NFL Forum

Nice work! Testing significance of correlation differences

Posted By: Colin Caster
Date: 19 Aug 01, 1:01 pm

In Response To: 1988 thru 2000 data (docriver)

Testing the significance of the differences in correlations is really tricky. It would be easier if, instead of breaking it apart by year, you aggregate the whole file, producing one r for win-win, and one r for Ew-win. Ordinarily, one can take the difference of the Fisher z-transformed rs, and test it for significance, but you've got a special case in which the two correlations share a variable. There is a procedure for testing the significance of so-called "correlated correlation coefficients," but it's too complicated for me to detail here. I would be happy to test it for you, or if you prefer, I can refer you to a print journal article with the proper formulae.

In summary, though, I don't see any down side to using Ew rather than w, given your demonstration.

Colin

Messages In This Thread

Pythagorean Formula for wins -- docriver -- 16 Aug 01, 12:02 pm

Pythagorean -- Editor -- 16 Aug 01, 1:33 pm

tails? -- docriver -- 16 Aug 01, 1:57 pm

Tails -- Editor -- 16 Aug 01, 3:50 pm

Starting point -- Colin Caster -- 16 Aug 01, 9:40 pm

1988 thru 2000 data -- docriver -- 17 Aug 01, 2:51 pm

also.. -- docriver -- 17 Aug 01, 3:01 pm

Nice work! Testing significance of correlation differences -- Colin Caster -- 19 Aug 01, 1:01 pm

email stats? -- docriver -- 19 Aug 01, 7:18 pm

Sure -- Colin Caster -- 19 Aug 01, 10:13 pm

Results (and computation) -- Colin Caster -- 20 Aug 01, 1:02 pm

Re: Bill James -- T. Hopper -- 19 Aug 01, 2:01 pm

I am told that the exponent for college football is 1.68 (nt) -- Colin Caster -- 19 Aug 01, 10:16 pm

NFL Forum is maintained by Pi Yee Press