Vote recall: a panel study on the mechanisms that explain vote recall inconsistency

Voters’ recall of past behavior is known to be inaccurate. Yet, owing to data limitations, the precise mechanisms behind recall inconsistency have not yet been pulled apart empirically. We analyze the Dutch 1VOP panel data set (29,955 respondents, 53 waves) to simultaneously test four explanations of recall inconsistency. We conclude that vote recall is explained by forgetfulness, nonattitudes, and cognitive bias, but find no evidence for measurement bias. Recall consistency is affected by current party preference and by volatility in party preferences. Both effects become stronger with longer time intervals between recalls. We discuss the implications for (cross-national) survey research methodology and for substantive research on electoral volatility and offer tentative solutions against the resulting validity, reliability, and equivalence issues.