Regression Discontinuity Design (RDD) is a quasi-experimental approach used in econometrics and social sciences to estimate causal effects. It relies on a predetermined cut off point to allocate subjects into different treatment groups. By comparing outcomes for individuals just above and just below this cut off, researchers can estimate the effect of an intervention or treatment.
Regression discontinuity analysis can be divided into two main types:
- Sharp regression discontinuity – In sharp RDD, the assignment to treatment is strictly determined by whether the running variable (e.g., test scores) crosses the threshold.
- Fuzzy regression discontinuity – Fuzzy RDD allows for some ambiguity, where crossing the threshold increases the probability of receiving the treatment, but does not guarantee it.
Visual representation is crucial in RDD analysis. A regression discontinuity graph displays the relationship between the running variable and the outcome, highlighting any discontinuity at the cut off point. This graph helps in visualizing the treatment effect clearly.
A practical regression discontinuity example can be seen in educational research where students scoring above a certain mark receive additional tutoring. By comparing the academic performance of students who score just above and just below this threshold, researchers can isolate the impact of the tutoring program.
Regression discontinuity design is a powerful method for identifying causal effects, especially when random assignment is not feasible. Its ability to produce credible estimates makes it a valuable tool in applied RDD econometrics.
List of recommended resources #
For a broad overview
The Regression Discontinuity Design
This chapter by Denis Fougere and Nicolas Jacquemet in Policy Evaluation: Methods and Approaches gives an overview of this quasi-experimental quantitative method for assessing impact of interventions. It also details how RDD is useful in policy evaluation.
What is the regression discontinuity approach? by Professor Mike Brewer
This lecture delivered by Professor Mike Brewer, as part of the Quantitative Data Handling and Data Analysis course by The National Centre for Research Methods (NCRM), provides research methods training of regression discontinuity design. Professor Brewer introduces regression discontinuity as well as the RDD principle along with examples. He also elaborates on implementation of RDD, along with explaining the different types of RDD in econometrics.
For in depth understanding #
Regression Discontinuity Designs in Economics
This paper by David S. Lee of Princeton University and Thomas Lemieux of University of British Columbia provides an introduction as well as a user guide to regression discontinuity for empirical researchers. It also explains why RDD is considered a “quasi-experimental” design and summarizes different ways of estimating RD designs and the limitations of interpreting these estimates.
A Practical Guide to Regression Discontinuity
This in-depth paper by Robin Jacob of University of Michigan and Pei Zhu, Marie-Andrée Somers, and Howard Bloom of MDRC serves as a practitioners’ guide to implementing RD designs. It seeks to explain things in easy-to-understand language and to offer best practices and general guidance to those attempting an RD analysis. In addition, the guide illustrates the various techniques available to researchers and explores their strengths and weaknesses using a simulated dataset.
Case study #
Regression discontinuity designs in agricultural and environmental economics
This study by David Wuepper and Robert Finger discusses how regression discontinuity designs (RDD) are increasingly being employed in agricultural and environmental economics to identify causal effects and shows how agricultural economists can leverage RDD in combination with remote sensing and environmental modeling.
This paper by Leah M Smith, Linda E Lévesque, Jay S Kaufman, and Erin C Strumpf implements RDD strategies to assess whether it is appropriate for a study of the impact of human papillomavirus vaccination on cervical dysplasia.
References #
Regression discontinuity design studies: a guide for health researchers
The Regression-Discontinuity Design