Date:         Sat, 25 Jul 1998 23:38:06 -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: Eta squared
Comments: To: roth jean ellen <j-roth2@UX6.CSO.UIUC.EDU>

The GLM procedure produces only the partial eta^2 measure, which will
only match "regular" eta^2 in a single factor design. The older MANOVA
procedure will give eta^2 values, but only if METHOD=SEQUENTIAL is used.
You'll want to be careful with SEQUENTIAL to make sure that you're
getting the sums of squares you want; it's likely to take several runs
to get desired etas for multiple terms.
 
David Nichols
Principal Support Statistician and
Manager of Statistical Support
SPSS Inc.
 
>----------
>From:  roth jean ellen[SMTP:j-roth2@UX6.CSO.UIUC.EDU]
>Sent:  Sunday, July 19, 1998 12:01 AM
>To:    SPSSX-L@UGA.CC.UGA.EDU
>Subject:       Eta squared
>
>Is Eta squared an option for the general linear models?
>A client of mine asked me to produce Eta squares with
>her analysis, but I wasn't sure what they were, and I
>did not find them in the documentation.
>Thanks,
>Jean Roth
>University of Illinois at Urbana Champaign
>