Basics and recent advances in regression discontinuity: Difference versus regression forms*

Jin Young Choi, Myoung Jae Lee

Research output: Contribution to journalArticlepeer-review

2 Citations (Scopus)

Abstract

This paper reviews the basics of regression discontinuity (RD) de-sign, whose hallmark is having a treatment determined by an underlying score (i.e., ‘running variable’) crossing a known cutoff or not. Following the basics, recent advances in RD are examined, where the topics are grouped into those related to score and those not. The former topics include multiple scores, measurement errors in score, integer score, and score-density continuity. The latter topics include regression kink (RK), high-order effects, and extending RD identification range (i.e., external validity). Detailed empirical examples are pro-vided for the RD topics, but not for the RD basics which are fairly well-known these days. RD is simple, which can thus appeal even to lay audiences, and this review accordingly emphasizes the intuitive nature of RD and its applicability in practice. Practical and widely applicable techniques are given more cover-age, whereas theoretically-motivated but less-practically-relevant ones are only briefly mentioned. The beauty of RD is in its simplicity, and temptation to make it too sophisticated should be resisted.

Original languageEnglish
Pages (from-to)1-68
Number of pages68
JournalJournal of Economic Theory and Econometrics
Volume32
Issue number3
Publication statusPublished - 2021 Sep

Keywords

  • Cutoff
  • High-order Effects
  • Instrument
  • Local Randomization
  • Regression Discontinuity
  • Regression Kink
  • Score

ASJC Scopus subject areas

  • Economics and Econometrics

Fingerprint

Dive into the research topics of 'Basics and recent advances in regression discontinuity: Difference versus regression forms*'. Together they form a unique fingerprint.

Cite this