Date: Wed, 25 Jul 2007 09:09:26 -0700
Reply-To: "Dennis G. Fisher, Ph.d." <dfisher@CSULB.EDU>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: "Dennis G. Fisher, Ph.d." <dfisher@CSULB.EDU>
Subject: Re: Type IV SS in GLM?
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
The zeros are real data as far as we are concerned. They are the number
of times having anal intercourse. Most of the straight men have not had
either insertive or receptive anal intercourse (gay/straight is my
2-level IV), but some of them have (which accounts for the two nonzero
but low cells). The four cells for the gay men have a lot of values in
them. I also ran this as just a one-way ANOVA for the gay men only. I
am trying to keep the analysis consistent for all of my dv's and one of
the dv's is anal intercourse with women which more of the straight men
have engaged in.
Nordlund, Dan (DSHS/RDA) wrote:
>> -----Original Message-----
>> From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On
>> Behalf Of Robin High
>> Sent: Wednesday, July 25, 2007 8:27 AM
>> To: SAS-L@LISTSERV.UGA.EDU
>> Subject: Re: Type IV SS in GLM?
>>> The usual advice for using type IV SS in PROC GLM is to use
>> them when
>>> you have a missing cell and there is more than a 1-way ANOVA. In my
>>> case I have a 2 x 4 factorial ANOVA in which two of the
>> cells have all
>>> values of 0. Is it appropriate, or even permissible, to
>> use type IV SS
>>> in this instance? TIA
>> Yes, and if you have access to "Analysis of Messy Data", Vol 1, by
>> Milliken and Johnson, Chapter 13-14 addresses these issues
>> quite nicely
>> with interpretation of a 'cell means' model.
>> Robin High
> Are you using zero (0) to represent missing values, or are these zeros real data values?
> Daniel J. Nordlund
> Research and Data Analysis
> Washington State Department of Social and Health Services
> Olympia, WA 98504-5204