안녕하세요.
5월 22일 금요일 오후 2시, 221동 110호에서 보건통계학 연자 초청 세미나가 진행됩니다.
관심 있는 분의 많은 참여 부탁드립니다.

○ 주제: Evaluating policy interventions using observational data
○ 연사: 이유진 교수 (Department of Biostatistics, Brown University)
○ 일시: 2026년 05월 22일(금) 오후 2시 - 4시
○ 장소: 서울대학교 보건대학원 221동 110호
○ 방식: 현장 참여
○ 문의: museum03@snu.ac.kr

 Abstract: This seminar introduces causal inference methods for evaluating policy interventions using observational data. Policy interventions are often implemented at the population or community level, making standard comparisons between exposed and unexposed groups potentially misleading because of confounding, time trends, and structural differences across populations. The seminar will discuss why different causal inference frameworks are needed depending on the study design and data structure. We begin with an introduction to difference-in-differences (DiD) methods, focusing on their role in estimating policy effects by comparing temporal changes between intervention and comparison groups. The discussion will highlight important distinctions between panel data and repeated cross-sectional studies and introduce different causal estimation approaches. The seminar will then introduce alternative approaches, including synthetic control methods, which construct data-driven comparison groups when suitable controls are not directly available. We will also discuss regression discontinuity designs and their applications in settings where treatment assignment is determined by cutoff-based policies or eligibility criteria. The seminar concludes with a broader discussion of practical considerations in policy evaluation, including study design, identifying assumptions, interpretation of causal effects, and challenges arising in real-world observational studies.