Best practices for inference in regression discontinuity designs

Determine econometric best-practice procedures for conducting inference in regression discontinuity designs, specifying how practitioners should implement methods for valid uncertainty quantification at the cutoff in applied settings.

Background

The paper notes that, despite the widespread use of regression discontinuity designs (RDDs), there is no consensus on the most reliable inferential procedures. The authors document in simulations calibrated to prominent applications that commonly used methods can systematically undercover or be excessively conservative, motivating the need for clearer guidance on best practices.

They introduce PLRD as a method that addresses shortcomings observed in existing tools, but the broader question of comprehensive best-practice guidance for RDD inference across diverse empirical contexts remains an explicit concern.