Short-Term Fluctuations in U.S. Voting Behavior, 1896–1964

This study is an attempt to employ some simple statistical models, motivated by certain assumptions about voting akin to those discussed by Downs and others, in an attempt to explain short-term fluctuations in the division of the national vote for the U. S. House of Representatives, over the period 1896–1964. The models will yield quantitative estimates of the impact of economic conditions on congressional elections, and of the effects of incumbency and presidential “coattails†as well.The notion that a vote represents a decision or rational choice between alternatives is an important theme in democratic theory. However, this rationality hypothesis has proved to be difficult to test empirically, particularly with survey data, from which most of our recent knowledge of individual voting behavior is drawn. The present study is an attempt to put a modified form of the rationality hypothesis to a different and in some respects more direct test than is readily possible with survey data.The analysis bears directly on the substantive question of the relationships between economic conditions and U. S. national election results.