MIRDD: An R package for multiple imputation regression discontinuity design

R package MIRDD provides a novel framework for the regression discontinuity design (RDD) by reinterpreting the estimation of treatment effects as a missing data problem. While standard RDD relies on local polynomial extrapolation at the cutoff, MIRDD uses multiple imputation (MI) to estimate unobserved counterfactual potential outcomes within a localized neighborhood. By integrating MI (R package Amelia) and RDD (R package rdrobust), the MIRDD package enables researchers to directly compare traditional RDD results with imputation-based estimates. The package features a comprehensive suite of diagnostic plots to visualize counterfactual distributions and regression stability, offering an essential tool for robust sensitivity analysis in causal inference.