ELECTORAL CHANGE AND THE FLOATING VOTER: The Reagan Elections.

This paper presents two perspectives on a fundamental issue of elections as mechanisms of democratic accountability. One is the interelection floating voter hypothesis, which implies that it is the least informed segment of the electorate that contributes most to electoral change. The second perspective is from V. O. Key's argument that vote switching is rooted in rational policy concerns. A direct test of Key's formulation of the problem on the Reagan election victories of 1980 and 1984 adds to the evidence supporting Key's perspective. The reasons why some voters hold firm to particular parties and candidates while others switch support is well explained by their different positions on matters of party, policies, and judgments of the candidates. Vote switching is not simply the by-product of an ill-informed segment of the electorate responding to its meager grasp of the short-term stimuli of a campaign. Vote switchers appear to judge the policies and the performance of an incumbent against their best estimates of these qualities in the competing candidate. The data are from the 1980 and 1984 CBS/New York Times exit polls. [ABSTRACT FROM AUTHOR]