Date:         Mon, 10 Feb 1997 15:59:47 GMT
Reply-To:     David Nichols <nichols@SPSS.COM>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         David Nichols <nichols@SPSS.COM>
Organization: SPSS, Inc.
Subject:      Re: Power in trend analysis

In article <199702100421.XAA21563@mail3.uts.ohio-state.edu>,
Joe Wheaton  <wheaton.3@OSU.EDU> wrote:
>I am running a trend analysis using oneway and I want to know the power of
>the trend. I typically run MANOVA's Power subcommand and have the program
>print the effect size, but these must be for the ANOVA, not the trend. Am I
>right, and can I still use these estimates for effect size and power?
>
>Thanks.
>Joe E. Wheaton, Ph.D.
>The Ohio State University
>356 Arps Hall, 1945 N. High St.
>Columubus, OH  43210-1172
>(614)292-8313; FAX: (614)292-4255
>Email: WHEATON.3@OSU.EDU
 
The POWER subcommand in MANOVA produces power estimates for just about all
of the test statistics the procedure gives. If you're looking at the linear
trend, for example, just specify POLYNOMIAL contrasts and POWER, and you'll
get power estimates for the individual parameters, including the linear
one. If you want to get it for some subset of the parameters in the term
larger than one, use PARTITION and specify subsets on the DESIGN and you'll
get it for them.
 
--
-----------------------------------------------------------------------------
David Nichols             Senior Support Statistician              SPSS, Inc.
Phone: (312) 329-3684     Internet:  nichols@spss.com     Fax: (312) 329-3668
-----------------------------------------------------------------------------