Regression discontinuity designs: a hands-on guide for practice

Abstract

Regression discontinuity (RD) designs have become increasingly popular in political science, due to their ability to showcase causal effects under weak assumptions. This paper provides an intuition-based guide for the use of the RD in applied research. After an intuitive explanation of how the method works, we provide a checklist that can help researchers understand the main robustness checks they should run, and a quick introduction to software implementing the design. We also provide a list of classic designs and examples of their application in political science. We hope this article can constitute a stepping stone from which researchers interested in RD can jump to more advanced literature; and which makes researchers not interested in implementing RDs better consumers of research employing this design.

Democratic transition and party polarization: A fuzzy regression discontinuity design approach

“…The optimal bandwidth estimator used for the previous calculations has been 0.292 as it is calculated according to the Imbens and Kalyanaraman procedure (2012) using the triangular kernel (Shang et al, 2022). Figure 2 reveals that our results remain to be statistically significant and substantially meaningful for a wide variety of bandwidths well beyond the suggested scope of half and double the optimal amount (Valentim et al, 2021). In addition to applying different bandwidths, we also test our models with different types of kernel which also produce significant and substantial results (see Online Appendix Table 4).…”

“…Broadly speaking the literature identifies two types of strategies for the estimation of the treatment impact: parametric and non-parametric strategies (Jacob et al, 2012). Considering the recent criticisms directed at parametric models in the literature (Gelman and Imbens, 2019;Valentim et al, 2021) and the increase in the ratio of notakers in our data as we move further away from the cutoff, we believe that the non-parametric strategy provides a better fit to our data. While we use the bandwidth procedure identified by Imbens and Kalyanaraman (2012), we report a variety of bandwidths below as it is recommended in the literature (Valentim et al, 2021), taking into consideration the small sample size around the cut-off point.…”

“…The fuzzy RDD has two additional assumptions which can be summarized under the concepts of non-zero probability of treatment around the cutoff and the exclusion assumption (Valentim et al, 2021). The former indicates the relatedness between the EDI score and the democratic transition.…”