Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | "Carlo Lazzaro" <carlo.lazzaro@tin.it> |
To | <statalist@hsphsun2.harvard.edu> |
Subject | R: st: Two factor anova post test |
Date | Tue, 19 Jul 2011 13:05:43 +0200 |
Dear Ricardo, why not performing a one-way ANOVA on the significant factor (race) and, if appropriate, perform a multiple comparison with correction? I agree with Angel that the risk of inflating Type 1 error is probably negligible. Anyway, you can explore in depth this issue with different critical values. Kindest Regards, Carlo -----Messaggio originale----- Da: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] Per conto di Ricardo Ovaldia Inviato: martedì 19 luglio 2011 12.19 A: statalist@hsphsun2.harvard.edu Oggetto: Re: st: Two factor anova post test Thank you Angel. The term that is significant is race, with three levels, therefore I am not completely sure which means are different, and I do not understand how to use -test-, or even if that is the correct way to compare them. Thank you, Ricardo. Ricardo Ovaldia, MS Statistician Oklahoma City, OK --- On Mon, 7/18/11, Ángel Rodríguez Laso <angelrlaso@gmail.com> wrote: > From: Ángel Rodríguez Laso <angelrlaso@gmail.com> > Subject: Re: st: Two factor anova post test > To: statalist@hsphsun2.harvard.edu > Date: Monday, July 18, 2011, 4:05 PM > Dear Ricardo; > > If you were having many factors and many interactions, you > should be > worried about and incresed alpha error level for the > F-test, but with > 2 factors I doubt this is a problem. Moreover, your p level > for the > significance of sex is far away from the conventional 0.05, > while the > interaction sex*race is non-significant. Therefore you can > conclude > that there is a significant difference only between sexes. > Because > there are only two levels, there is no need to carry out > any posthoc > test like Scheffe, because you already know which means are > different, > those of men and women. > > Best regards, > > Angel Rodriguez-Laso > > 2011/7/18 Ricardo Ovaldia <ovaldia@yahoo.com>: > > Hello, > > > > I performed a two factor ANOVA; one factor (sex) has 2 > levels and the other factor (race) has 3 levels. The > interaction term is not significant but one of the factors, > race, is (p=0.0004). How can I determine which means are > different while maintaining the experiment wise type I error > rate. > > > > If I had only one factor, I could use -oneway- and > then use, for example, scheffe for multiple comparisons. But > not sure what to do with two factors. > > > > Thank you in advance, > > Ricardo > > > > Ricardo Ovaldia, MS > > Statistician > > Oklahoma City, OK > > > > > > * > > * For searches and help try: > > * http://www.stata.com/help.cgi?search > > * http://www.stata.com/support/statalist/faq > > * http://www.ats.ucla.edu/stat/stata/ > > > > * > * For searches and help try: > * http://www.stata.com/help.cgi?search > * http://www.stata.com/support/statalist/faq > * http://www.ats.ucla.edu/stat/stata/ > * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/