The Conflict-Oriented Group Identity of Partisanship

This dissertation analyzes partisanship in America. I lay out a new theory of partisanship that shows that how strongly an American identifies with the Democrats or Republicans is due to how likely he or she is to take sides in a conflict and join groups generally. I use data from the American National Election Studies (ANES) and a variety of statistical techniques to demonstrate this. Behavior scholars know much about factors that pull Americans toward the Democrats or Republicans, but little (until now) about how strongly an American is pulled toward either party.

Strength of partisanship influences almost all aspects of political behavior. Pure independents, independents who lean toward one party, and weak and strong partisans vary considerably in turnout, vote choice, and political knowledge. My theory explains these differences and improves our understanding of political participation, attitudes, and elections. These findings provide answers for why some people are rabid partisans and others don't care much about their party, and why many Americans who favor one party prefer to remain nominally neutral.

Additionally, I use the conflict-oriented, group identity theory of partisanship to explain differences in strength of partisanship between women and men. I analyze ANES data, and use difference-of-means tests and logistic regression to compare partisanship between genders. I show that men's greater tendency to judge things and take sides in a conflict, and women's greater tendency to join groups, explain why women are more likely to identify as weak partisans and men are more likely to identify as independent leaners.

Finally, I explain gender differences in the social identity and rational choice theories of partisanship. I examine data from the ANES, National Annenberg Election Survey, and Cooperative Congressional Election Studies, and compare models of partisanship using regression techniques and model tests. Ideology and demographics both significantly influence an American's partisanship, but their relative importance is different for women and men. The rational choice model is comparatively better for men, and the social identity model is comparatively better for women. These results increase our knowledge of the gender gap in political behavior.