Attitudes to immigration and human values – a repeated cross-sectional study in three European countries

This doctoral research investigates the relationship between attitudes towards immigration and human values in three European countries; UK, Germany and Sweden. I use ESS survey data from 2002 – 2014 and new methodological approaches to examine this link and further explore if there is country specific differences and whether attitudes to immigration changes over time. In addition to this I also add various socio-demographic and political factors to the analysis. Previous research in this area is scarce and generally assumes the effect of human value dimensions on attitudes is universal and the there are no country-specific differences. To answer the question if human values shape attitudes to immigration and whether there are country specific differences I used cumulative link ordinal regression modelling which is the most appropriate statistical method for analysing such data. Analysis was carried out on a country by country basis, looking separately also at the majority native population as well as the minority ethnic population. Items contributing to the human value dimensions of universalism and conservation showed a clear association with attitudes towards immigration, which was in line with previous research. However, not all human value-attitude relationships were in agreement with Schwartz’s theory of effect direction, but with some value items an opposite relationship was observed. A new latent class approach was used to examine whether human values change over time and there was is no evidence that the latent classes themselves change definition over time, implying that Schwartz is correct when he states that human values are invariant and do not change over time. While individuals may change their view and move classes, there is no evidence of the value items realigning themselves around a new latent class structure. The developed latent class model used to investigate whether the latent class definitions or profiles change over time, in the case of human values, is new and extents the multi-group latent class model into continuous groups where a linear change is expected. The finding of this research concludes that, contrary to Sagiv and Schwartz (1995), the effect of human values on attitudes appears not to be universal, changing both over the subsample considered and from country to country. In general, the thesis concludes that clear country differences are to be expected.