2017

Matias D. Cattaneo (Michigan), Sebastian Calonico, (University of Miami), Max H. Farrell (Chicago), Michael Jansson (Berkeley), Xinwei Ma (Michigan), Rocio Titiunik (Michigan) and Gonzalo Vazquez-Bare (Michigan) for their work on a group of open source packages (rdrobust, rdlocrand, rddensity, rdpower) for regression discontinuity problems, written for both R and Stata.  Those packages are described and listed here: https://sites.google.com/site/rdpackages/

Selection committee: Arthur Spirling (NYU, Chair), Justin Esarey (Rice) and Jamie Monogan (Georgia)

Citation:

The centrality of Regression Discontinuity Designs (RDD) to the causal inference goals of social science is not in doubt: however, the practical implementation of such approaches can be very challenging, involving technical knowledge of bandwidth selection, randomization methods, appropriate interval estimation and so on.  This elegantly written, NSF supported suite of software allows researchers to focus on the theory and substance of their problem, while relieving them of otherwise complicated programming.  The committee was especially impressed with the options available for systematic neighborhood selection---thus greatly decreasing the possibility of arbitrary 'researcher degrees of freedom' in those choices. Along with its other features, which include extremely helpful plotting routines, this product will aid replication and communication for applied researchers.  The software is open-source in its R version, and is also implemented in the proprietary package, Stata: we therefore anticipate that it will be widely-used, extended and improved upon by others. The committee is unanimous: the RD packages are a worthy winner.