Matching for causal inference
Holder： Fang Han（University of Washington, Seattle）
Time：2025-08-29 10:30-11:30
Location：Siyuan Lecture Hall，Zhi Hua Building-225
Abstract:
In two landmark Econometrica papers, Abadie and Imbens established that the nearest neighbor (NN) matching estimator for the average treatment effect is asymptotically normal when using a fixed number of NNs, but it remains semiparametrically inefficient and bootstrap inconsistent. In this talk, I will demonstrate that these limitations can be overcome by simply allowing the number of NNs to grow with the sample size. Under this modification, the NN matching estimator becomes asymptotically normal, doubly robust, semiparametrically efficient, and bootstrap consistent. These results are published in Econometrica and other leading journals.
About the Speaker:
Dr. Fang Han is a full professor and as the Job & Gertrude Tamaki Endowed Professor in Statistics, in economics (by courtesy) at the University of Washington, and an affiliated investigator in Fred Hutchinson Cancer Research Center. He obtained his Ph.D. (Biostatistics) from Johns Hopkins University in 2015. Previously, he received his B.S. (Mathematics) from Peking University and M.S. (Biostatistics) from University of Minnesota. Dr. Han is a current associate editor for Bernoulli and an editorial advisory board member of Dependence Modeling.
Your participation is warmly welcomed!
欢迎扫码关注北大统计科学中心公众号，了解更多讲座信息!
