Hi everyone!

I have a data bank of a cross-sectional study where each person has tree measures (same kind of measure in different regions). The analysis was performed with logistic or linear regressions with cluster adjustment (robust SEs). Now, I’m trying to calc the post-hoc power of sample for repeated measures with this sintax:

power repeated, ngroups() varwithin() nrepeated() corr() varerror()
where “corr(numlist): correlation between repeated measures; one of corr() or covmatrix() is required”

However, I found a difficult to obtain a unique value of correlation of the repeated measures. How I can obtain?
I tried a matrix, but I have the correlations of pairwise and not a global value.
Is it possible?
Thank you!