Thanks for help and advices. Gabriel - I also thought about wilcoxon test, but there is one problem with this.. my groups have not equal number of observations. Does this test is good option in the fact of that? Cheers, Jakub