The Minimax Concave Penalty Variable selection regularization
method for Regression Discontinuity Designs
Bahr kadhim Mohammed
Ashwaq Abdul Sada Kadhim
Al-Qadisiyah University – College of Administration and Economics
Corresponding Author: Ashwaq Abdul Sada Kadhim
Abstract : The classical method faced a big problem with estimating and selecting important variables when the
dataset has a cut-off point. Therefore, we propose a new method to solve these problems. In this paper we suggested a
new approach by combining the Regression Discontinuity Designs (RDD) with the Minimax Concave Penalty (MCP)
method. Local linear regression (LLR) method was used to estimate the effect of processing on the cut-off region of
the observations within the optimum bandwidth selection for the RDD design to obtain the best model. Three models
were used to determine the IK (Iembens and kalyanman) bandwidth, cross-validation (CV) method, and The CCT
(Calonico, Cattaneo & Titiunik) bandwidth. A simulation study and real data are conducted to investigate the
performance of the proposed method. The mean squared errors (MSE) is used to choose the best model.
Keywords: Regression Discontinuity Designs (RDD), Minimax Concave Penalty (MCP), variable selection, Local
linear regression, bandwidth selection, IK, CV, CCT.
