Several months ago I performed a Kruskall-wallis test and a dwass steel critchlow-fligner test on some data. A few days ago, I wanted to check those results, and I conducted the exact sames tests on the exact same data (as we can see on the picture the descriptive stats are exactly the same).
The weird thing is that I got different results. The statistics are the same, but not the indices of significance. Hopefully, I had a screenshot of my old results to compare. As you can see on the picture, the diffrence between 1-3 and 2-3 is significant on the old results, but not on the new one.
How can you explain this difference ?
Thank you for your help.
