讲座时间：6月10日9:00-12:00

讲座地点：科学园2H425

讲座主题：Causal Inference with Observational Data: A Comprehensive Guide to Quasi-Experiments（观察性数据的因果推断：准实验方法全面指南）

讲座嘉宾简介

Hongfei Li is an Assistant Professor in the Department of Decisions, Operations and Technology (DOT) at The Chinese University of Hong Kong (CUHK) Business School. Before joining CUHK, he received his PhD from the School of Business at the University of Connecticut and his BS and MS from Renmin University of China in Beijing. His current research focuses on business analytics in emerging online platforms, applications of artificial intelligence and machine learning, and statistical methodology.

Professor Li has extensive teaching experience in the MIS domain such as Business Information Systems, Computer-based Information Systems, Econometric Theory and Application, Advanced MIS Research Seminar, and Database and Big Data Management. He is interested in teaching technical courses related to mathematics, statistics, econometrics, and computer language in business application.

讲座摘要

Quasi-experimental methods are essential tools for causal inference in observational studies, particularly when randomized controlled experiments are not feasible or ethical. This presentation provides a comprehensive introduction to the theory and application of quasi-experimental designs, highlighting their critical role in rigorous empirical research. We begin with a fundamental discussion of counterfactual reasoning, clarifying key concepts such as treatment and control groups, and addressing common misconceptions about causality. Following this foundational overview, the presentation systematically explores several prominent quasi-experimental approaches, including multiple regression, fixed effects models, matching techniques (propensity score matching and coarsened exact matching), regression discontinuity designs (sharp and fuzzy), difference-in-differences, and the synthetic control method. Each method is illustrated through practical examples, demonstrating their respective strengths, limitations, and conditions necessary for valid application. Emphasis is placed on understanding critical assumptions and implementing robust empirical strategies, accompanied by guidance on performing sensitivity analyses and robustness checks. Concluding with best practices and methodological advice, this presentation equips researchers with a nuanced understanding of quasi-experimental techniques, fostering informed methodological choices for empirical investigations in business analytics, social sciences, and related disciplines.