Mixed Effects Models vs Hierarchical Models

This chapter explains why it is necessary 1) to extend linear models, used to describe the variation in a sequence of observations, to nonlinear models, useful to represent numerous biological and physical phenomena, 2) to extend these models to (linear and nonlinear) mixed effects models in order to simultaneously model several sequences of observations coming from different individuals, and 3) to extend these models for continuous data to generalized models for discrete data. We will then consider hierarchical models, which allow us to systematically adopt one standard methodology for the entire population approach, no matter the model or data type.