![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
ABSTRACT
Generalised trust promotes social interactions and may well be a crucial component of immigrant integration. Recent immigrants in particular are likely to be viewed by themselves and others as ‘outsiders’ who are unfamiliar with the expectations and norms that structure day-to-day social interactions in the host country. This study relies on a unique combination of three sources of data all derived from World Values Surveys to examine levels of trust, and its sources, among newcomers in one country with a large immigrant population, Canada. The evidence indicates that recent immigrants to Canada make a clear distinction between trust in other people in general, and trust in Canadians in particular: the former is grounded in pre-migration cultural influences, while the latter is grounded in immigrants’ experiences in the new host country. Moreover, the evidence suggests trust in Canadians is a crucial component of immigrant integration.
Acknowledgements
We would like to thank Alain Deschamps for his research assistance and the anonymous reviewers for their very helpful comments.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. We have excluded from the analyses 393 immigrants from the WVS because the WVS questionnaire did not allow us to identify their specific country of origin. These include a small number of immigrants who have also been in Canada for 10 years or less.
2. A precise match of the geographic locations of WVS respondents with those of NIS respondents is not possible. We restrict the WVS sample to respondents in Quebec, Ontario, and British Columbia who resided in cities with populations of more than 500,000. All of the British Columbia respondents in this restricted sample lived in Greater Vancouver; based of the size of cities with population of more than 500,000 people in Quebec and Ontario, we estimate that more than 85% of respondents in these two provinces lived in greater Montreal and Toronto.
3. Responses were elicited for a list of 10 groups. The results of a factor analysis of the NIS responses to this battery of questions reveal that only two items have strong loadings on the first factor: ‘Canadians' and ‘French Canadians' (). Neither of these items have strong loadings on any other factor (indeed, none of the other loadings generated eigenvalues above the standard 1.0 cut-off). ‘French Canadians’, of course, is likely to hold different meaning in Montreal, where the overwhelming majority of the population is French-speaking, than in Toronto or Vancouver. As it turns out, levels of trust in ‘Canadians' and ‘French Canadians' are very closely related among Montreal respondents (correlation = .87) but less so among people in Toronto (.67) and Vancouver (.62). Consequently, we decided to focus only on trust in ‘Canadians’. The wording of the question was different in 2000; we therefore do not use it for this part of the analysis.
4. For the NIS sample, the original distribution is the following: ‘Trust completely' (8.6%), ‘trust somewhat' (76.1%), ‘do not trust very much' (14.3%), ‘do not trust at all' (1.0%). For the analyses that follow, we also analysed other transformations of the variable, but they resulted in broadly similar conclusions as those reached using the current transformation. Those alternative models are not presented.
5. This proportion is for the Canadian-born population residing largely in Montreal, Toronto, and Vancouver. Additional analyses (not presented here, but available from the authors upon request) indicate that the results are not significantly different when we broaden the analyses to the entire Canadian-born sample. This holds for the entire set of analyses presented in this study.
6. We are primarily concerned with the impact of early socialisation experiences of immigrants. Accordingly, all foreign-born respondents are assigned the mean score from earliest WVS survey available for their country of origin (starting from 1981). Canadian-born respondents are assigned the mean score for the Canadian population score for 2000 or 2006, corresponding to their year of interview. We have not weighted these data.
7. An additional question we explored is why generalised trust is relatively lower among the Canadian-born population, when there is essentially no difference between recent immigrants and those born in Canada with respect to trust in Canadians. A possibility we considered is that negative attitudes towards outgroups might lower Canadian-born respondents’ generalised trust. We investigated this possibility by examining whether controlling for trust in recent immigrants and trust in people from ethnic groups other than one's own reduces the gap in generalised trust between recent immigrants and the Canadian-born population. It does not.
8. There was no evidence of a gap between recent immigrants and the Canadian-born population, even when we control for socio-economic factors. Thus, we did not proceed with further analysis.
9. In our study, that challenge is further complicated by the fact that our dependent variables and many of our independent variables are dichotomous, and there is no generally accepted statistical procedure for dealing with endogeneity under these kinds of conditions (Angrist Citation2001). One frequently used approach is to instrument the endogenous variables in a two-stage multivariate model. Quite aside from the absence of any commonly accepted strategy for dealing with both dichotomous dependent variables and dichotomous endogenous variables, we were unable to identify suitable exogenous variables to use as instruments. As such, we interpret the multivariate results with caution, noting instances where reciprocal causation and biased results are possibilities.
10. It is not clear why perceived mistreatment by other Canadians has no significant effect. It may well be the case that asking respondents whether they have been mistreated by other Canadians is too general a question. We also explored the possibility that mistreatment by the government and by other Canadians might be collinear, thereby making it more difficult to assess their independent impacts; it turns out however that there is no collinearity problem between these two variables.
11. We tested whether the length of residence effects are curvilinear by adding a squared term for length of residence. The coefficient was not significant and there was no improvement in model fit. We also examined interactions between length of residence and our other independent variables; none of these were significant.
12. Another specification of the model was performed for trust in Canadians, this time including generalised trust as an independent variable. However, because there is probably a reciprocal causal relationship between generalised trust and trust in Canadians we excluded generalised trust from the final model. The results for all other coefficients in this alternative model specification are virtually the same as those presented in .
13. However, just as the effects of citizenship may be different in other national contexts, they may also vary across time in Canada, depending on the policy context. For instance, important restrictive citizenship reforms have been implemented in Canada since the data were collected for this project (see Winter Citation2014). The effects of citizenship might be different in this new policy context.
14. We assessed the impact on the findings of Chinese immigrants, a major source of immigration in Canada, using two alternate specifications of the models presented in this article. First, we incorporated a dummy variable for Chinese immigrants. Second, we excluded Chinese immigrants and focused only on immigrants from other countries. None of the results in these models were substantially different from those presented in this article.