A Spatial Model for Legislative Roll Call Analysis

A general nonlinear logit model is used to analyze political choice data. The model assumes probabilistic voting based on a spatial utility function. The parameters of the utility function and the spatial coordinates of the choices and the choosers can all be estimated on the basis of observed choices. Ordinary Guttman scaling is a degenerate case of this model. Estimation of the model is implemented in the NOMINATE program for one dimensional analysis of two alternative choices with no nonvoting. The robustness and face validity of the program outputs are evaluated on the basis of roll call voting data for the U.S. House and Senate. Extensive Monte Carlo studies are also presented. Substantive applications using the results for the Senate are briefly illustrated.