Abstract
This contribution investigates the socially conservative attitudes of newly arrived immigrants from Poland in two Western European countries, Ireland and the Netherlands, with a particular interest in the selective nature of out-migration along non-economic factors and attitude change after migration. It focuses on attitudes towards homosexuality, which remain on average less accepting amongst Polish natives than those of these two residence countries. First, we infer from comparisons between migrants and non-migrants residing in the origin country whether a selection effect of migrants with a more liberal attitude exists. We find that there is evidence for selective out-migration that remains when controlling for pre-migration characteristics. Second, using data collected shortly after arrival of immigrants and from a second wave one and a half years later in the residence country, we show whether immigrants adapt to the norms of the residence country and to what extent they maintain the dominant attitudes of the origin country. We find that attitudinal changes after migration are associated with the level of social integration in the country of residence and maintenance of religious involvement.
Disclosure Statement
No potential conflict of interest was reported by the authors.
Notes
[1] From the ESS data-set, we excluded non-citizens considering that the inclusion criteria for the SCIP study specified that only Polish nationals would be interviewed.
[2] All analyses were checked to determine whether there was an effect of sampling method. While migrants in the referral sample had more negative attitudes overall than those from the non-referral sample, findings presented here are robust regardless of sampling method.
[3] Using replacement means that an ESS respondent can be used more than once as a match for a SCIP respondent, as the emigrants are a highly selective group.
[4] A range of robustness tests were performed that show that differences in sampling method in Ireland (referral versus non-referral methods), previous visits in the residence country and previous contacts in the residence country did not alter results.
[5] Results are robust if using the variables prayer and attendance separately.
[6] As above with the propensity score models, controlling for previous contacts with people in the residence country and sampling method in Ireland did not alter results in the regression models.
[7] As language may mediate residence country contacts, additional models were estimated that control for language ability of the respondent, but these do not differ in any substantial way from the models presented here. Interaction effects between media usage and language ability were not significant.
[8] A separate model that interacts the residence country exposure variables (media use, contact, work/study) with the residence country shows that patterns do not differ significantly between Ireland and the Netherlands.