Abstract: Legislative parties respond to the changing preferences of the citizens they represent through the adaptation and replacement of their members. This creates an indirect electoral connection between the aggregate preferences of citizens and the aggregate behavior of legislative parties. In this research, I argue that legislators from moderate districts are the least likely to support their legislative parties and most likely to vote moderately during roll call votes. I also argue that states with low ideological variance among citizens are the most likely to have moderate districts. This implies that states with ideologically heterogeneous populations are more likely to have homogeneous, ideologically extreme legislative parties. Using measures of legislative ideal points and party cohesion from U.S. state legislative parties, empirical evidence largely supports my expectations. The polarization and internal homogeneity of state legislative parties is at least partly attributable to the citizens those parties represent.
Published: Political Research Quarterly. 2014. Vol 67 (3). Find the online version here!
Find the supplemental appendix here.
Find the replication materials here.