ABSTRACT
There is an on-going debate whether a ‘clash of civilizations’ exists between Islam and the West, in particular in relation to support for democracy and endorsement of liberal sexual mores. This study aims to explain Muslim minorities’ support for democracy and their attitude towards these mores by making a comparison with majority members and by considering the level of religiosity and sense of national belonging. Using data from four European countries (Belgium, Germany, U.K., Switzerland), we found that Muslims compared to majority members were less supportive of democracy and of liberal sexual mores. Furthermore, these differences could be explained by Muslims’ higher level of religiosity and weaker national belonging (for liberal sexual mores). These findings demonstrate that the so-called ‘clash’ is more prominent in the case of liberal sexual mores than democracy and that not only religiosity but also the development of a sense of national belonging is important.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 A shorter version of the survey was also used in the Netherlands and France. However, since many of the questions related to the variables in this study were not included in this short version, these countries could not be included in this study.
2 The information on the language that the interview was conducted in was missing for 21.6% of the participants
3 Participants were identified as being Muslim minorities if they stated their denomination as Muslim and if they and/or their parents had an immigrant background. Furthermore, participants who had Muslim immigrant parents were also categorised as Muslim minorities. In other words, non-religious individuals were also included among the Muslim sample. Majority members were defined as native-born (with native-born parents), with a Protestant or Catholic background.
4 The following formula was used to calculate rho = (variance of the factor * sum of the item loadings)2 / ((variance of the factor * sum of the item loadings)2 + sum of the residual variances of the items + (2 * sum of the residual covariances of the items)).
5 It has been proposed that religiosity can be distinguished in terms of the three dimensions of religious belief, belonging and behaviour (Kellstedt et al. Citation1996). We investigated whether this distinction could be made empirically with the current items and across the four countries. This turned out not to be the case, with for example problematic measurement equivalence across countries. Furthermore, there were considerable cross-loadings with means that various theoretically uninformed modifications would have to be made in order to achieve a good model fit.
6 Furthermore, this category was predominantly formed of Muslim immigrants, who stated not having a diploma from their country of origin. Further analyses demonstrated that some of them had completed schooling in the host country. Nevertheless, in order to not create false truths, these participants were treated as ‘missing’ only on this education variable
7 The chi-square test indicated that the model did not fit the data perfectly. However, with large sample sizes, model fits are often significant, as no model is perfect. RMSEA = acceptable fit <.08, good fit <.05, CFI >.900, TLI >.900 (Kline Citation2015)
8 The effect of Muslim was freed on the following items: frequency of religious service attendance, expressing religious beliefs by abstaining from drinking alcohol, and following dietary regulations. In the case of the first item, majority members demonstrated attending the church more often than Muslims. This may be surprising due to the otherwise lower religiosity of majority members. However, mosques can be less locally available to Muslims (e.g. Carol and Koopmans Citation2013). The other two items are (nowadays) associated more with Islam than Christianity (e.g. distinctions of “halal/haram” in terms of dietary practices, in the case of both food and alcohol, Hämeen-Anttila Citation2004). The items on which the effect of gender was relaxed were “showing religious beliefs by covering hair” and frequency of church attendance. Veiling is a practice more specific for women and mosques tend to be visited more by men than women.
9 As suggested by one of our reviewers, we also examined whether items tapping into other aspects of support for democracy, such as free speech, separation of church and state, and authoritarian leadership could be examined as an additional dimension. We ran an exploratory factor analysis with our original items included which resulted in a single factor, with the exception of the free speech item. However, the reliability of this four item measure was poor and importantly the pattern of findings was similar as with the two item measure that we used. We also did a factor analysis in which we forced the four democracy items on two separate factors (excluding the free speech item). This showed that one item had an unacceptable cross-loadings (>. 39) and that the differences between one of the items cross-loadings was only .101. These cross-loadings mean that various theoretically uninformed modifications would have to be made in order to achieve a good model fit. Nevertheless, we re-examined our structural model with two measures of support for democracy and all the associations remained the same as in our original analysis. Thus, we can be confident of the findings presented in our model.