ABSTRACT
This chapter focuses on the analysis of variance. It presents a covariance model is a mixed analysis of variance and linear regression model. Its design matrix can be partitioned into two parts. The first part consists of 0 and 1, which is the design matrix of an analysis of variance model. The second part is similar to the design matrix of a linear regression model which can take any real value. For a covariance model, the primary analysis is the analysis of variance. The purpose of a linear regression model is to remove the variability due to some uncontrollable independent variable. Statistical inference of an analysis of covariance model can be obtained based on statistics obtained from its corresponding analysis of variance model. The chapter analyses the computational procedure for a two-way classification model with a covariate by introducing some useful notations.
