Regression discontinuity: review with extensions (Q1685300)

In treatment effect analysis, often the treatment takes a particular structure: an underlying continuous variable crosses an institutional/legal cutoff to get treated. This occurs in various settings such as a test score crossing a threshold to graduate, or income falling below a threshold to qualify for an aid. In such cases the study design is called \textit{regression discontinuity} (RD). The paper reviews RD to show its essentials: (a) the main RD idea is introduced based on local randomization due to an institutional/legal break, (b) treatment effects identified by RD are studied, (c) popular RD estimators are mentioned including LSE and instrumental variable estimator, and (d) main specification tests are examined. Special RD topics are reviewed as well, including multiple scores, measurement error problem, and quantile-based RD.