home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!uwm.edu!spool.mu.edu!olivea!pagesat!spssig.spss.com!nichols
- From: nichols@spss.com (David Nichols)
- Newsgroups: bit.listserv.spssx-l
- Subject: Re: manova interpretation
- Message-ID: <C0Enrv.Dyp@spss.com>
- Date: 5 Jan 93 23:55:06 GMT
- References: <1992Dec31.135743.18937@polaris.utu.fi> <C0CvIA.BHy@spss.com>
- Sender: news@spss.com (Net News Admin)
- Organization: SPSS Inc.
- Lines: 43
-
- In article <C0CvIA.BHy@spss.com> nichols@spss.com (David Nichols) writes:
- >In article <1992Dec31.135743.18937@polaris.utu.fi> Timo Toivonen <ttoivonen@abo.fi> writes:
- >>
- >>How to interprete MANOVA
- >>
- >>Manova gives interaction parameters in a known order. What is this order.
- >>Is it so that the first parameter is the value of effect when both (or
- >>more) variables are in the first category. But the second parameter? Is
- >>it so that the first mentioned variable (for example, if variable names
- >>are USOS and HESU and the interaction term USOS BY HESU) is still in the
- >>first category, but the second one moves to the second category or vice
- >>versa.
- >
- >The interaction contrasts are created by elementwise multiplication of
- >the main effect parameters involved. The incrementing of subscripts
- >runs fastest on the last factor listed, so that if A and B are the factors,
- >then AB(1,1) is a contrast with coefficients equal (or proportional) to
- >the product of those for the first contrasts of A and B. The next parameter
- >for the interaction would be AB(1,2), then on to AB(i,j-1), then back to
- >AB(2,1) up to AB(2,j-1), etc., up to AB(i-1,j-1). The interaction parameters
- >created by the default deviation contrasts are generally not of supreme
- >usefulness. Often it certain forms of interaction are postulated, you'll
- >have to do one of two things: figure out what kind of main effect contrasts
- >will produce the desired particular interaction contrasts, or simply lay
- >the design out as a big oneway and specify all the contrasts in the model
- >and partition the pieces if you want to get the factor effects in addition
- >to the individual 1 df contrasts.
-
- Oops, sorry, there's a typo here. The line above that reads:
-
- >for the interaction would be AB(1,2), then on to AB(i,j-1), then back to
-
- should read:
-
- >for the interaction would be AB(1,2), then on to AB(1,j-1), then back to
-
- the difference being the 1 for the i.
-
- --
- David Nichols Senior Statistical Support Specialist SPSS, Inc.
- Phone: (312) 329-3684 Internet: nichols@spss.com Fax: (312) 329-3657
- *******************************************************************************
- Any correlation between my views and those of SPSS is strictly due to chance.
-