The analysis of variance

Abstract

In earlier chapters we have defined the model linear in the parameters, commonly called the linear model, and discussed the theory of estimating parameters of the model by the method of least squares. Our applications of the theory so far have been to simple linear regression, polynomial regression and multiple regression. However, there is a very wide class of models which can be analysed on the basis of the theory we have covered. In this chapter we apply our theory to data arranged in balanced arrays or cross classifications; it is part of a very big topic usually referred to as ‘the analysis of variance’. The following example shows the type of problem we shall be dealing with.