PhD Project

Despite the extensive literature on values and welfare attitudes, the empirical analysis of the relationship between both concepts falls short systematic reviews. Most research captures only one side of the equation; employs umbrella concepts; or assumes a simplified relationship in order to explain welfare attitudes. The thesis challenges these perspectives from a theoretical and an empirical angle by explicitly restricting the attention to value attitude links.

The cross-disciplinary theoretical reflection on applications of the attitude concept in psychology, political science and sociology revealed similarities, differences and, in a historical account, the appearance of the recent divisions in the field. A main conclusion from the theoretical considerations was that multidisciplinary approaches are particularly fruitful in attitudinal research as intra-disciplinary research agendas often neglect the complexity of attitude formation. In the synthesis merge the theoretical reflections into a micro-macro framework containing alternate micro predictors such as political trust, political rationality and perceived material vulnerability. The macro context was set up to test for cross-sectional variation with two hypotheses in mind. Firstly, some authors postulate a deterministic understanding of the value attitude relationship depending on democratic development. Hence a rather homogeneous region like the EU should show great similarities across countries. Secondly, the clustering of countries along the popular welfare regime argument was tested.

Combined path diagram of three models explaining attitudes towards gender equality. Blue frame: BM19; red frame: POAC19; green frame: UNBE19. For matter of simplicity, error
terms are represented by ingoing arrows, covariances are in light gray and variances are not specified. Reprinted from “Values and Political Attitudes - a Systematic and Empirical Review of a Simple (?) Association”, by A.Seymer, 2013, p. 121

A methodically rigorous review was chosen to assess the theoretical framework empirically. Hence, an array of ten different value attitude models was defined in order to systematically assess similarities and differences in value attitude links. All relevant micro-level predictors including value, attitudes and control variables were captured by multidimensional latent constructs delivering improved measurement quality and cross-sectional comparability. Finally, the comparison of 19 EU member states referred to the two macro-context orientated hypotheses. The analysis was carried out with data from the fourth wave of the European Social Survey applying Multi-Group Structural Equation Modeling (MGSEM). A multivariate regression analysis was carried out accounting for the cross-sectional invariant measurement models of values and attitudes. In a second step, the value-attitude links were related to contextual variables controlling for alternate macro predictors.

The results indicate clearly that values, political trust and political rationality are the key predictors of political attitudes. Despite this clear result in line with previous research, the results provide little support for the hypothesized relationships. Social structural determinants are largely irrelevant in the different models and the macro context resembles a heterogeneous landscape of value attitude links across countries. The existence of multiple clusters contradicts claims of universal associations of values and political attitudes, while the clustering itself follows no known or theoretically derivable pattern.

The research bears important implications for future research. Firstly, values need to be considered more thoroughly in the context of attitudinal research, as they are amongst the few consistent predictors in all models. Secondly, the lack of explanatory power found in mainstream macro variables indicates the importance of considering alternate macro predictors such as elite discourses. Finally, alternative considerations of social structure that tend towards a milieu approach have probably the most promising potential for further insights in value-attitude links.

Download the PhD Thesis

Alexander Seymer
Social Scientist and Data Analyst