In most experiments, there is structure to the treatments that are applied. In factorial experiments, this structure consists of the separate factors and their interactions. However even in experiments with a single factor, the factor levels may have some structure. If an analysis of variance test concludes that a factor affects the response, it is possible to further investigate its effect, guided by what is known about the structure of the factor levels.
The factors in an experiment are variables that are controlled by the researcher. Sometimes additional information is obtained from the experimental units that is uncontrolled but helps to explain differences between them. Such variables are called covariates and are the subject of the last section in this chapter.