Nonlinear Models for Clustered Data

Chapters 5 and 6 discussed general nonlinear models and generalized linear models (GLM) for independent data. It was emphasized in §6 that GLMs are special cases of nonlinear models where a linear predictor is placed inside a nonlinear function. Chapter 7 digressed from the nonlinear model theme by introducing linear models for clustered data where the observations within a cluster are possibly correlated. To capture cluster-to-cluster as well as within-cluster variability we appealed to the idea of randomly varying cluster effects which gave rise to the Laird-Ware model

Y X Z b e .