Conferences

Due to globalisation and immigration, the ethnic diversity of European countries increases. Is-sues of multiculturalism and migration policies are high on the political agenda, fuelled by negative immigrant stances among the general public. In this presentation we focus on the public opinion regarding immigrants and ethnic minorities in European countries. That is, we are out to compare countries on their mean level of anti-immigrant attitudes, without explain-ing why there are differences.

The research field of ethnic relations and anti-immigrant attitudes offers a relevant and inter-esting application for methods of cross-cultural research. Even more so than in many other re-search fields, the question of cross-cultural equivalence of anti-immigrant measures is press-ing here. Inter-group relations and individual’s attitudes towards immigrants are shaped by individual characteristics as well as the specific national context. Countries differ for instance in the ethnic composition, the history of immigration waves, and the national discourse on citizenship and multiculturalism. Hence, can identical survey questions regarding attitudes towards immigrants really be compared across countries?

The main research question is whether or not survey items that are suppose to measure anti-immigrant attitudes really measure anti-immigrant attitudes in the same way across different countries? This is tested using the approach suggested by Van der Veld and Saris (work in progress). That means that we test for measurement equivalence in a bottom-up approach, i.e. we first test for configural invariance, then for metric invariance, and finally for scalar invari-ance. If the latter type of invariance holds, then it is meaningful to compare means and rela-tionships across countries. The estimation of the factor models is done with LISREL (Jöre-skog & Sörbom) on data from the European Social Survey. The testing of the factor models is however done with JRule (Van der Veld, Saris, and Satorra, 2008). The results indicate that the set of items in the scales that we studied, i.e. ethnic threat, ethnic distance, and civil rights, can and should be reduced in order to obtain valid, reliable, and comparable measures.