Party images and party identification in Canada

The notion that electors acquire an enduring allegiance to a political party has guided the study of voting behaviour and elections in western democracies for almost three decades. Yet when used outside the United States, where the concept was developed, party identification has been greeted with less unanimous enthusiasm. For Canadian scholars, debate over the concept revolves primarily around (1) the empirical independence of party identification and the vote and (2) the stability of such attachments. Indeed, this debate has occupied much of the efforts of previous analysts in Canada. In this thesis, we reexamine the conventional wisdom about party identification in Canada by looking at two questions. First, are party identification and the vote distinguishable empirical referents in Canada? And secondly, how can we account for changes in the party identifications of Canadians? Thus, the first question seeks to address the property of independence, and the second, the property of stability. The first hypothesis is that the party identification and the vote are not the same thing, but differ by the impact of short-term forces that contradict partisan attachments. This hypothesis is supported—for all types of partisans, the differences between vote and party identification can be understood, at least in part, as a function of short-term forces. The second hypothesis is that levels of stability and change in party identification are related to individual party images. That is, when these “mental pictures” of the parties serve to reinforce past party identification, identifications will be stable. On the other hand, when these images are in conflict with the party one identified with in the past, there is a smaller likelihood that the individual will remain a stable identifier. We test this hypothesis in two ways. The first relates image of one’s own party to patterns of change and stability in identification, while the second correlates the relative image of all three parties with patterns of partisan stability. In both analyses the patterns that emerge are generally consistent with the expectations derived from the hypothesis. All partisans are more likely to change their party identifications when their images of the parties are in conflict with past party attachments. However, when party images reinforce past identifications, identifications are likely to endure