Party differentiation in Congress

At times, the American political par-ties are so close in terms of policy positions that critics denounce the lack of a "dime's worth of difference" between them. At other times, the gap between them on a left-right dimension is huge. How can we explain this variation? We argue that par-ties can behave rationally as collective units, and that shifts in divergence and convergence can be explained as rational responses to changes within governmental institutions and to shifts in conditions outside. We analyze this argument using adjusted ADA scores (Groseclose, Levitt and Snyder 1999) to compare voting score differences between the Democratic and Republican parties in Congress from 1952 to 1996. We pose specific hypotheses for potentially important factors shaping party behavior and test them with a multivariate model. Our results support the argument that the variation in the behavioral gap between the two parties in Congress can be explained as rational party responses to internal and external stimuli.