1

I guess this is more of a statistical question. I'm using scipy to conduct a non parametric hypothesis test. Distribution below. My question is I got a p-value of almost 0.8. The distributions look almost equal though. The test I'm using is Krusal-Wallis. Anyone know how the p-value is calculated here? I would've expected the distributions to be completely different at a p-value of 0.8 I don't want to just reject the null unless I'm absolutely sure.

enter image description here

user10939484
  • 167
  • 2
  • 13
  • 2
    no that looks right, the closer the distributions are the more p tends to 1. Try it out using exactly the same distributions you should get p == 1 – DrBwts Oct 23 '19 at 22:11
  • I was looking at the example on SciPy's page and just got a little confused. I thought a high p-value rejects the null where it's actually a low p-value that does. Thanks! – user10939484 Oct 23 '19 at 22:27

0 Answers0