Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
nonparametric stats
Message posted by Holly (via 24.24.8.30) on February 24, 2001 at 4:05 PM (ET)
I have three independent samples and am interested in knowing whether they differ on 5 different dependent variables, all of which are nominal. I would like to be able to do simple contrasts. Which nonparametric stat is most appropriate for this data?
Thanks!
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: nonparametric stats
Message posted by Bill (via 192.231.71.108) on February 26, 2001 at 3:21 PM (ET)
Your independent variable is group (with 3 levels i.e., 3 independent samples) and your dependent variable is a nominal scaled variable x5. If you want to test for an association between the IV and DV (x5) (or test for independence) you may use a chi-square test. Most texts do not suggest how to do specific comparisons following a significant chi-square (if that's what you want). The only reference I've seen is to a Ryans procedure in a book that is now out-of-print (Linton and Gallo, The Practical Statistician, 1975, Brooks-Cole). If you can't find it and can't live without it then send the web-master an email and ask him to forward it to me (I hope he won't mind). I'll send you a copy of that portion of the text. Good luck.
Re: nonparametric stats
Message posted by Holly (via 24.24.8.30) on March 9, 2001 at 10:45 PM (ET)
Hi... Thanks so much for your clear and helpful answer. I would be more than thrilled if you were to send me portions of the relevant text on Ryan's procedure. You can use this address:
holly@twcny.rr.com
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.