Date:         Mon, 11 May 1998 15:52:31 -0500
Reply-To:     "Nichols, David" <nichols@SPSS.COM>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         "Nichols, David" <nichols@SPSS.COM>
Subject:      Re: Post-hoc comparisons in RM ANCOVA
Comments: To: rfoster <rfoster@LOON.NORLINK.NET>

You want to be comparing the estimated/adjusted means under the model
rather than the observed means. If you take out the covariate and fit a
full factorial model (or have balanced data) they'll be the same.
 
David Nichols
Principal Support Statistician and
Manager of Statistical Support
SPSS Inc.
nichols@spss.com
 
----------
From:  rfoster [SMTP:rfoster@LOON.NORLINK.NET]
Sent:  Monday, May 11, 1998 3:29 PM
To:  SPSSX-L@UGA.CC.UGA.EDU
Subject:  Post-hoc comparisons in RM ANCOVA
 
I am doing a repeated measures ANOVA, with one within factor (4 levels)
and
two between factors (4 levels each), and one covariate.  I used EMMEANS
COMPARE (with command language) to examine simple main effects.  Is there
any way to compare observed (rather than estimated) marginal means?  Or is
it really the estimated means that I should be comparing to see where
significant differences lie?
 
Thanks for any help.
 
Rob Foster