ABSTRACT
In this chapter, the authors investigate the performance of low-dimensional spatial models and discuss the substantive meaning of the dimensions. With respect to performance, they show that a simple spatial model adequately accounts for the roll call data. The authors' preferred model has only two dimensions; it limits temporal change in the positions of individual legislators to simple linear functions of time. They address the issue content of the first and second dimensions; the first dimension almost always picks up the fundamental economic issues that separate the two major political parties of the time, while the second dimension divides the parties internally over regional issues. The authors offer supporting evidence for their basic finding of low dimensionality. One important observation is that our low-dimensionality result applies not only to the House, but also to the Senate, where gatekeeping is less prevalent. Another is that certain legislation, particularly in regard to appropriations, must be considered annually and cannot be screened.
