Response privacy and elapsed time since election day as determinants for vote overreporting

Survey respondents have consistently been found to overreport their participation in political elections. Lots of research has been done about the sociodemographic correlates of vote overreporting, but only a few studies analyzed determinants which survey researchers have under their control in order to reduce the problem. Theoretical explanations have assumed memory failure and social desirability bias to explain overreporting. Taking these explanations as a starting point, we tested firstly whether asking retrospective questions about the participation in elections which date back longer in time leads to more overreporting. The second hypothesis is that ensured response privacy eliminates the need for positive self-presentation and thus reduces overreporting. Using data from a field experiment, we found that even when already a substantial time has elapsed between the election under consideration and the survey interview, increasing this time lag still increases the probability of vote reports. This was as well the case when the data was collected interviewer- rather than self-administered. Thus, conducting the survey interview as soon as possible after the political election and using an administration mode which ensures response privacy successfully reduces vote overreporting.