Abstract
Despite the centrality of party identification in understandings of political behavior in the United States, there is an unacknowledged disparity between our theories and measurement of the phenomenon. The traditional method of measuring party identification relies on supplying cognitive cues by explicitly asking respondents to “think” about their partisanship. The Michigan theory of party identification, in contrast, assumes that partisanship is primarily affective. Using a survey experiment, we explore the effects of asking respondents to feel rather than think about their party identification. The new questions reveal that the electorate is more Republican than previously thought. Response timers show that respondents take longer to answer the new items, suggesting that they are surveying a wider and deeper array of considerations. These results serve to revive many of our traditional conceptions of how party identity works while also opening the door for new research questions.
