A1574
Title: Handling correlation in stacked difference-in-differences estimates with application to medical cannabis policy
Authors: Nicholas Seewald - University of Pennsylvania (United States) [presenting]
Beth McGinty - Weill Cornell Medicine (United States)
Kayla Tormohlen - Weill Cornell Medicine (United States)
Ian Schmid - Johns Hopkins Bloomberg School of Public Health (United States)
Elizabeth Stuart - Johns Hopkins Bloomberg School of Public Health (United States)
Abstract: Health policy researchers often have questions about the effects of a policy implemented at some cluster-level unit, e.g., states, counties, hospitals, etc., on individual-level outcomes collected over multiple time periods. Stacked difference-in-differences is an increasingly popular way to estimate these effects. The approach involves estimating treatment effects for each policy-implementing unit, then, if scientifically appropriate, aggregating them to an average effect estimate. However, when individual-level data are available, and non-implementing units are used as comparators for multiple policy-implementing units, data from untreated individuals may be used across multiple analyses, thereby inducing a correlation between effect estimates. Existing methods do not quantify or account for this sharing of controls. A stacked difference-in-differences study is described, investigating the effects of state medical cannabis laws on treatment for chronic pain, a framework for estimating and managing this correlation due to shared control individuals is discussed, and how accounting for it affects the substantive results is shown.
