A Test of Downsian Voter Rationality: 1964 Presidential Voting

Abstract

Systematic testing of Downsian voter rationality is accomplished using a computer simulation technique on the 1964 SRC voting survey. The simulation tests both the hypotheses predicting whether an individual will vote and for whom an individual will vote. To evaluate the results of the tests we develop a statistic analogous to Pearson's r. This statistic measures the percentage improvement over a random guess technique. Utilizing this statistic, Downs explains 68.5 percent of the unexplained variance in the voters' choices of party. Three alternative interpretations of the turnout decision are then considered, each premised on a different notion of how the costs of voting are distributed among the voters. Here we use an Engel Curve technique to develop the turnout decision and explain 92 percent of the variance. The importance of the various elements of the Downsian theory are evaluated and, in contrast to some recent conjectures, the probability of making a difference on the outcome of the election is shown to have an effect on the turnout decision. Finally, to determine the viability of the results, the SRC “6 factor” model is developed in an analogous fashion and used to predict both turnout and direction of vote.