Vote Overreporting and a Survey Experiment: The Case of the Taiwan National Elections

Abstract

The misreporting of voter turnout, prevalent in survey data across the world, has received comparatively little attention anywhere apart from in some western countries. This study evaluates the use of questions specifically designed to mitigate the level of vote overreporting for the 2012 national elections in Taiwan. After a theoretical examination of social desirability and memory failure, the two primary causes of misreporting, we present the results of a split-question experiment featuring two questions designed to mitigate overreporting. While the findings reveal that the experiment with changes to the questionnaire context was far from successful because of a low reported turnout for the control question, it is the case that, as hypothesized, reported voter turnout differs vastly among the different questions, with the question mitigating for social desirability resulting in higher figures than that for memory failure.