Online Program Home
  My Program

Abstract Details

Activity Number: 270 - Intersection of Econometrics and Biometrics in Making Policy and Treatment Determinations
Type: Invited
Date/Time: Tuesday, August 1, 2017 : 8:30 AM to 10:20 AM
Sponsor: Business and Economic Statistics Section
Abstract #322186
Title: Learning heterogeneity in causal inference using sufficient dimension reduction
Author(s): Wenbo Wu* and Wei Luo and Yeying Zhu
Companies: University of Oregon and Baruch College and University of Waterloo
Keywords: Causal inference ; Central causal effect subspace ; Heterogeneity ; Model-free ; Sufficient dimension reduction

Often the research interest in causal inference is to see how the covariates affect the mean difference in the potential outcomes. In this paper, we use sufficient dimension reduction to estimate a lower dimensional linear combination of the covariates that are sufficient for this purpose. To enhance interpretability of the results, we further modify the estimator using sparse sufficient dimension reduction, which selects an active set of covariates for variable selection as a by-product. The estimator can also be used to test the heterogeneity of the causal effect. Compared to the existing methods, our approach is model-free, and avoids separate regression modeling in different treatment groups. Thus it can be more applicable and effective. These advantages are supported by both simulation studies and a real data example.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2017 program

Copyright © American Statistical Association