Regression discontinuity design (RDD) and difference-in-differences (DID) are among the main methods for evaluating the effects of policy interventions employed without conducting experiments.
RDD is used when an intervention is implemented only if a certain variable (the running variable) exceeds a specific threshold (cutoff)—for example, vaccination eligibility determined by date of birth, the selection of granted firms for subsidy programs, or health guidance following health checks. By comparing units just above and below the cutoff, RDD enables impact evaluations that operate similarly to randomized controlled trials (RCTs).
DID is used when an intervention is introduced at a certain point in time for only a portion of a population (e.g., policies implemented in only some prefectures). It evaluates whether the change in an outcome variable before and after the intervention differs between the treatment group and a control group (those not exposed to the intervention).
