ABSTRACT
In the general case, there will be m random samples from different groups, with sizes n1, n2, … nm, and values will be available for p variables X1, X2, …, Xp for each sample member. Thus, the data for a discriminant function analysis takes the form shown in Table 8.1. The data for a discriminant function analysis do not need to be standardized to have zero means and unit variances prior to the start of the analysis, as is usual with principal components and factor analysis. This is because the outcome of a discriminant function analysis is not affected in any important way by the scaling of individual variables.
