"Modeling Longitudinal and Panel Data" covers the extension of Generalized Linear Models (GLM) to model varieties of longitudinal and clustered data, called panel data. Specifically, the course treats generalized estimating equations (GEE), a population averaging method that models panel data in which the response is a member of the exponential family of distributions; e.g., continuous, binary, grouped, and count. The instructors will discuss GEE theory, relevant correlation structures, and differences in both theory and application between population averaging GEE (PA-GEE) and random effects or subject specific panel models (SS-GEE). This course covers model construction, how to estimate the equations, different types of models, how to deal with missing data, testing of models, model assumptions, and more. Familiarity with GLM (Generalized Linear Models) is a prerequisite.
Dr. Hilbe and Hardin are the authors of "Generalized Estimating Equations" (Chapman & Hall/CRC). Joseph Hilbe is Emeritus Professor at the University of Hawaii and Solar System Ambassador with NASA's Jet Propulsion Laboratory at California Institute of Technology. He is an elected Fellow of both the American Statistical Association and Royal Statistical Society. Dr. Hardin is on the faculty at the University of S. Carolina, and also on the board of the "Stata Journal."
- Theory and history of GLM
- Development of methods to analyze panel data
- Software used for GEE and related models
- Model Construction and Estimating Equations for Panel data in general and PA-GEE specifically
- Parameterization of the working correlation matrix
- Scale variance estimation
- Alternating logistic regression models
- SS-GEE models (random effect)
- GEE2 models
- Generalized and cumulative logistic regression
- Problems with missing data
- Residual analysis
- Comparative testing of models
- MCAR assumption for PA-GEE models