Activity Number:
|
323
- Estimating Treatment Effects: Applications in Health Policy
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 5, 2020 : 10:00 AM to 2:00 PM
|
Sponsor:
|
Health Policy Statistics Section
|
Abstract #313784
|
|
Title:
|
Estimating Causal Effects in Observational Studies: A Comparison of the Use of Boosting for Modeling the Assignment Mechanism, the Response Surface, or Both
|
Author(s):
|
Daniela Golinelli* and Greg Ridgeway
|
Companies:
|
Mathematica and University of Pennsylvania
|
Keywords:
|
Causal inference;
Observational studies;
boosting;
bart
|
Abstract:
|
Several authors have shown that modeling well the response surface using non-parametric approaches produces better treatment effect estimates than those obtained by modeling the treatment assignment mechanism. With this paper we investigate the use of boosting in modeling the assignment mechanism, the response surface, and/or both in a doubly robust setting. We use the same simulation set-ups that have appeared in the literature to assess and compare the performance of the various boosting approaches to more traditional approaches (such as linear regression for modeling the response surface) and Bayesian Additive Regression Trees.
|
Authors who are presenting talks have a * after their name.