ABSTRACT
The candidate-centered view of congressional elections has two fundamental problems. First, the evidence for the view that electoral outcomes were changed in the 1960s by legislative and campaign resources are very weak. Second, the framework provides no explanation for why more party members in the House would begin to vote together. In the 1950s there was limited class division in American politics. The portrait of candidate-centered elections conveys a sense that personality, campaign funds, and the exploitation of office resources drive election outcomes and that somehow issues and representation of the public get lost. While the candidate-centered view was prompted in large part by evidence of trends in electoral behavior, this reconsideration of the role of constituencies has been prompted by evidence of the increasing party polarization in Congress. At the district level, parties in Congress increasingly derive their seats from districts with differing populations.
