Is it possible to obtain partial correlation coefficients in SAS surveyreg? If so, how? The model statement is not allowing pcorr. Is there a back door or is it just wrong to request partial correlations in SAS Survey?
Well, I wouldn't call it wrong, but I would call it difficult to do, I think you would have to do the heavy lifting in PROC IML to get the weighted quadratic form matrix, and then do the reduction to partial out each variable. I don't see any way to do it solely in SURVEYREG--you could export the X'WX matrix to IML and calculate partial correlations there.
ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.