the surveys subject to a shift from face-to-face interviews to self-administered modes.

This study investigated the potential effects of a survey mode switch in ESS Switzerland in 2027 on the quality of some political measures, the indicators most used by ESS-based authors and subject to nonresponse bias and measurement error.

Our findings suggest that transitioning to self-administered questionnaires in the ESS could yield political estimates that are comparable in quality to CAPI data, and in some cases, even more advantageous. Our results cannot be used to determine whether any observed differences between the experimental modes (i.e., mail, CATI, and web) and the CAPI mode in ESS are due to the mode of data collection or the context in which the survey was conducted. It is even more difficult to attribute the observed differences between the different rounds of MOSAiCH to the survey mode alone, as it is confounded by other design features. However, the results still give us a descriptive indication of the low difference between the four modes with respect to the measurement of political interest and political behavior in ESS 2012 and the irrelevant quality impact of the mode change in MOSAiCH on the same measurements.

While the data we rely on are somewhat dated and based on the survey setting in Switzerland, where online surveys were not the mainstream mode at that time and thus their sample is biased towards younger and more educated individuals, the findings do not show large survey mode differences in favor of face-to-face surveys. On the contrary, the use of online and paper survey modes may be more advantageous to attract young individuals, women, low-educated individuals, individuals with a migration background from non-neighboring countries, and individuals living in rural areas who are more likely to be politically disengaged. However, since the sociodemographic segments addressed by online and paper surveys are slightly different, using these modes as a mixed mode rather than using them alone may provide better results in terms of sample representativeness and quality of political measures.

Our research therefore represents a small contribution to the thesis that switching from face-to-face mode to web-paper is possible without a significant loss of measurement quality, at least with respect to political interest and political behavior considered and in the context of a country with high Internet penetration.

Data availability statement

Publicly available datasets were analyzed in this study. This data can be found here: https://www.swissubase.ch/fr/catalogue/studies/11517/14976/overview, https://www.swissubase.ch/fr/catalogue/studies/12370/18307/overview and https://www.swissubase.ch/fr/catalogue/studies/20135/18873/overview.

Author contributions

NA: Formal analysis, Writing – original draft, Writing – review & editing. ME: Writing – original draft, Writing – review & editing.

Funding

The author(s) declare that no financial support was received for the research and/or publication of this article.

Conflict of interest

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.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpos.2025.1425888/full#supplementary-material

Footnotes

1. ^https://www.europeansocialsurvey.org/sites/default/files/2023-12/ESS_Annual_Report_2022-23_0.pdf

References

Bowling, A. (2005). Mode of questionnaire administration can have serious effects on data quality. J. Public Health 27, 281–291. doi: 10.1093/pubmed/fdi031

PubMed Abstract | Crossref Full Text | Google Scholar

Bytzek, E., and Bieber, I. E. (2016). Does survey mode matter for studying electoral behaviour? Evidence from the 2009 German longitudinal election study. Elect. Stud. 43, 41–51. doi: 10.1016/j.electstud.2016.04.007

Crossref Full Text | Google Scholar

De Leeuw, E. D., and de Heer, W. (2002). “Trends in household survey nonresponse: a longitudinal and international comparison” in Survey Nonresponse, eds. R.M. Groves, D.A. Dillman, J.L. Eltinge, and R.J.A. Little (New York: Wiley), 41, 41–54.

Google Scholar

De Leeuw, E. D. (1992). Data quality in mail, telephone and face to face surveys. Plantage Daklaan 40, 1018CN Amsterdam: TT Publikaties.

Google Scholar

Dillman, D. A. (1978). Mail and telephone surveys: The total design method, vol. 19. New York: Wiley.

Google Scholar

Dillman, D. A. (2017). The promise and challenge of pushing respondents to the web in mixed-mode surveys. Surv. Methodol. 43, 3–31.

Google Scholar

Dillman, D. A., Smyth, J. D., and Christian, L. M. (2014). Internet, phone, mail, and mixed-mode surveys: The tailored design method. Hoboken: John Wiley & Sons.

Google Scholar

Dixon, J., and Tucker, C. (2010). “Survey nonresponse”. In Handbook of Survey Research (2nd ed.), eds P. V. Marsden and J. D. Wright (Bingley: Emerald), 593–630.

Google Scholar

Ernst Stähli, M. (2012). “Spotlight on Switzerland” in Telephone surveys in Europe: research and practice. eds. M. Häder, S. Häder, and M. Kühne (Berlin Heidelberg: Springer), 25–36.

