Regression discontinuity design (RDD) is a causal inference technique that can be used with observational data. Social scientists are often faced with challenges when using observational data in contexts where an experimental design is infeasible. In such contexts, under fewer assumptions relative to other causal inference techniques, an RDD may be useful to isolate causal effects. RDD has gained traction over the past decades in various disciplines, including economics, political science, sociology, and epidemiology, to analyse questions related to labour, health, crime, voting behaviour, and environment, to name a few.
This entry discusses the practicalities of implementing an RDD with a focus on the intuition behind its applicability. It first discusses its origins and details the identification strategy. This is followed by a discussion of some ...