Abstract
When asked whether they voted in the last election, many survey respondents say they did even though they did not. Although misreporting has the potential to produce specious findings in candidate choice models, researchers have ignored its effects on our understanding of candidate choice. We test an established candidate choice model first with and then without self-reported voters whose turnout could not be validated, in order to gauge whether misreporters undermine empirical tests of candidate choice models. The results for three different presidential elections indicate that misreporters have little impact on the overall performance of the model or on the impacts of most predictors. Misreporting does produce some misleading findings, but its disruptive effects are confined to factors peripheral to the model; misreporters have little or no effect on party identification or candidate evaluations, the main factors shaping candidate choice.
Get full access to this article
View all access options for this article.
