Ethnic Affinity Voting in Canada
Voting , Race and Ethnicity , Social Identity Theory
As immigration transforms the electorates of contemporary democracies, the political activity of newcomers will have increasingly powerful effects on political life. This thesis examines one important dimension of this process: the role of ethnic affinity in electoral politics. Affinity effects lead to voters being more likely to support members of the same ethnocultural group - which, as well as vote choice effects, might form the basis of support for leadership contests, counterbalance discrimination, and underpin more general issue coalitions. However, data from the United States, and some from Europe, suggests that inter-minority conflict, rather than cooperation, is the norm. This severely limits the breadth - and hence influence - of potential “rainbow coalitions”. To test for the existence and dimensions of ethnic affinity effects in Canada, this dissertation uses data from a web-based survey experiment with a large panel of racialized respondents (~1500). Respondents evaluated candidates in a hypothetical election, with candidate ethnicity experimentally manipulated. The analysis shows that respondents have strong affinity for their own ethnocultural group, some affinity for other minority candidates, and certainly no inter-ethnic discrimination. Ethnic affinity effects not only apply to Canada, broader “rainbow coalitions” seem much more likely than previously suggested. Moreover, a contextual analysis suggests that Canada is not an outlier – these effects should apply to many countries. The second part of the dissertation explores the causes and motivations for affinity effects. Two categories of explanations are explored – interest-based explanations, and identity-based explanations. For this, the dissertation uses a novel application of the Identification with a Psychological Group Scale, drawn from social psychology, a manipulation that primes self-interest, and other measures. In a sharp contrast to previous research, these results show strong evidence for identity effects and a significant amount of evidence against interest-based explanations. Finally, a new data set on the ethnicity of several thousand candidates over four federal elections, combined with riding level census data, demonstrates that the context for ethnic affinity effects is widespread.