ABSTRACT
We now turn our attention to the study of two-way factorial experiments; that is, experiments where the mean response is a function of two factors: factor A with a levels denoted by i = 1, . . . , a, and factor B with b levels denoted by j = 1, . . . , b. We define the (i, j)th cell to be the set of observations taken at the level i of factor A and the level j of factor B. We assume that n experiments (called replicates) have been performed at each possible combination of the factor-levels. For instance, a computer engineer might be interested in the execution times (the response variable) of five different workloads (factor A, a = 5) on three different configurations of a processor (factor B, b = 3). The goal here is to study how the processor performance is affected by the different processor configurations.
