Jing Tao - University of Washington
(RESCHEDULED TO SPRING 2024) Inference for Heterogeneous Quantile Treatment Effects in Observational Studies with High-Dimensional Covariates
Date: 05/11/2023 (Thu)
Time: 3:30pm- 5:00pm
Location: Seminar will be held on-site: Social Sciences room 113
Organizer: Adam Rosen
Meeting Schedule: Login or email the organizer to schedule a meeting.
All meetings will be held in the same location as the seminar unless otherwise noted.
9:30am - Adam Rosen (pick-up from 21c)
11:30am - Michael Pollmann
12:00pm - Lunch w/Matt, Michael, Adam, ...
1:00pm - Matt Masten
1:30pm - Xinyue Bei
3:00pm - Seminar preparation
3:30pm - Seminar Presentation (3:30pm to 5:00pm)
6:00pm - Dinner at Fairview Dining Room with Adam, Matt, Michael, ...
Additional Comments: Abstract: In this study, we investigate heterogeneous local quantile treatment effects for observational data with high-dimensional covariates, without relying on the strong ignorability assumption. Using a binary instrumental variable, we identify parameters of interest in a population subgroup (compliers) through a two-stage regression model. We develop Lasso estimation with a non-convex objective function to estimate these parameters and propose a de-sparsifying estimator for both pointwise and uniform inference. Moreover, we obtain uniform strong approximations to the local quantile treatment coefficient process by conditionally pivotal and Gaussian processes. Based on these strong approximations, we develop bootstrap resampling methods that can be used for constructing uniform confidence bands for the heterogeneous/conditional local quantile treatment effects given high-dimensional covariates. The proposed approach works for both continuous and categorical response variables under the framework of generalized linear models. We derive theoretical properties, evaluate performance through simulation studies, and apply our method to real data from the Oregon Health Insurance Experiment.