Reporting on one's behavior: a survey experiment on the nonvalidity of self-reported COVID-19 hygiene-relevant routine behaviors

Surveys based on self-reported hygiene relevant routine behaviours have played a crucial role in policy responses to the COVID-19 pandemic. In this article, using anchoring to test validity in a randomised controlled survey experiment during the COVID-19 pandemic, we demonstrate that asking people to self-report on the frequency of routine behaviours are prone to significant measurement error and systematic bias. Specifically, we find that participants across age, gender and political allegiance report higher (lower) frequencies of COVID-19 relevant behaviours when provided with a higher (lower) anchor. The results confirm that such self-reports should not be regarded as behavioural data and should primarily be used to inform policy decisions if better alternatives are not available. To this end, we discuss the use of anchoring as a validity test relative to self-reported behaviours as well as viable alternatives to self-reports when seeking to behaviourally inform policy decisions.