Multilevel models

These are models for grouped data in which datapoints may be correlated, and there is a need to treat variance differently among the groups. They are variously called multilevel, mixed, or heirarchical/ models.

Suitable for longitudinal data in which one is interested in estimating coefficients for groups or individuals within the population, rather than for the population at large.

There is a chapter (10) in Ben Bolker's book about them.

A few resources: