Attrition in a Longitudinal Online Randomized Controlled Trial with Prospective Homeowners

Longitudinal survey and experimental research is essential for understanding psychological and economic processes, but attrition is a concern. We examine attrition in a 2016 longitudinal randomized controlled trial of more than 13,000 prospective U.S. homeowners, in which participants were invited to up to seven online surveys delivered every two weeks. Our research has two aims. First, we document overall attrition, finding that 60% of participants completed the study and attrition primarily occurred soon after enrollment. Second, we examine correlates of attrition, examining not only relationships with demographic and study characteristics (enrolling for text message reminders, survey delivery day) but also considering under-studied economic and psychological characteristics (time preferences, subjective numeracy, and financial literacy). Overall, we find that many variables, including some psychological characteristics, predict attrition. Our results suggest specific variables for researchers to measure in adjusting for attrition and identifying interventions to reduce attrition.

Automated summary

The study found that 60% of participants completed the study, with most attrition occurring soon after enrollment; Factors related to attrition included demographics and study characteristics, as well as under-studied economic and psychological characteristics; Many variables, including some psychological characteristics, were found to predict attrition.