Gary King on Simplifying Matching Methods for Causal Inference
作者：Gary King
摘要 / PDF 全文下載
On May 30, 2018, Gary King, the Albert J. Weatherhead III University Professor at Harvard University, gave a speech at the International Conference about Innovations in Political Methodology and China Study, which was held at National Taiwan University. His second keynote speech focused on the benefits of using alternative matching methods for statistical analysis, and the dangers of using the most commonly used matching method, Propensity Score Matching (PSM).
King showed how to use matching in causal inference to reduce model dependence and bias. He introduced two matching methods: Mahalanobis Distance Matching (MDM) and Coarsened Exact Matching (CEM), and explained how these methods are simpler, more powerful, and easier to understand than existing approaches. The discussion that followed addressed some of the concerns with using matching methods, and why King believes matching to be the superior way to preprocess data before running regressions.
This speech was transcribed and edited by Stephen B. Reynolds, a Ph.D. Candidate at the Department of Political Science at National Taiwan University. For more information on Simplifying Matching Methods for Causal Inference, please refer to: GaryKing.org
Gary King on Simplifying Matching Methods for Causal Inference