0

I have one set of data. I have two procedures for making 1-10 ranking of the data (one was used previously and one is a new procedure). I want to do a hypothesis test to see if the ranking is the same or not.

SciGuyMcQ
  • 113
  • You can't actually test to see if the rankings are the same (see here). You can test if they are different, or if their difference is less than some amount you find negligible, or you can test if their agreement is greater than chance. Are any of those what you want? If #2, you need to specify a level of difference you wouldn't care about. – gung - Reinstate Monica Apr 30 '15 at 15:22
  • This is a good point. Basically, I want to test whether the differences are not based on a chance event (ie. alpha of 0.05). – SciGuyMcQ Apr 30 '15 at 15:30

1 Answers1

1

I would advise you to have a look at one of the follow tests:

  • Wilcoxon signed-rank test
  • Mann–Whitney U test
  • Kruskal–Wallis test

These tests are generally used when testing with rankings.

I'm pretty sure one of them matches your data and does everything you ask for.

Chen Orihara
  • 148