Key points are not available for this paper at this time.
The research reported in this article involved tests of a model by which voting decisions can be explained and predicted. Data for the tests came from surveys conducted in five presidential elections by the Survey Research Center of the University of Michigan. Predictions made in terms of the model show it to be a good basis both for predicting the division of the vote and for predicting the votes of individual voters. Extensive analyses of incorrect predictions suggest them to be in great part the sort of errors one would expect, were voters arriving at their voting decisions in the manner described by the model. The validity of the model has implications of importance for practical politics, political history, and political theory.
Kelley et al. (Sat,) studied this question.