@ssch, As far as I understand DistributionFitTest gives you back the p-value from the test which is going to vary for each set but most of the results of the actual test are what you expect sets=100;size=10000;data=RandomVariate[NormalDistribution[x,{sets,size}];Tally[??????(DistributionFitTest[#,Automatic,ShortTestConclusion] & /@ data)]