11-04-2011 02:22 PM
PROC CORR and PROC FREQ give equivalent point estimates for Kendall's tau-b, but I can't find anywhere in the documentation that explains why the aysmptotic standard errors and p-values would be different. Can someone point me to an explanation for this? I have a very small (n=7) data set but can't get the two procedures to agree. Any help would be greatly appreciated.