NetNews Usenet Archive 1992 #26/NN_1992_26.iso/spool/bit/listserv/statl/1952

Internet Message Format | 1992-11-12 | 1.9 KB

Path: sparky!uunet!utcsri!torn!nott!dgbt!netfs!ub!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!eff!news.oc.com!spssig.spss.com!nichols
From: [e-mail] (David Nichols)
Newsgroups: bit.listserv.stat-l
Subject: Re: power for repeated measures/split plot designs
Message-ID: <[e-mail]>
Date: 12 Nov 92 22:34:40 GMT
References: <[e-mail]>
Sender: [e-mail] (Net News Admin)
Organization: SPSS Inc.
Lines: 25

In article <[e-mail]> [e-mail] writes:
>I am interested in software that will do power analyses for repeated
>measures designs(behavioral science language) or split plot and
>split-split plot designs.
>Programs can be either PC or mainframe.  Our mainframe is a VAX
>configuration v5.1.  Our sas software does NOT include IML.

SPSS on Windows, Mac or OS/2, as well as mainframe, will do post-hoc
power analysis on any design MANOVA will handle (including those listed
above). PC+ will be able to do this down the road, but cannot do it now.
What I mean by post-hoc power analysis is that for a given data set with
a particular noncentrality parameter, etc., MANOVA will give the power
for each test. It will not, however, do all of the kinds of calculations
you might want if you wanted to use it to design studies and wanted to
know what the sample sizes would need to be in order to get a certain
level of power, unless you're willing to do some trial and error with
the post-hoc setup stuff (that is, you would have to dummy up data with
the characteristics assumed for the study and see what the power would
be).

--
 David Nichols        Senior Statistical Support Specialist         SPSS, Inc.
 Phone: [phone]     Internet:  [e-mail]     Fax: [phone]
*******************************************************************************
 Any correlation between my views and those of SPSS is strictly due to chance.