Generalized Estimating Equations (GEE)

This session will extend the material from the Mixed Models training on June 25 (available at    • Mixed Models  ) to outcomes that are not normally distributed, e.g., count and binary outcomes. To recap, mixed models are one method for analyzing correlated data. This can include longitudinal data, when multiple observations are taken on the same unit over time; hierarchical data, such as when multiple patients are seen by the same doctor; and other forms of grouped data. We will cover implementation in R with a real data set and enough theory to guide decision-making.