Issue Voting, Projection, And Persuasion: Who Does What?

Abstract

This dissertation focuses on the psychological processes which connect people's own positions on political issues, their evaluations of candidates, and their perceptions of candidates. A critical review of the recent literature reveals that political scientists have not determined the types of people who are most likely to use issues in formulating evaluations of candidates. The goal is to determine if projection or persuasion undermine the ability of more politically involved people to vote on the basis of issues. Hypotheses regarding who should be most likely to project or be persuaded are drawn from cognitive dissonance theory and a political involvement model. The cognitive dissonance hypothesis suggests that those who believe the choice between candidates important should be most susceptible to projection and persuasion. Cognitive dissonance also suggests that the intensities with which one holds issue opinions may be affected by the relationship between individuals' opinions and those of candidates. The political involvement model suggests that persuasion should be weaker among the more politically involved. The simultaneous equation models estimated with data from the Survey Research Center's 1972-74-76 American National Election Panel Study suggest little support for the cognitive dissonance hypothesis. There is more support for the political involvement hypothesis, but only on the redistributive issues that are examined. On the moral issues, lesser involved people are less likely to be persuaded and also less likely to project. More involved people are also found to use issue evaluations more in determining their candidate evaluations than lesser involved people are. However, the differences in the cognitive processes are not large, and suggest that when lesser involved people have positions and perceptions of candidates on issues, they are almost as likely to use them in evaluating candidates. The relationship between the issue intensities and issue distances was null. Implications for the understanding of public opinion are discussed.