General linear models (GLMs) were introduced to describe situations in which a response variable is linearly related to one or more numerical explanatory variables. We have also shown how the same class of GLMs can be used to model:

In this chapter, we show that GLMs can also be used to model the relationship between a response and one or more categorical explanatory variables. These GLMs involve numerical 'explanatory variables' called indicator variables that are not direct measurements but are defined in a way that gives the GLMs useful properties.

Using these extensions, it is largely possible to ignore the distinction between numerical and categorical explanatory variables when modelling a response.