Output From DFA Advanced
Use the option that lets you input the group membership. If you have 4 clusters label them 1, 2, 3, and 4. Prepare a list of sample numbers in each cluster.
The program DFA Advanced will allow you to perform two major types of analyses. First, perform a Full-Model which will make use of all of the variables. How successful are the three variables in assigning these samples? Closely examine the individual assignments. Is there a pattern? Second, use the step-wise model in which each variable is added, one at a time. Is there a smaller number of variables which produce an efficient discriminant function. What variables are included?
Interpret your cluster assemblages using the results of DFA.