Valid Instrumental Variable Selection for Causal Analysis

• In Event: Addressing Challenges in Causality: Free Will, Mediation and Latent Factors, Heterogeneity, and Variable Selection

Abstract

Valid instrumental variable (IV) selection method has attracted attention in biostatistics and epidemiology, particularly for causal inference in high-dimensional and cross-sectional datasets. This method reduces reliance on costly longitudinal studies by identifying valid instruments directly from available datasets. While recent approaches like agglomerative hierarchical clustering under the family plurality rule address multiple endogenous variables and heterogeneous treatment effects, they are computationally intensive and limited in interpreting non-categorical heterogeneity. This study introduces Lasso under the control function approach to handle multiple endogenous regressors, non-normal endogenous variables, and complex treatment heterogeneity via interactions and nonlinear terms.

Authors

• Jiachen Liu, Michigan State University

• Yuehua Cui, Michigan State University

• Haolei Weng, Michigan State University

• William H. Schmidt, Michigan State University