Understanding Participation in a Web-Based Measurement Burst Design: Response Metrics and Predictors of Participation

Abstract

Measurement burst designs are becoming a popular method for examining within- and between-person associations; however, little is known about response patterns in these designs. Specifically, a thorough characterization of response metrics at various stages of the data collection has not yet been documented. Additionally, although substantive studies using these designs often mention that differential participation might affect the representativeness of their analytic samples, a better understanding of where in the data collection process these differences arise is needed. Using data from a measurement burst study examining substance use in the year following high school graduation, we demonstrate how response metrics commonly used by panel studies can be adapted for use in such designs. We then evaluate the representativeness of a hypothetical analytic sample and use these new response metrics to identify where in the data collection process these differences arise. Consistent with previous research, we find that sociodemographics and substance use predict inclusion in the analytic sample; however, the exclusion of particular subgroups is due to differential participation at different stages of the data collection. To improve the representativeness of the analytic samples used in measurement burst designs, researchers should consider using this analytic approach to inform tailored data collection methods.