Beyond Spatial Models: A New Perspective on American Operational Ideology Using Latent Class Analysis

Abstract

Analyses of operational ideology—the pattern of correlations between different political attitudes—in the American public generally assume “spatial” models of ideology. Using Latent Class Analysis, we relax many of these assumptions by treating operational ideology as a latent categorical variable and analyze the changing structure of American operational ideology between 2004 and 2020. We find that some Americans during this period held consistently liberal or conservative views and were well sorted into the “correct” political parties. For other Americans, however, we observe complex and shifting relationships between partisanship and economic, moral, and racial attitudes. We find that Racial Justice Communitarians consistently prefer to identify as Democrats, while Nativist Communitarians and Libertarians both tended to identify with whatever party won the most recent presidential election. Future studies of operational ideology should be wary of simplifying assumptions that obscure important dynamics in American politics.