Political involvement and memory failure as interdependent determinants of vote overreporting

Survey respondents have been found to systematically overreport their participation in political elections. Although the sociodemographic correlates of this response bias are well known, only a few studies have analysed the determinants predicted by two prominent theoretical explanations for vote overreporting: memory failure and social desirability (SD) bias. Both explanations have received empirical support in studies in which the probability of vote overreporting was found to increase (a) with the time between the election and the survey interview and (b) when respondents were politically involved to a larger extent. In the present paper, we argue that the effect of each of these determinants is not simply additive but depends on the value of the respective other factor. This interaction effect has been found with data from the American National Election Studies: The probability of vote overreporting increases significantly more strongly with the respondents' political involvement when more time has elapsed since the election day. Copyright (c) 2007 John Wiley & Sons, Ltd.