Heterogeneity and representation: The senate and free trade

Theory: Constituent pressures on senators vary by party and heterogeneity of state. Studies of representation based on analysis of roll call voting which fail to take these factors into account will underestimate constituency influence and overemphasize the impact of party and interest group ratings. Hypotheses: (1) The greater the heterogeneity of the electorate, the less predictable will be the relationships between constituency characteristics and roll call votes. (2) The effects of some constituency characteristics on roll call votes are party-specific. Methods: Ordered probit analysis of Senate voting on international trade issues in 1993-94 was conducted for the whole Senate and for the Senate split into homogenous and heterogeneous states. Results: For homogenous states, general and party-specific constituency characteristics perform very well, while party and interest group ratings perform poorly. For heterogeneous states, general and party-specific constituency characteristics perform poorly, while party and interest group ratings perform well. These results indicate, first, that heterogeneity of state electorates and party strategies is an important factor in determining roll call votes and, second, that there is strong evidence of dyadic representation on roll call voting.