Regions, social structure and value orientations: a comparative study of 15 West European countries

In: European political science review: EPSR, Band 1, Heft 3, S. 401-434

Abstract

This article is a comparative study of regional differences in social and political value orientations. I identify four central sets of value orientations: two Old Politics orientations – religious–secular and economic left–right; and two New Politics orientations – libertarian/authoritarian and green values. I use the international value studies from 15 West European countries as my empirical base. The article addresses three major research problems: (1) Which of the four value orientations is most strongly anchored in regional differences, and in which countries do we find the largest value differences between the regions? (2) Do we find the same ranking of the regions across the four value orientations, or do the value orientations group the regions in separate ways? (3) Can the other socio-structural variables explain the impact of region on value orientations, or is that impact unique? The average correlations between regions and each of the four value orientations are similar but somewhat larger for religious value and libertarian/authoritarian values. With regard to the second research problem, I hypothesize one-, two-, and three-dimensional solutions based on different spatial configurations. The one-dimensional configurations implied that there exist some (centre) regions that have secular, economic rightist, green, and libertarian regions, and (peripheral) regions with opposite value orientations. This pattern is clearly found in four countries. The expected two-dimensional solution with an Old Politics and a New Politics dimension was found in four countries, whereas the expected three-dimensional solution with two Old Politics and one New Politics dimension was found in three countries. In the multivariate analysis examining the causal impact of region, only a small portion of the correlation between region and value orientations was spurious when controlling for the other (quasi-)ascriptive variables. Furthermore, only a small portion of the impact of region was transmitted via social class variables.