Google Scholar

Ernst Stähli, M., Joye, D., Pollien, A., and Sapin, M. (2012): MOSAiCH: Enquête sur la santé, and lenvironnement et la politique - 2011 [Dataset]. Centre de compétences suisse en sciences sociales - FORS, Lausanne. Distributed by FORS, Lausanne. doi: 10.23662/FORS-DS-587-2

Crossref Full Text | Google Scholar

Ernst Stähli, M., Joye, D., Pollien, A., Sapin, M., Dümmler, K., Kessler, D., et al. (2015a). European Social Survey, Switzerland - 2012 [Dataset]. FORS - Centre de compétences suisse en sciences sociales, Lausanne. Distributed by FORS, Lausanne. doi: 10.23662/FORS-DS-648-3

Crossref Full Text | Google Scholar

Ernst Stähli, M., Joye, D., Sapin, M., Pollien, A., Dümmler, K., and Kessler, D. (2014): MOSAiCH: Enquête sur la famille et le changement des rôles homme-femme, andlidentité nationale et la citoyenneté - 2013 [Dataset]. Centre de compétences suisse en sciences sociales - FORS, Lausanne. Distributed by FORS, Lausanne. doi: 10.23662/FORS-DS-645-2

Crossref Full Text | Google Scholar

Ernst Stähli, M., Joye, D., Sapin, M., Pollien, A., Ochsner, M., Nisple, K., et al. (2015b): MOSAiCH: Befragung zur Staatsbürgerschaft und zum Sinn der Arbeit – 2015 [Dataset]. FORS - Schweizer Kompetenzzentrum Sozialwissenschaften, Lausanne. Distributed by FORS, Lausanne. doi: 10.23662/FORS-DS-683-3

Crossref Full Text | Google Scholar

Ernst Stähli, M., Sapin, M., Pollien, A., Ochsner, M., and Nisple, K. (2023). MOSAiCH 2022. Measurement and Observation of Social Attitudes in Switzerland. Study on Family and Changing Gender Roles and related topics (1.0.0) [Dataset]. FORS - Swiss Centre of Expertise in the Social Sciences. Distributed by FORS, Lausanne. doi: 10.48573/6ezs-eg91

Crossref Full Text | Google Scholar

Ernst Stähli, M., Sapin, M., Pollien, A., Ochsner, M., and Nisple, K. (2022). MOSAiCH 2021. Measurement and Observation of Social Attitudes in Switzerland. Study on Health and Health Care and related topics [Dataset]. FORS – Swiss Centre of Expertise in the Social Sciences. Distributed by SWISSUbase, Lausanne. doi: 10.48573/t659-e039

Crossref Full Text | Google Scholar

Ernst Stähli, M., Sapin, M., Pollien, A., Ochsner, M., and Nisple, K. (2021): MOSAiCH 2020. Measurement and Observation of Social Attitudes in Switzerland. Study on Environment and related topics [Dataset]. FORS - Swiss Centre of Expertise in the Social Sciences. Distributed by FORS, Lausanne. doi: 10.23662/FORS-DS-1232-1

Crossref Full Text | Google Scholar

Ernst Stähli, M., Sapin, M., Pollien, A., Ochsner, M., and Nisple, K. (2020): MOSAiCH 2019. Measurement and Observation of Social Attitudes in Switzerland. Study on Social Inequality and related topics [Dataset]. FORS - Swiss Centre of Expertise in the Social Sciences. Distributed by FORS, Lausanne. doi: 10.23662/FORS-DS-971-1

Crossref Full Text | Google Scholar

Ernst Stähli, M., Sapin, M., Pollien, A., Ochsner, M., Nisple, K., and Joye, D. (2019b). MOSAiCH 2018. Measurement and Observation of Social Attitudes in Switzerland: Study on Religion and related topics [Dataset]. Distributed by FORS, Lausanne. doi: 10.23662/FORS-DS-962-2

Crossref Full Text | Google Scholar

Ernst Stähli, M., Sapin, M., Pollien, A., Ochsner, M., Nisple, K., and Joye, D. (2019a). MOSAiCH 2017. Messung und Observation von sozialen Aspekten in der Schweiz: Studie zu sozialen Netzwerken und zur Rolle des Staates [Dataset]. Distributed by FORS, Lausanne, 2018. doi: 10.23662/FORS-DS-925-1

Crossref Full Text | Google Scholar

Faas, T., and Schoen, H. (2006). Putting a questionnaire on the web is not enough-a comparison of online and offline surveys conducted in the context of the German federal election 2002. J. Off. Stat. 22:177.

Google Scholar

Galesic, M., and Bosnjak, M. (2009). Effects of questionnaire length on participation and indicators of response quality in a web survey. Public Opin. Q. 73, 349–360. doi: 10.1093/poq/nfp031

Crossref Full Text | Google Scholar

Groves, R. M., and Couper, M. P. (2012). Nonresponse in household interview surveys. New York: John Wiley & Sons.

Google Scholar

Groves, R. M., Fowler, F. J. Jr., Couper, M. M., Lepkowski, J. M., Singer, E., and Tourangeau, R. (2009). Survey methodology, New York: John Wiley & Sons.. John Wiley & Sons. 561.

