Confidence Interval for Wilcoxon Rank Sum

[hsaul]

Hi All,

I am wondering what the best approach would be to calculate p-values and 95% CIs for the Wilcoxon rank sum test on ordinal data. I'm looking into the QUANTREG and NPAR1WAY (with the wilcoxon option) procedures, but not sure which is best for CIs. The latter uses the Hodges-Lehmann estimation method. Any help would be appreciated!

[Ksharp]

Bootstrap method ? @Rick_SAS  wrote many blogs about this kind of method.