Re: NPAR TESTS

On Sunday 28 March 2004 23:44, Ben Pfaff wrote:
> Can you think of a few of the others off-hand?  I like to try to
> look forward when I write code, so that I can generalize as much
> as necessary (but not more).

Off the top of my head, I'd say there's only Mann-Whitney U and 
Kruskal-Wallis.  W. S.-R. doesn't work that way -- I was incorrect: you only 
have to rank the *differences* between pairs of scores within cases.  I was 
actually thinking of the M.-W. U at the time and I stopped trying to code 
when I realised that assigning ranks was more work than I thought.  

However, about half of them require ranks of some kind, including Spearman 
which is in a different procedure.  Maybe some others too.  I can't think of 
any other odd twists to it though.

- M.