Overview
Kung-Yee Liang is a biostatistician (Johns Hopkins University; later National Yang-Ming University, Taiwan) best known, with Scott Zeger, for inventing generalized estimating equations (GEE) — one of the most widely used methods for the regression analysis of longitudinal and clustered data.
Key Contributions / Features
- Generalized estimating equations (Liang–Zeger 1986): The foundational GEE method — extend the GLM score equation with a working correlation structure and a robust sandwich variance, giving consistent marginal-mean coefficient estimates without a full joint distribution.
- SS vs. PA models — Zeger–Liang–Albert (1988): With Zeger and Paul Albert, formalized the subject-specific vs. population-averaged distinction for longitudinal GLMs and the Gaussian-mixing relationship between their coefficients. See Zeger-Liang-Albert (1988) and Generalized Estimating Equations.
- Multivariate categorical GEE (Liang–Zeger–Qaqish 1992): Extended GEE to categorical data and the joint modeling of mean and association.
Related