Abstract
European nations are undergoing increasing cultural and religious pluralization. Yet, we know little about how crime victimization relates to religion. Different theories suggest that religion might protect from or, on the contrary, be a risk factor for victimization. Drawing on a youth survey (ISRD–3), we examine Finland and Switzerland, two nations with different histories with respect to religious pluralism. We did not observe associations suggesting that membership in minority religions would protect from victimization. The risk of hate crime victimization was elevated among Finnish Muslim youths, while in Switzerland, there appears to be a more general association between ‘other’ religious identification and victimization risk. We conclude by discussing avenues for future research.
Acknowledgements
We thank Ivo Staub and Mikko Aaltonen for their comments and support at various stages of this research.
Notes
1. In 2013, there were 11,125 members of Islamic communities (Statistics Finland, http://www.stat.fi/til/vaerak/2014/01/vaerak_2014_01_2015-12-10_tau_007_en.html), a figure which underestimates the prevalence of Muslims. Based on the ISRD–3 measurement itself, 4–5% of the youths in the 13–16 age bracket self-identified as Muslims at the two Finnish research sites.
2. The data on the current research were collected before the extensive influx of refugees to Finland in 2015.
3. The merged ISRD–3 data-set was not available to us, so we used the original national data. There may be differences with respect to the standard merged data-set, mainly due to different inclusion criteria.
4. Originally, a community size larger than 10,000 inhabitants was the only criterion for the label ‘cities’. The new definition from 2014 is based on the density and number of inhabitants, employees and equivalences from overnight stays.
5. For both countries, we excluded cases with missing predictor variables.
6. Due to the overall size of the data, each group is too small to be used as a stand-alone predictor.
7. We give prevalence findings in the text only, while our tentative regression analyses are shown in Table .
8. In extortion/robbery, assault, and theft, the difference was similar but non-significant (p < .10).
9. As the data have a clustered structure due to the class-based sampling design, the standard errors of the models were adjusted accordingly.