107 3 In Figure 6 we analyse whether in-group favouritism explains voting for DENK. To do this, we added ethnic identification variables and in-group favouritism scales to the model with religion and ethnicity (Model 4 of Figure 3). Turkish, Surinamese and Hindustani identification does not have an independent effect on voting for DENK, but Moroccan identification does. At the same time, however, Moroccan in-group favouritism has a negative effect on DENK vote propensity. Thus, there is a positive baseline effect of identifying as Moroccan, but the stronger a Moroccan identifier expresses in-group favouritism, the more this positive effect is cancelled out. Less surprisingly, Turkish and Surinamese levels of in-group favouritism impact voting for DENK positively. Most importantly, though, Muslim identification and in-group favouritism impact voting for DENK positively, by 1.91 and 4.88 points, respectively. Muslim in-group favouritism has by far the highest effect size of all models. The more a voter favours their Muslim in-group, the more likely they are to vote for DENK. Most notably, the effect of identifying as Muslim drops substantially when adding these identification and in-group favouritism measures to the model: from 3.56 to 1.91 points, or a reduction of 46%. Moreover, the effect of being of Turkish descent drops by more than half a point and is no longer significant when adding these variables. The impact of being of Moroccan descent also loses significance, though this effect was not high to begin with. As the effect of being Muslim is still statistically significant, ingroup favouritism does not explain all of the Muslim effect, but the difference is still considerable. Indeed, in-group favouritism explains why Muslims vote DENK much more clearly than discrimination and issues do. The analyses that we have presented thus far comprised the entire sample of respondents, with non-religious respondents without a migration background forming the reference category to the migration background and religious variables. To zoom in specifically on why Muslims vote DENK, we now analyse a subset of only the Muslim respondents. Figure 7 comprises all the explanatory independent variables of our analyses in one model: issues, discrimination and in-group favouritism. Taken together, we find only two variables significantly explain why Muslims vote for DENK: position on immigration (2.36 points) and Muslim in-group favouritism (4.83 points). We also analyse the impact of issues, discrimination and identification on Turkish, Moroccan, Surinamese and Dutch subsets of the sample. For the minority subsets, Muslim in-group favouritism also turns out to be the explanation for DENK voting with the highest explanatory power, over all issue and discrimination variables.
RkJQdWJsaXNoZXIy MjY0ODMw