Immigrant generation, religiosity and civic engagement in Britain

Abstract

Immigrant integration appears to be generational in the USA, and further facilitated by religious involvement. We examine whether similar patterns exist in Britain. We find evidence for secularization across ethnic minority groups, measured by private religious practice and religious salience. Communal religious practice appears robust to generational decline. Ethnic minority members of the second generation exhibit lower social trust; for the 1.5 generation, being more religious is associated with lower trust. However, members of the 1.5 and second generation are more civically involved than the first and religiosity further increases civic involvement. While anecdotal accounts suggest that religiosity has a particularly dissociative effect on the second generation, we find no evidence for this. In sum, successive generations of ethnic minority respondents appear to be secularizing; successive generations are more civically involved than the arriving generation, although less trusting; and immigrant religiosity promotes civic integration.

Keywords:

• civic engagement
• ethnic minorities
• generations
• integration
• religiosity
• trust

Notes

1. The Kaiser-Meyer-Olkin measure is acceptable at 0.643 and Bartlett test of sphericity significant (p<.001); we accordingly reject the null hypothesis that the correlation matrix is an identity matrix.

2. Simultaneous estimation of the three outcomes via multivariate regression, taking into account the full covariance structure, allows efficient estimates of model coefficients and standard errors.

3. It is also possible to calculate predicted values for an ‘average individual’ (e.g. one who is 47 per cent male and 53 per cent married) or by averaging over individuals in the sample. We chose to calculate marginal effects at representative values.

4. We use the ‘margins’ suite in Stata 11.0 to calculate pairwise comparisons of predictive margins for the separate effects of generational and ethnic status; then the conditional marginal effect of generation for each ethnic group in turn. We thus minimize the ‘multiple comparisons problem’ by restricting the number conducted. While we do not correct the p values for the multiple comparisons, a significant omnibus F test of the joint significance of the interaction terms (not reported here) protects against an inflated type I error.

5. The majority of the analysis conducted here uses the Ethnic Minority British Election Survey 2010. However, responses to the trust question (‘most people can be trusted’, ‘you can't be too careful’ and unprompted responses of ‘it depends’) were not directly commensurate with that for the main British Election Study, which used a 0–10 scale. Accordingly, the Citizenship Survey 2010–11 was chosen to allow a comparison with whites.

6. We suspect differential item functioning between whites and ethnic minority respondents for the option ‘it depends’, where response rates are considerably higher for the latter. Accordingly, we group those replying ‘most people can be trusted’ with those replying ‘it depends’. To simplify the graph we group the 1.5 with the second generation.

Additional information

Notes on contributors

Siobhan McAndrew

SIOBHAN MCANDREW is Marston Research Fellow at the Institute for Social Change, University of Manchester

David Voas

DAVID VOAS is Professor of Population Studies at the Institute for Social and Economic Research, University of Essex