The rdrobust packages provide R and Stata implementations of statistical inference and graphical procedures for Regression Discontinuity designs. These implementations include local-polynomial-based point estimators, confidence intervals estimators and bandwidth selectors, as well as automatic RD-plots and other features. The papers below discuss several aspects (theoretical, methodological and practical) of the procedures implemented in the rdrobust packages.

This research is supported by the National Science Foundation through grant SES-1357561.

AuthorsSebastian Calonico, Matias D. Cattaneo, Rocío Titiunik

Comments, questions and suggestions:

Background references:

Robust Nonparametric Confidence Intervals for Regression-Discontinuity Designs, Econometrica 82(6): 2295–2326, November 2014. [Supplemental Appendix]

Optimal Data-Driven Regression Discontinuity PlotsJournal of the American Statistical Associationforthcoming. [Supplemental Appendix]

rdrobust: An R Package for Robust Inference in Regression-Discontinuity Designs, R Journal, conditionally accepted.

Robust Data-Driven Inference in the Regression-Discontinuity Design, Stata Journal 14(4): 909-946, 4th Quarter 2014.