Regression Discontinuity Designs

Abstract

The Regression Discontinuity (RD) design is one of the most widely used non-experimental methods for causal inference and program evaluation. Over the last two decades, statistical and econometric methods for RD analysis have expanded and matured, and there is now a large number of methodological results for RD identification, estimation, inference, and validation. We offer a curated review of this methodological literature organized around the two most popular frameworks for the analysis and interpretation of RD designs: the continuity framework and the local randomization framework. For each framework, we discuss three main areas: (i) designs and parameters, which focuses on different types of RD settings and treatment effects of interest; (ii) estimation and inference, which presents the most popular methods based on local polynomial regression and analysis of experiments, as well as refinements, extensions and other methods; and (iii) validation and falsification, which summarizes an array of mostly empirical approaches to support the validity of RD designs in practice.

Regression discontinuity designs in agricultural and environmental economics

“…Under clearly defined conditions, this set-up provides a natural experiment, with some similarities to an actual randomised experiment: The units above the cutoff are in the 'treatment group', the units below the cutoff are in the 'control group' (Lee and Lemieux, 2010). Causal identification of treatment effects comes from the discontinuous change in the probability of receiving the treatment whereas all potential confounders are either continuously changing across the cutoff, or do not differ just above and below the cutoff (Cattaneo and Titiunik, 2021).…”

“…These two assumptions lead to two distinct frameworks: The continuity framework and the local randomisation framework(Cattaneo and Titiunik, 2021). We discuss examples for both frameworks in Section 2.Downloaded from https://academic.oup.com/erae/advance-article/doi/10.1093/erae/jbac023/6749527 by ETH Zurich user on 09 October 2022…”

Effect Heterogeneity and the Challenges of Causal Attribution in Regression Discontinuity Designs

“…As such, it makes sense for researchers to approach the assumption of continuity in the same ways they would for the standard RD design, theorizing on the usual concern of sorting across the threshold and applying standard diagnostics (e.g. McCrary, 2008;Cattaneo et al, 2020;Hartman, 2021; for an overview, see Cattaneo and Titiunik, 2021).…”