Drug policy laws vary widely across states, providing a unique opportunity for policy evaluation studies to shed light on which policies are most effective. These studies employ a range of analytic methods, including difference-in-differences, interrupted time series, synthetic control, and autoregression models, among others. Recent advances in difference-in-differences methods have addressed issues such as staggered policy adoption and heterogeneous treatment effects, making them increasingly popular in policy evaluation studies. However, with so many methods available, selecting the most appropriate one for a given context can be challenging for applied researchers.
This talk will provide a brief overview of the most commonly used analytic methods in policy evaluation studies, highlighting the relative advantages and shortcomings of each method. In addition, a methodological "decision tree" framework will be presented to help researchers identify the most appropriate method(s) for their specific policy evaluation context. This framework will highlight key decision points regarding study design (e.g., presence of a comparison group?, more than one state adopted the policy?, staggered policy adoption across states?) that impacts which method(s) may be appropriate. A case study evaluating state opioid policy laws will be used to demonstrate the decision tree framework.
Overall, this talk will provide valuable insights for researchers who are interested in conducting policy evaluation studies. No prior experience with policy evaluation studies is assumed.