The Influence of World Societal Forces on Social Tolerance. A Time Comparative Study of Prejudices in 32 Countries

Abstract

Societal variation in xenophobia, homophobia, and other prejudices is frequently explained by the economic background and political history of different countries. This article expands these explanations by considering the influence of world societal factors on individual attitudes. The empirical analysis is based on survey data collected within the World Value Survey and European Values Study framework between 1989 and 2010. Data are combined to a three-wave cross-sectional design including about 130,000 respondents from 32 countries. Results show that xenophobia and homophobia are influenced by the national political history, societal affluence, and the presence of international organizations. Global forces, however, are of particular importance for homophobia.

ACKNOWLEDGMENTS

Thanks to the four anonymous reviewers at The Sociological Quarterly. Special thanks to reviewer 1 who provided comments regarding the data used, and to reviewer 4 regarding the organization of this article.

Work on this research was supported by the Austrian Science Funds. Project J2500.

NOTES

Notes

¹ Explanatory factor analyses with principal component extraction. Eigenvalue greater or equal 1 results in the 2-factor solution of (1) different race and immigrants/foreign workers and (2) individuals who have AIDS and homosexuals in 22 countries. A forced two component solution results in the two named dimensions in 94 out of 96 country–time combinations. Only in Slovakia and Argentina, a xenophobia item loads at both dimensions but only at a single time point. Also, the changes over time within the two items of each factor are parallel in most countries. Substantial divergent trends can be found only from wave 1 to 2 in (a) Turkey, Lithuania, and Finland, where tolerance to individuals of different race increases, while foreign workers receive more hostility and (b) Northern Ireland where tolerance toward homosexuals increases, while individuals with AIDS face more hostility.

² Changes in these affluence and competition indicators were also considered as the ratio of (current level/level in preceding wave). The underlying idea was that the most recent changes are more salient to respondents than the absolute level. The inclusion of these ratio variables, however, did not add any additional insight and results are not reported.

³ This observation was made by reviewer 1.

⁴ There is no indication of multicollinearity problems among the presented models: The strongest bivariate correlation between two macrovariables within the models of Table 3a and 3b is −.45 between GDP and Communist Legacy. Models including GDP and the peak of the authoritarian past, for example, were not estimated, because of a correlation of .7 between the macro-level variables.

⁵ For example, INGOs could have been used instead of IGOs; the explained variances, however, would have been smaller.

⁶ Not shown in the tables. Xenophobia: B of Communist Legacy = 0.32; of GDP (West) = −.017; and of interaction term GDP* Communist legacy = .013.

⁷ Upon recommendation of a reviewer, a membership/no membership variable was considered alternatively for EU membership, because the founding members of the EU are liberal democratic countries and, therefore, could cause a bias in the duration variable. This 0/1 variable, however, yields similar results.

⁸ One can also argue that this is related to poor data quality in the present study, especially for postcommunist countries in the first wave of 1989 to 1994. However, entering interaction terms between communist history and several competition measures does not result in any significant effects for Western democracies, either.

⁹ Interaction terms between communist past and time are not significant for any dimension.

¹⁰ Following the logic of Snijders and Bosker (1999), the following formulae is used 100*(1 − [sum of both variances in final model/sum of both variances in baseline model]).

¹¹ Xenophobia: B of Education = −.035, of IGO = −.287, of interaction term Education*IGO = −.01; Homophobia: B of Education = −.013; of NGO = −.583; and of interaction term Education*NGO = −0.01.