报告题目：Semiparametric estimation for average causal effects usingpropensity score-based spline
报告人：童行伟（北京师范大学统计学院）
报告时间：2023年6月22日（周四）上午10:00-11:00
报告地点：腾讯会议号：678 567 152，会议密码 3207
报告摘要: When estimating the average causal effect in observational studies, researchers have totackle both self-selection of treatment and outcome modeling. This is difficult becausethe parametric form of the outcome model is often unknown and there exists a largenumber of covariates. In this work, we present a semiparametric strategy for estimatingthe average causal effect by regressing on the propensity score. Furthermore, we showthat regression error terms usually depend on the propensity score as well, which couldcause heteroscedastic variances, and thus construct a refined estimator to improve theestimation efficiency. Both estimators are shown to be consistent and asymptoticallynormally distributed, with the latter one having a smaller asymptotic variance. Thesimulation studies indicate that our methods compare favorably with many competingestimators. Our methods are easy to implement and avoid hazardous impact due toextreme weights as often seen in weighting estimators. They can also be extended tohandle subgroup effects with known structure. We apply the proposed methods to datafrom the Ohio Medicaid Assessment Survey 2012, estimating the effect of having healthinsurance on self-reported health status for a population with subsidized insurance planchoices under the Affordable Care Act.
专家简介：童行伟，北京师范大学教授，博士生导师，目前担任北京师范大学数理统计系系主任、教育心理与数据科学技术与应用广东省普通高校重点实验室主任。博士毕业于北京大学数学科学学院，美国University of Missouri, Columbia 博士后，长期从事生物统计、金融统计、因果分析及稳健统计领域前沿研究。现担任中国因果推断分会常务理事；中国工业统计学教学研究会副会长；“应用概率统计”杂志的编委；国际生物统计学会（International）中国分会常务理事，北京大数据协会副会长等。主持科技部重点研发课题1项，1项国家自然科学重点子课题和多项国家自科面上项目等，在Annals of Statistics, Biometrika, Statistica Sinica等顶尖期刊发表学术论文50余篇。
