Programme 2013

Tuesday 16th July       Wednesday 17th July       Thursday 18th July       Friday 19th July      

Download the conference book
Download the program

Tuesday 16th July 2013, 11:00 - 12:30, Room: No. 22

European Values Study - methodological and substantive applications 1

Convenor Dr Ruud Luijkx (Tilburg University)
Coordinator 1Professor Wolfgang Jagodzinski (University of Cologne)

Session Details

The European Values Study (EVS) is a unique research project into Europe's basic values. First, it spans a period of almost 30 years with surveys in 1981, 1990, 1999, and 2008. Second, EVS has an extensive geographical coverage. In Europe, the survey has gradually been expanded from mostly Western European countries in 1981 to the whole of Europe in 2008. Third, even though several items have been changed in the consecutive waves, EVS still includes an impressive number of unchanged questions. Fourth, researchers have combined survey data with macro-level data so that multi-level models can be estimated. Finally, and perhaps most importantly, the questionnaires pertain to a very broad spectrum of life domains: family and marriage, economics, work, leisure, politics, religion, morality. This allows to introduce overarching concepts and to examine their effects on attitude and reported behavior in different domains. Such a rich data source also offers a unique chance for substantive and methodological investigations. We are particularly interested in papers which make use of the comparative potential of EVS from a methodological and a substantive perspective. To give a few examples: Do the measurement instruments that have been used in EVS guarantee comparability across time and space? In which domains do we observe a generational change during the last decades and in which domains are life-cycle effects more plausible. Can we reliably estimate the long-term change? Which models are suited best for that purpose? To what extent can differences between countries traced back to cultural influences? How can the latter be measured? What are the main problems of these multi-level models? However, other empirical and methodological topics are possible too.

Paper Details

1. Does measurement stop at borders? A latent class IRT scale analyses of political participation items from EVS 2008
Dr Markus Quandt (GESIS Leibniz-Institute for the Social Sciences)

When the measurement (non-)invariance of attitude or value scales contained in comparative surveys is investigated, the method chosen is almost always a multi-group factor analytic approach. Another commonality of the vast majority of such analyses is the nearly exclusive focus on national borders as the sole demarcation of potential measurement non-invariance between subgroups of the overall samples. The present paper will follow an alternative strategy to identify sample sub-groups for which different measurement models hold, by applying a simultaneous latent class analysis of measurement model parameters (using Item Response Theory instead of Factor Analysis). Compared to the fields of sociology, psychology, and political science, this strategy is more common in organisational and educational research. Substantively, the paper will explore the degree to which the more easily observable national sample distinctions contribute to explaining the diagnosed non-invariance of scales, in comparison to response sets and respondent attributes. A battery of items on unconventional political participation from the European Values Survey 2008 that has been frequently used as base for scaling political behaviour will provide the data for the scale analysis.

2. Contextual determinants of perceived immigration related threat. A cross-cultural comparison in and outside Europe.
Miss Marie-sophie Callens (University of Leuven & CEPS/INSTEAD)

This paper analyzes the determinants of perceived immigration related threat across countries, both European as well as non-European countries. It was hypothesized that people in vulnerable societal positions and people living in countries with large and heterogeneous immigrant groups and bad economic conditions show higher levels of threat feelings.
In order to measure the perceived group level threat a composite measure containing seven items is used. Furthermore, we applied hierarchical structural equation modeling on data from the European Value Study together with data of the Worldbank, resulting in a sample with only native residents of 42 countries.
The outcomes of our preliminary analyses reveal that respondents who hold a vulnerable societal position (low educational level, experience with unemployment, laborers) display more feelings of threat, whereas the economic conditions of the country and the proportion of the immigrant groups appear to have almost no effect on threat perceptions.
Implications of the findings for further research and for policymakers are discussed.

3. Religiosity and Traditional Values in East Asia: A Data Analysis of Multi-National Comparative Surveys
Professor Kazufumi Manabe (Aoyama Gakuin University)

One of the most significant developments in contemporary social sciences is the implementation of large-scale multi-national comparative surveys using questionnaire methods.

The datasets of these kind of multi-national surveys analyzed in this paper are as follows:

1. European Values Study (1999) and World Values Survey (2000-2002)
2. International Social Survey Programme (1998)
3. AsiaBarometer (2003)
4. Asian Barometer (2001-2003)
5. The East Asia Values Survey (2002-2004)

The purpose of data analysis (secondary analysis) in this paper is to examine the relationships between people's religiosity and traditional values. The stepwise method of multi-regression analysis is used for this purpose.

The results of the analyses show that the explanatory power of traditional values in East Asia is weak, which is surprising considering the proposition widely accepted by scholars of qualitative methods that religiosity and general traditional values are strongly related.

Further examinations in the wording of each variable, the method of operationalization of each indicator, and the measurement of the relationships between variables are needed.