Longitudinal Data III: Generalized Estimating Equations and Generalized Mixed Models: Treating Toenail Infection
The longitudinal data generated by the clinical trial of treatments for toenail infection differ in an important respect from the longitudinal data considered in the previous chapter, namely that the response variable is dichotomous making models assuming normality for this variable of little use. In the linear mixed models for Gaussian responses described in Chapter 13, estimation of the regression parameters linking explanatory variables to the response variable and their standard errors needed to take account of the correlational structure of the data, but their interpretation could be undertaken independent of this structure. When modelling non-normal responses, this independence of estimation and interpretation no longer holds. Different assumptions about how the correlations are generated can lead to regression coefficients with different interpretations. The essential difference is between marginal models and conditional models.