Can question wording reduce vote overreporting in post-election surveys?

This research note investigates whether question wording can help reduce vote overreporting in post-election surveys. It draws on a survey experiment embedded in the 2019 Swiss Election Study (Selects) and leverages a rich dataset of individually validated vote records for both respondents and non-respondents. This allows us to isolate vote overreporting from the other major source of turnout bias – namely, voter overrepresentation. Importantly, our empirical test represents an “extra-hard case," as the standard turnout question used in Selects already includes a face-saving preamble and response options designed to reduce social desirability bias. The three alternative wording formats – normalizing abstention, emphasizing the importance of the turnout question, and highlighting anonymity – perform slightly better than the standard version, though none reach conventional levels of statistical significance. The anonymity condition comes closest. These near-null results carry important implications for survey research and offer insights into improving the measurement of voter turnout, in Switzerland and beyond.