2015.12.22
2015年12月19日に久留米大学バイオ統計センター公開セミナーを開催しました。
バイオ統計センター公開セミナー
日時：2015年12月19日（土）15:00-17:00
場所：筑水会館 1階 中会議室
15:00-16:15 講演者１：Xiao-Hua Andrew Zhou (Professor, Department of Biostatistics, University of Washington Associate Director, U.S. National Alzheimer’s Disease Coordinating Center) 演題：「Propensity score based subclassification estimators for the average causal effect」 概要： Inferring the causal effect from observational studies is challenging due to difference in pre-treatment variables. Subclassification on quintiles of estimated propensity scores is often used in practice to balance observed covariates. However, residual confounding of the subclassification estimator with a fixed number of subclasses remains even with an infinite sample size. Here we consider increasing the number of subclasses with sample size and show that this enables consistently estimating the average causal effect via subclassification at a square root convergence rate. This is a joint work with Linbo Wang.
16:30-17:00 講演者２：Satoshi Hattori (Professor, Biostatistics Center, Kurume University) 演題：「Stratified doubly robust estimators for the average causal effect」 概要： Doubly robust estimators are being of great interest in estimating the average causal effect. In this talk, we discuss a hybrid estimator of the doubly robust estimator and the propensity score stratification. The new estimator allows one more propensity score model than the doubly robust estimator and this addition introduces further robustness to the doubly robust estimator. We discuss theoretical and numerical properties and issues in the new estimator. This is a join work with Masayuki Henmi.
場所：筑水会館 1階 中会議室
15:00-16:15 講演者１：Xiao-Hua Andrew Zhou (Professor, Department of Biostatistics, University of Washington Associate Director, U.S. National Alzheimer’s Disease Coordinating Center) 演題：「Propensity score based subclassification estimators for the average causal effect」 概要： Inferring the causal effect from observational studies is challenging due to difference in pre-treatment variables. Subclassification on quintiles of estimated propensity scores is often used in practice to balance observed covariates. However, residual confounding of the subclassification estimator with a fixed number of subclasses remains even with an infinite sample size. Here we consider increasing the number of subclasses with sample size and show that this enables consistently estimating the average causal effect via subclassification at a square root convergence rate. This is a joint work with Linbo Wang.
16:30-17:00 講演者２：Satoshi Hattori (Professor, Biostatistics Center, Kurume University) 演題：「Stratified doubly robust estimators for the average causal effect」 概要： Doubly robust estimators are being of great interest in estimating the average causal effect. In this talk, we discuss a hybrid estimator of the doubly robust estimator and the propensity score stratification. The new estimator allows one more propensity score model than the doubly robust estimator and this addition introduces further robustness to the doubly robust estimator. We discuss theoretical and numerical properties and issues in the new estimator. This is a join work with Masayuki Henmi.
