How to run a correlation analysis between a cluster solution and a variable?

Question

I've seen a professor illustrated a couple of analysis in SPSS, and I need to preform the same analysis deriving the same results, but don't know how; Could you advice on how to perform the following analysis (which functions or steps in SPSS) ?

1) After performing cluster analysis (using Ward's method), an output table showed the frequency of each clusters, cluster 1 has X numbers, cluster 2 has Y numbers.

2) The more difficult part is, he used the cluster solution to perform R square correlation and pearson correlation; in the pearson correlation analysis, he used "Ward" (the cluster solution which was based on few variables) as independent factor, and another variable as the dependent factor. I cannot find anywhere to include the "Ward" as a factor, and cannot see either where to include dependent and independent variables in pearson correlation.

3) The same goes for R square correlation, the "Ward" serves as a factor.

Answer 1

Ward's method is available for Hierarchical Clustering. You can save the cluster assignment via the Save dialog box. You can then do any analysis you want using that generated variable (or variables if you save a range of solutions).

But it doesn't make sense to use the cluster number, which is an arbitrary nominal variable, in a linear relationship. You can compute a correlation in the usual way, but it would be meaningless. Try doing a boxplot with cluster number as the grouping variable and a continuous variable as the other variable. You could also do a one-way anova with it.