Uninformed Votes? Reappraising Information Effects and Presidential Preferences

Abstract

Previous work on information effects and preferences has used the technique of statistical imputation to estimate the impact of political ignorance on presidential preferences, suggesting that the electorate would vote differently if more informed. In this paper, I challenge that assertion by disputing the extent to which the changes in preferences generated by imputation are interpretable as information effects. Using data from the 1992–2008 National Election Surveys, I show that the changes in preferences resulting from imputation fail to support a number of hypothesized relationships between political knowledge and preferences. I suggest that the resulting shifts in preferences are most likely attributable to the psychological traits of the more informed rather than to information itself.