Affiliation: College of Arts and Sciences, Department of Political Science
Contemporary accounts of elections characterize voters who participate in primaries as the ideologically extreme subset of their party. However, previous methodological limitations have impeded efforts to directly measure the ideological extremity of party primary voters at the district level. Utilizing recent methodological innovations (multilevel regression with synthetic postratification, MrsP), I create a direct measure of ideology for primary and general election constituencies for both parties in each congressional district. My estimates demonstrate (1) that the ideological extremity of the primary constituency varies between districts, and, (2) primary election constituents are more ideologically extreme than general election constituents within districts. These estimates and future applications of MrsP will allow for more thorough investigation of subnational public opinion in the United States.