s-news
[Top] [All Lists]

multivariate test for homogeneity of variances

To: s-news@lists.biostat.wustl.edu
Subject: multivariate test for homogeneity of variances
From: Karl "Hülber" <huelber@yahoo.com>
Date: Wed, 15 Feb 2006 11:04:15 -0800 (PST)
Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:Received:Date:From:Subject:To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=dPudmCyfRR8B4LX4hcsmS45bRND+othITxYdJ1u4LcK6N31FruLMJaonwNvm+M+Mkxa+OZEl98fIIdoGyFiFiikNZUHOIugqYgEEugWNXf9CC3oi4FjaVd8ohxyJsKJ3ntqcuZDA/6beKbEtQCu79tNtk7HiLn/4yjhEYxYyxmI= ;
Hi everyone!

One way of testing for homogeneity of variances is to
use the absolut values of residuals (e. g. from a
one-way ANOVA) and use these as response in a further
one-way ANOVA. Significant differencies indicate
non-constant variances (Levene test).
Is this also possible in a multivariate design like
one-way MANOVA. Are there better solutions to the
problem?

Thanks for your help
Karl

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

<Prev in Thread] Current Thread [Next in Thread>