Proc CORR

This procedure helps to find the Pearson correlation coefficient and other measures of association between the variables. After you get the covariance matrix, you can check for the heterogeneity looking at the diagonal values in the covariance matrix. If the values are much different then there is quite an amount of heterogeneity. 

(this can be used when you would like to do repeated measures analysis and would like to know whether to use a compound symmetry matrix using say GLM procedure or an unstructured variance/covariance matrix, or use an autoregressive structure by looking at the coefficient table for the repeated observations)

Proc corr data = dataname1 

                  outp = dataname2 

                  cov 

                  vardef = n 

                  plot = scatter 

                  best = 2 

                  fisher (alpha = 0.1 type= twosided) 

                  nocorr 

                  noprob

                  nosimple

                  noprint;

                  var var2 var3 var4 

                  by var1; 

run;