Date: Wed, 12 Apr 2006 16:03:43 -0500
Reply-To: Anthony Babinec <email@example.com>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Anthony Babinec <firstname.lastname@example.org>
Subject: Re: MANOVA and Discriminant Analysis
Content-Type: text/plain; charset="us-ascii"
Sounds like you're on the right track.
Huberty is the author of a nice text on applied
discriminant analysis, which is evidently about
to come out in a second edition titled "Applied
MANOVA and Discriminant Analysis," co-authored with
Stephen Olejnik. Huberty draws the distinction
between descriptive and predictive discriminant
analysis, and you're interested in the former.
James Stevens' book on Applied Multivariate Analysis,
now in its 4th edition, is another good reference.
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of
Sent: Wednesday, April 12, 2006 2:02 PM
Subject: MANOVA and Discriminant Analysis
I've been doing some reading (Huberty & Morris, 1989; Bray & Maxwell,
1982) trying to understand how to best understand the results of an
MANOVA. It seems that one is well advised to do a discriminant function
analysis (DFA) after a significant MANOVA to better understand the
nature of the variate that is created in MANOVA. This is rarely done,
but it seems necessary to be able to interpret MANOVA results. If I do
a dfa after my manova what can I do with those results? For example, I
have a situation where my MANOVA consists of 11 DVs. I have a
significant between subjects effect for gender. Then when I look at the
DFA it appears that only 6 of the 11 DVs are really loading on the
variate. What do I do with this information? Does this lead me to
reanalyze the data with only 6 DVs? Or does this information only tell
me how to interpret the variate on which males and females differ? Any
refs advice, much appreciated.
Matthew Pirritano, Ph.D.
National Science Foundation Post-Doctoral Fellow
College of Education
Department of Individual, Family & Community Education
1 University of New Mexico
Albuquerque, NM 87131-0001