Why Bother Asking? The Limited Value of Self-Reported Vote Intention

How accurate are responses to questions about intentions to vote in an upcoming election? Questions of this type are studied in a range of work in political science to understand the effects of other factors on political engagement, as well as in public opinion research. We analyze six phone surveys conducted over two elections which include pre-election vote intention and postelection vote validation (N=24,303). As expected, many who report intending to vote actually do not vote (13% and 54% for the two elections). More surprisingly, high rates people who predicted they would not vote actually do vote (56% and 39%). For both forms of inaccurate self-prediction, respondents were much more accurate when predicting that they would behave consistently with their past behavior than when predicting that they would behave inconsistently with their past behavior. We discuss implications for political science research, behavioral prediction, election administration, and public opinion.