Google Scholar

Holbrook, A. L., Green, M. C., and Krosnick, J. A. (2003). Telephone versus face-to-face interviewing of national probability samples with long questionnaires: comparisons of respondent satisficing and social desirability response bias. Public Opin. Q. 67, 79–125. doi: 10.1086/346010

Crossref Full Text | Google Scholar

Joye, D., Schöbi, N., Leuenberger-Zanetta, S., Pollien, A., and Sapin, M. (2010): MOSAiCH: Enquête sur la religion, les inégalités sociales et la citoyenneté - 2009 [Dataset]. Schweizer Kompetenzzentrum Sozialwissenschaften - FORS, Lausanne. Distributed by FORS, Lausanne. doi: 10.23662/FORS-DS-559-2Joye,D

Crossref Full Text | Google Scholar

Joye, D., Schöbi, N., and Kaenel, C. (2007) MOSAiCH: Erhebung zu allgemeinen politischen Fragen, sozialen Netzwerken, zur Bürgerschaft und Arbeit - 2005 [Dataset]. Schweizerischer Informations- und Daten-Archivdienst für die Sozialwissenschaften - SIDOS, Neuchâtel. Distributed by FORS, Lausanne. doi: 10.23662/FORS-DS-462-3

Crossref Full Text | Google Scholar

Joye, D., Schöbi, N., Leuenberger-Zanetta, S., and Pollien, A. (2008): MOSAiCH: Enquête sur les loisirs et le sport, le rôle du gouvernement, la politique suisse et les relations avec l’Europe - 2007 [Dataset]. Schweizer Kompetenzzentrum Sozialwissenschaften - FORS, Lausanne. Distributed by FORS, Lausanne. doi: 10.23662/FORS-DS-530-2Joye,D

Crossref Full Text | Google Scholar

Klausch, T., Hox, J. J., and Schouten, B. (2013). Measurement effects of survey mode on the equivalence of attitudinal rating scale questions. Sociol. Methods Res. 42, 227–263. doi: 10.1177/0049124113500480

Crossref Full Text | Google Scholar

Lipps, O., and Pekari, N. (2016). Sample representation and substantive outcomes using web with and without incentives compared to telephone in an election survey. J. Off. Stat. 32, 165–186. doi: 10.1515/jos-2016-0008

Crossref Full Text | Google Scholar

Luijkx, R., Jónsdóttir, G. A., Gummer, T., Ernst Stähli, M., Frederiksen, M., et al. (2021). The European values study 2017: on the way to the future using mixed-modes. Eur. Sociol. Rev. 37, 330–346. doi: 10.1093/esr/jcaa049

Crossref Full Text | Google Scholar

Malhotra, N., and Krosnick, J. A. (2007). The effect of survey mode and sampling on inferences about political attitudes and behavior: comparing the 2000 and 2004 ANES to internet surveys with nonprobability samples. Polit. Anal. 15, 286–323. doi: 10.1093/pan/mpm003

Crossref Full Text | Google Scholar

Martin, Peter, and Lynn, Peter (2011). The effects of mixed mode survey designs on simple and complex analyses. ISER working paper series, no. 2011–28.

Google Scholar

Murphy, J., Biemer, P., and Berry, C. (2018). Transitioning a survey to self-administration using adaptive, responsive, and tailored (ART) design principles and data visualization. J. Off. Stat. 34, 625–648. doi: 10.2478/jos-2018-0030

Crossref Full Text | Google Scholar

Olson, K., Smyth, J. D., Dykema, J., Holbrook, A. L., Kreuter, F., and West, B. T. (2020). Interviewer effects from a total survey error perspective: New York: CRC Press.

Google Scholar

Revilla, M., and Ochoa, C. (2017). Ideal and maximum length for a web survey. Int. J. Mark. Res. 59, 557–565. doi: 10.2501/IJMR-2017-039

Crossref Full Text | Google Scholar

Roberts, Caroline, Joye, Dominique, and Michèle, Ernst Stähli (2016). “Mixing modes of data collection in Swiss social surveys: methodological report of the LIVES-FORS mixed mode experiment.” In LIVES working papers (Vol. 48, pp. 1–42).

Google Scholar

Roberts, C., Ernst Stähli, M., Joye, D., and Sanchez Tome, R. (2022). Well-being and Unease in French-speaking Switzerland. The LIVES-FORS Mixed Mode Experiment 2012 (restricted access) [Dataset]. FORS - Swiss Centre of Expertise in the Social Sciences, University of Lausanne. Distributed by SWISSUbase, Lausanne, 2022. doi: 10.48573/yx0e-qs55

Crossref Full Text | Google Scholar

Schaeffer, N. C., Dykema, J., and Maynard, D. W. (2010). “Interviewers and Interviewing” in Handbook of Survey Research, 2nd ed., eds. Marsden P. V., Wright J. D. (Bingley: Emerald), 437–471.

Google Scholar

Vehovar, V., Batagelj, Z., Lozar Manfreda, K., and Zaletel, M.. (2002). “Nonresponse in web surveys” in Survey nonresponse, eds. R.M. Groves, D.A. Dillman, J.L. Eltinge, and R.J.A. Little (New York: Wiley), 229–242.

Google Scholar

Villar, A., and Fitzgerald, R. (2017). “Using mixed modes in survey research: evidence from six experiments in the ESS” in Values and identities in Europe. (London: Routledge), 299–336.

Google Scholar