All Times EDT
Virtual
Propensity Score Matching Concurrently with Participant Recruitment: Examining Cardiovascular Health and PTSD (309924)
*Kathleen Lamarque-Navarrete, University of California, Los AngelesAmanda Kay Montoya, University of California, Los Angeles
Keywords: Propensity Score, Matched Pairs Design, Concurrent Validity
When randomization of experimental variables are not possible, matched pairs design (MPD) can be used to pair treatment and control participants who share covariate values, but requires an exact match in covariates and limits the total number of possible pairs. Propensity score matching (PSM) expands on MPD by aggregating participants’ covariates into a single dimension, and matching based on those scores. However, PSM traditionally utilizes retrospective data, forcing researchers to rely on data that has already been gathered and reduces control over the recruitment process. This work aims to fill the gap between these two methods by proposing PSM that is concurrent with participant recruitment. We demonstrate this method with a secondary dataset that explores Post Traumatic Stress Disorder covariates and diagnosis, and uses PSM on a weekly basis as participants are recruited. We compare the weekly recruitment to the retrospective approach on balance of covariates, matching accuracy, and number of participants in the final sample. This work aims to maximize the number of matched pairs and increase the accuracy of each pair while minimizing cost.