★摘要：The causal inference framework has risen to be a widely accepted framework for evidence generation in the field of epidemiology and public health. Under the framework, response surface modeling (i.e., outcome regression) and propensity score methods are two mainstay approaches to causal effect estimation. However, the assumptions underpinning the validity of these two approaches are often overlooked. In this talk, we will give a brief overview to both approaches, how they can be combined into doubly robustness estimators aiming for robustness to model misspecification and slow convergence rates. We will also discuss the large-sample and finite-sample properties of these estimators and provide suggestions for how to choose between them in practical analysis.

★講者介紹：國立清華大學生命科學暨醫學院施銘杰 助理教授

★專長：因果推論、證據整合

★日期：2024年3月18日（一）

★時間：中午12點入場、12:20PM-13:20PM演講

★地點：公共衛生學院1樓拱北講堂

★報名網址：https://forms.gle/bb46aRNK2UtrwbgNA

★截止時間：3月12日（二）

★本活動中午提供簡易輕食，並請自行攜帶水杯

主辦單位：國立臺灣大學健康資料研究中心

協辦單位：國立臺灣大學公共衛生學院、臺大公共衛生學院流行病學與預防醫學研究所