Regression Discontinuity Designs in the Evaluation of Section 1115 Demonstrations
U.S. Department of Health and Human Services, Centers for Medicare & Medicaid Services, Center for Medicaid and CHIP Services
Medicaid section 1115 demonstrations often use thresholds or cut-offs based on continuous variables such as age or income to determine whether beneficiaries are subject to certain policies. Eligibility thresholds present an opportunity for states to evaluate the causal impact of demonstration components using a regression discontinuity design (RDD). This white paper provides states with an accessible resource to support sound implementation of RDD in their evaluations of section 1115 demonstrations. We give an overview of RDD using section 1115 demonstration policies as examples; provide a nontechnical summary of best practices; describe ways to check model assumptions and sensitivity to key design choices; and describe recent extensions to the method that might be useful for evaluating section 1115 demonstrations.