All time references are in CEST
European Values Study and World Values Survey: Exploring New Survey Findings and Addressing Methodological Challenges 3
Dr Vera Lomazzi (University of Bergamo, Italy)
Dr Kseniya Kizilova (Institute for Comparative Survey Research, Austria)
Professor Ruud Luijkx (Tilburg University; University of Trento)
|Time||Thursday 20 July, 09:00 - 10:30|
The European Values Study (EVS) and the World Values Survey (WVS) are two large-scale comparative time-series survey research programs studying people’s values, norms and beliefs. Since 1981, these programmes have jointly carried out representative national surveys in over 120 countries and societies containing 92 percent of the world’s population representing an invaluable data source for a global network of scholars and international development agencies, including the World Bank, the UNDP, the WHO, regional development banks etc.
Over the years, the EVS and the WVS have proven the importance of population value study and have demonstrated that people’s beliefs play a key role in economic development, emergence and flourishing of democratic institutions, rise of gender equality, and the extent to which societies have effective government.
We welcome submissions based on EVS/WVS data addressing substantive and/or methodological aspects of value research.
The recently published joint EVS-WVS dataset (2017-2022) and the EVS-WVS trend file (1981-2022) allow social and political sciences to broaden and deepen their analysis. Present session invites papers which make use of the EVS/WVS data -solely or in combination with other types of data- to address a broad scope of issues, including political culture and political attitudes, support for democracy and political participation, perceptions of gender equality and moral values, identity and trust, civil society, corruption, solidarity, and migration among the others.
We also invite papers addressing the projects’ methodological aspects, including challenges and limitations such as reliability and equivalence of employed scales and indicators, non-responses, combining self- and interviewer-administered mode and other. The panel particularly invites papers comparing findings collected via different survey methods in the same countries allowing to estimate the reliability of online surveys and discuss challenges and prospects of their combined use.
Keywords: values; EVS; WVS
Miss Anastasiia Volkova (University of Helsinki) - Presenting Author
This paper investigates the differences in moral values across 28 European countries, using data from the Morally Debatable Behaviors Scale (MDBS) in the European Values Study (EVS). The MDBS measures moral values by asking for justifications for different actions and events, from claiming social benefits to homosexuality and suicide. Results of multi-group confirmatory factor analysis followed by validity tests prove that the MDBS successfully measures the leniency of moral judgments. However, for a deeper understanding, the basis for intergroup analysis in the case of personal-sexual morality may be not countries but other cultural clusters. This paper explores alternative groupings that may explain differences in personal-sexual moral values and similarities between countries, such as religious constructs (based on denominations) and cultural zones (based on historical context, both religious and political). The preliminary findings of this paper support the results of previous research, arguing that while the legal-fair dimension is universal, the personal-sexual dimension depends on cultural context. It is demonstrated how researchers can use the MDBS to analyze and visualize value differences across Europe, drawing on insights from survey methodology and comparative social research.
Mr Manuel Holz (TU Chemnitz) - Presenting Author
Ms Sandra Jaworeck (TU Chemnitz)
Many experimental studies aim to estimate whether a treatment has an effect on a particular population. In this context it is important to consider that heterogeneity can be both present as unobserved differences in individual attributes, as well as a highly varying effect of the treatment on different groups within a population. Yet, many studies rely solely on average treatment effects, which mostly assume homogeneity of the treatment.
International comparisons of migrant health inequalities require special attention to reducing unobserved heterogeneity since migrant and native populations vary across countries on important socioeconomic and cultural characteristics. Choosing the right statistical method to compute health outcome differences is therefore of most importance. There is a growing interest in machine learning approaches when dealing with treatment effects. One of them is the Causal Forest framework, which is well suited to measure treatment heterogeneity. On the other hand, conventional and more frequently applied approaches like Propensity Score Matching minimize unobserved heterogeneity between treatment and control groups by balancing their distribution of covariates. In both methods, conditional treatment effects as a measure of treatment heterogeneity can be retrieved, while depending on different estimation processes.
In this analysis we address the question whether differences in health outcomes between migrant and native populations converge when applying different statistical methods, taking into account the methodological similarities and differences. Using the latest World Values Survey (WVS) wave, we apply both methods to calculate the average treatment effect on the treated (migrants vs. natives) in the self-reported health status (5-point scale).
We conclude our analysis by recommending an integrative approach of both methods that reduces treatment heterogeneity even more than relying on one method.
Ms Andrea Turković (Università degli Studi di Milano Statale) - Presenting Author
European societies are witnessing an ongoing and extensive secularization while, due to immigration, simultaneously becoming increasingly ethnically and religiously diverse. Differentiating between the intrinsic religiosity related to altruistic values and tolerance, and extrinsic religiosity related to in-group identity and negative out-group stereotypisation (Allport & Ross, 1967) study will focus on the influence of religion on inter-group relations and religion as a source of “othering” by locating both within social context of culturalized religion (Astor & Mayrl, 2020) acknowledging both its complexity and transformation.
In our previous work we focused on the influence of religiosity on intolerance towards immigrants and the way this relationship was shaped by the religious context. Consequently, we applied the multilevel regression modelling on the data on 29 countries from the ninth wave of European Social Survey (ESS-9, 2021). Accounting for different types of relationships between religious (non-) practice and adherence, 5 different types of religiosity were yielded, including those who culturally affiliate with religion without practice and those who culturally practice religion by celebrating religious holy days but do not affiliate. We found that, when compared to the non-religious, both the culturally affiliated and culturally practicing were more intolerant towards immigrants, but no mediation with religious context was found except in Orthodox countries. In this paper we wish to replicate the same analyses using the fifth wave of the European Values Study (EVS, 2022) using the same approach to the religiosity and religious context. Aim of this paper is to compare the results of the analyses based on the data from two different surveys to understand whether such typology might better explain the influence of religiosity on intolerance and whether the lack of its mediation with the religious context can be explained by methodological issues or whether religious context actually holds no influence.
Dr Michael Ochsner (FORS) - Presenting Author
Dr Jessica M. E. Herzing (University of Berne)
Mr Alexandre Pollien (FORS)
Dr Michèle Ernst Staehli (FORS)
Given financial pressure, technological advancement, and the impact of the pandemic, many survey infrastructures plan to switch from interviewer-based to self-completion modes. While web surveys have become more feasible also in cross-cultural settings given the rapid advancement of internet penetration and digital literacy in the wake of the recent pandemic, established survey infrastructures face the challenge of comparability between interviewer-based and self-completion surveys because they need to ensure comparability across a time-series that spans several years or even decades.
Using experimental data from the European Value Study (EVS) in Switzerland in 2017, we compare substantial research questions across four designs of push-to-web and a face-to-face survey. We use an encompassing framework to investigate possible representation biases and their effects on results from real-life applications while considering different levels of complexity of analyses users might apply. Preliminary results suggest that face-to-face and push-to-web surveys are generally comparable. However, the likelihood of bias depends on the analysis type applied. Therefore, we formulate recommendations on how users can address such potential differences in their analyses.