When does religiosity matter for attitudes to immigration? The impact of economic insecurity and religious norms in Europe

ABSTRACT

Religious identity and practice has been associated with attitudes to immigration in Europe, but it is not known how this relationship varies between different cultural and economic contexts. Analysing data from seven waves of the European Social Survey (ESS) 2002–2014 we examine the association cross-nationally and over time, in what was a financially unstable period for many European countries and households. We have two main findings. Firstly, it is not religion per se, but rather conformity to national rates of religiosity which is associated with concern about the economic and cultural consequences of immigration. Secondly, the association between religion and anti-immigration is strengthened in contexts of economic uncertainty. These findings suggest that while religion does not predict immigration attitudes uniformly across countries, when religion reflects cultural conformity, it could become an expression national or ethnic group identity in times of economic insecurity.

KEYWORDS:

• Religion
• immigration attitudes
• economic insecurity
• unemployment
• social norms
• Europe

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributor

Ingrid Storm is a British Academy Postdoctoral Fellow at the University of Manchester. She has published extensively on religious change, and on attitudes to immigration and ethnic and religious minorities, in Journal for the Scientific Study of Religion, European Sociological Review and British Journal of Sociology among others. She is currently researching the relationships between economic insecurity and religion in Europe.

ORCID

Ingrid Storm http://orcid.org/0000-0003-4185-9893

Notes

1 Information about sampling and survey response rates for each country and survey year is available in the Survey Documentation Reports and Fieldwork Summaries http://www.europeansocialsurvey.org/data/round-index.html

2 Turkey is the only Muslim majority country. Excluding it does not substantially affect the results.

3 The other six categories are Catholic, Protestant, Orthodox, Other Christian, Muslim and Other non-Christian. The original variable had separate categories for ‘Jewish’, ‘Eastern’ and ‘Other non-Christian’ religions, but these were combined as each amounted to less than 0.4% of the sample.

4 A previous model included weekly prayer, but this variable was mostly insignificant and excluded from the final analysis to avoid multicollinearity. There is no measure of religious belief available in the dataset.

5 While there is an additional category of ‘Once a day’ in the original variable, less than 1% of the sample (and less than 10 respondents in some countries) reported this level of attendance frequency, and it was recoded as ‘More than once a week’ to increase comparability across countries and denominations. Recoding frequency of religious practice as a dichotomous variable of once a month or less often, makes no significant difference to the model fit.

6 The country’s major denomination was not included as a contextual variable due to high correlations with individual denominational affiliation (Cramer’s V = 0.68). In an alternative model a variable of whether the respondent belonged to the dominant country denomination or a minority religion, was included instead of proportion of adherents to the same religion. The results were largely similar, with ‘majority’ affiliates being consistently more concerned, and ‘minority’ affiliates less concerned about immigration compared to non-affiliates. In order to test the ‘conformity’ hypothesis for affiliates and non-affiliates with the same measure, proportion of adherents to the same religion was included in the final model. Results of the alternative models are available from authors on request.

7 Age is multiplied by 100 to make the tables easier to read.

8 The economic variables are not available for Russia and Ukraine.

9 The models in Tables 2–4 are not directly comparable, since the sample size varies depending on which variables are included. Log-likelihood tests have been performed by comparing the model with the less complex model conducted on the same sample, that is, model 7 is compared to model 2 without the cases missing values on GDP.

10 Other country level variables that have been controlled for in alternative models are per cent foreign citizens, per cent foreign-born populations, the Gini coefficient, and the Human development index, but neither of them were significant, and they were thus excluded from the analysis.

11 In an additional analysis of changes over time (fixed effects model of country-level variables), we found that increases in unemployment were associated with increased economic concerns about immigration, but only in religious countries. The results are available on request.

12 It is difficult to separate the effects of economic and religious contextual variables as they are correlated. However, including the interactions simultaneously for model 4 and model 6, confirms that the interaction with national rates of religiosity accounts for most, but not all of the variances. The interaction with unemployment levels was significant but small for both outcome variables. These models are not included here due to collinearity.

Additional information

Funding

This work was supported by the British Academy under The Postdoctoral Fellowship Scheme.