@JeffMeyers  Yes you are right. In my case the exact p-value of the Kruskal-Wallis test was not in the output, because my group variable had only two categories. Then only the assymptotic p-value P_KW is calculated. Only if I the variable has more than two categories the exact Kruskal-Wallis p-value XP_KW is calculated and added to the temporary dataset _num_p1e.       If the group variable has only two categories, the p-value XP2_WIL of the exact Wilcoxon Test is added to the temporary dataset _num_p1e. The assymptotic Wilcoxon p-value is P2_WIL. I think the solution is to add a condition when you want the exact Kruskal-Wallis test and use the correct variable names for the exact Wilcoxon and Kruskal-Wallis test p-values:       %if &_nexact=1 %then %do; %if &nby > 2 %then %do; when 2 then p1e.xp_kw %end; %if &nby=2 and &_nbymiss=0 %then %do; when 4 then p1e.xp2_wil %end; %end;     ... View more

Thank you. I have looked in the tempory dataset num_p1e. Both variables exist in the temporary data.    Therefore, I am surprised by the error. ... View more

Hello @JeffMeyers .Thank you very much. I had already found that section.   However, when I change the values there, the MACRO no longer runs and I get the follwoing error messages:  In the SAS dataset _num_p1e the variables exists. I would be grateful if you could provide further suggestions as to what the errors might be. ... View more

Error in the selection of the correct exact p-values   Hello everyone,   I have just discovered that when calculating the exact Wilcoxon rank test (P_VALS=4 for continuous variables), unfortunately the exact p-value is not displayed. The asymptotic two-tailed p-value P2_WIL is selected instead of the exact p-value, which is called XP2_WIL in the generated data set of the npar1way procedure. The same applies when calculating the exact Kruskal-Wallis test. The asymptotic p-value P_KW is also selected, rather than the exact p-value XP_KW. It would probably be necessary to check whether the other exact p-values have been selected correctly.  Unfortunately, I can't find all the places in MACRO where something needs to be changed so that the entire MACRO runs.   It would be nice, if the MACRO could be corrected or if anybody can say where to correct the MACRO. Thanks! ... View more

Thank you for your answer @JeffMeyers. But is it possible that you can update the post again in the menu and then just click save? Thanks again for your fantastic macro! ... View more

Hello @JeffMeyers. Are there any news about the update yet? When is the new version of the macro available? It would be really nice if you could make the macro available soon. Many many thanks! ... View more

Hello @JeffMeyers, I have a problem with the macro variable BY_INCMISS. I have missing values in the BY variable (3 categories). I do not want to take the missing category into account when calculating the p-values of the Kruskal-Wallis test. However, no matter how I set BY_INCMISS, the missing category is always taken into account. I have an additional ROWBY variable. The same occurs when I specify an additional COLBY variable.   Do you have any suggestions? Thank you! ... View more