Date:         Fri, 13 Dec 2002 16:31:48 -0500
Reply-To:     Art@DrKendall.org
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         "Arthur J. Kendall" <Art@DrKendall.org>
Organization: Social Research Consultants
Subject:      Re: MANOVA: post-hoc testing
Comments: To: Joel Snyder <jss47@cornell.edu>
Content-Type: text/plain; charset=us-ascii; format=flowed

It has been a while since I used MANOVA, but I believe it has a MATRIX
out sub command.

A kludge would be to multiply all DVs by 1000.


see /posthoc= under the GLM procedure


Art

Joel Snyder wrote:
> I am trying to do post-hoc tests with MANOVA in SPSS 10 and can't figure
> out what the best option is. My design has three repeated measure factors,
> one of which is nested in another. Here is the basic syntax I am using:
>
> MANOVA TO111 TO TO338
> /WSFACTOR=ST(2) VER(3) EXC(4)
> /WSDESIGN=ST VER EXC W ST  ST*VER VER*EXC W ST
> /MEASURE=TWOONE.
>
> This gives me all of the main effects and interactions specified but I now
> would like to test for differences in means between levels of my factors
> with >2 levels (e.g., VER and EXC). Is this possible using MANOVA?
>
> Alternatively, is it possible to get a precise estimate of the MS error
> values from the output. As it is now, I can only get two decimal places.
> This prevents me from using the MS error value to calculate my own post-hoc
> tests (e.g., Bonferonni).
>
> Thanks in advance,
> Joel
>
>
> -----------------------------------------------------------------------
> Joel Snyder
> Cornell University
> Uris Hall
> Dept. of Psychology
> Ithaca, NY 14853
> USA
>
> (607) 255-6398 (o)   (607) 255-8433 (fax)
> email: jss47@cornell.edu
> http://www.people.cornell.edu/pages/jss47/
>