A Structural Analysis of Class Voting*

Studies of the issue disagree about the reasons for—even the existence of—a low level of class voting in Canada. In order to study class voting, three items must be conceptualized and measured adequately: the class of the voters, the class of the parties, and the structural constraints on political partisanship. Central to our concerns is that the only true leftist party, the New Democrat party (NDP), does not field a candidate in every constituency and is not really a viable party everywhere it does contest the election. Thus, the working classes do not always have the realistic option of voting NDP and research which does not take this into account attenuates the relationship between class and voting. We hypothesize, therefore, that (a) the level of class voting is higher in constituencies in which the NDP is viable and (b) the level of nonvoting among the working classes is higher.where the NDP is not viable. Reconceptualizing class in Marxian terms, our results indicate that class is related to voting and that there is a statistical interaction among class, voting, and the viability of the NDP. Furthermore, nonvoting among the working classes increases in nonviable constituencies and decreases in those constituen-cies where the NDP is viable, indicating that nonvoting may be a class re-sponse similar to voting NDP. Implications of this analysis for other Western democracies are noted. Social class, the principal structural cleavage in Western democracies, has a substantial political importance that typically is reflected in the party sys-tem (Maclver). Alford contended that an association between class and voting choice is "natural and expected. " Rose and Urwin analyzed support for 25 leflist parties and found that 20 of them drew two-thirds or more of their support from the working classes. In fact, Lipset (a) maintained tha