Contexts of immigrant receptivity and immigrant religious outcomes: the case of Muslims in Western Europe

Abstract Among migration scholars, immigrant religiosity has become an important variable in understanding immigrant incorporation into the new society, but less studied are determinants of varying immigrant religious outcomes. Using a subsample of immigrant Muslims within the European Social Survey (2002, 2004, 2006), contexts of immigrant receptivity as less or more welcoming are tested on immigrant Muslim religious outcomes using multi-level modelling. Results confirm the hypothesis that less welcoming immigrant contexts are associated with higher religious outcomes among Muslim immigrants in comparison to the host region's religiosity. Keywords: Contexts of receptionimmigrationimmigrant religiosityreligionMuslimsWestern Europe Acknowledgements The author gratefully acknowledges the comments of Alejandro Portes, Richard Alba, Rafaela Dancygier, Robert Wuthnow, Pyong Gap Min, Gordon De Jong, and two anonymous reviewers on previous versions of this manuscript. Appreciation is expressed to the principal investigators of the European Social Survey (ESS) for making their data available for this analysis. The ESS can be accessed through the Norwegian Social Science Data Services NSD. By making this data freely available, it is recognized that any errors committed in this analysis are the sole responsibility of the author. Notes 1. There is an ever-expanding literature on contexts of immigrant receptivity and immigrant incorporation policy more generally; however, space constraints do not permit a thorough review. For a good introduction to pertinent issues, see Bauböck, Heller and Zolberg (Citation1996), The Challenge of Diversity: Integration and Pluralism in Societies of Immigration. 2. Although public policy in the United States is often reflective of public opinion (Page and Shapiro Citation1983), this may not be true for other liberal democracies. European states sometimes aim to educate their native populations through their policies. 3. Participating Western European countries included in this analysis are: Austria, Belgium, Switzerland, Denmark, Germany, Spain, Finland, France, Great Britain, Ireland, Italy, Luxembourg, the Netherlands, Norway, Portugal, and Sweden. 4. Actual response rates by country range considerably, from nearly 60 per cent among some countries to slightly over 70 per cent for others. 5. The selection of regions for analysis was further restricted by a minimum number of fifty natives within the larger sample in determining mean levels of immigrant receptivity by region. The number of immigrant Muslim respondents by region varies from one to forty-seven with a mean count of seven. 6. This ordinal scale was also recoded into a dichotomous variable of monthly or more frequency as well as representative annual counts of frequency. Both logistic and poisson regression models provide similar results to the OLS regression results used in this paper. 7. Other measures of immigrant sentiment were included within the survey as well, including: allow many/few immigrants of same race/ethnic group as majority, allow many/few immigrants of different race/ethnic group as majority, and allow many/few immigrants from poorer countries outside Europe. However, factor analysis indicates that these variables do not favourably correlate with the other three selected variables. Independent variable means by region were calculated using the probability and design weights provided within the ESS. 8. This measure of religious concentration as well as proportion foreign-born were derived by calculating proportions by region found in the full survey sample. 9. The measure for religious pluralism is calculated using the Herfindahl Index – one minus the sum-squared proportion of each religious group. Religious groups were coded according to global religious categories (i.e. Catholic, Protestant, Eastern Orthodox, other Christian, Jewish, Muslim, Eastern religions, other non-Christian religions, and no religion). 10. Although race can influence both religious outcomes and native perceptions of immigrants, race was unavailable in the ESS. 11. Given the variety of sampling strategies employed by each participating country, ESS principal investigators recommend using population and household weights for any analysis. Unfortunately, the xtreg procedure in STATA are not friendly to weights. Additionally, the limited subsample used in this analysis defies the suitability for weights. 12. A country level of analysis was selected since some regions only have a handful of immigrant Muslims; therefore, any testing of a reverse-causation hypothesis requiring mean levels of immigrant Muslim religiosity by region could be unrepresentative of immigrant Muslim religiosity. 13. Means for immigrant Muslims were retained for regions with ten or more Muslim respondents. 14. All three indicators of immigrant receptivity are highly correlated with each other and would portray little difference if mapped separately. 15. Due to the varying level of regions by country, this map was produced by hand; therefore, division lines between regions are approximate. Any errors in these boundaries are the author's responsibility and not that of ESS principal investigators. 16. I am indebted to Alejandro Portes for this valuable insight.