Date: Tue, 13 Dec 2005 10:18:35 -0800
Reply-To: David L Cassell <davidlcassell@MSN.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: David L Cassell <davidlcassell@MSN.COM>
Subject: Re: Error in log - ??
In-Reply-To: <200512130149.jBD0mYKF013193@mailgw.cc.uga.edu>
Content-Type: text/plain; format=flowed
hari_s_nath@YAHOO.COM wrote back:
>David,
> There were 3 time points, lets say, t0, t1, t2. The timepont t1, is not
>recorded/ missing for one dataset, but all three timepoints including
>baseline are present for the other dataset. And so to balance the two
>datasets, we have added a dummy time variable with missing values for one
>dataset (the log is with regard to this dataset).
> The by groups used here has two cells(cell 1 and 2) and three time
>points. The hypothesis is to see if there are product differences at the
>end of each time points and thatswhy GLM.
> Should i just consider by removing the dummy time variable and do the
>analysis assuming just the two time points?
I can't answer this question. But I do have a lot of questions in return.
Perhaps,
if you write back to SAS-L and actually explain what you are trying to do
(the
big picture, not 'I want an ANOVA'), we ought to be able to help you better.
Here are some of the questions you need to consider answering:
Why do you need to 'balance' the data sets?
What are your two data sets and how do they get combined?
What are you trying to test?
Why are you doing these tests and CIs separately by CELL and TIME, instead
of
incorporating CELL and TIME effects in the model(s)?
Are these all fixed effects, or do you have random effects in there?
(Random effects
are basically effects where you have some of the levels you are interested
in, but not
all, so your TIME variable may be a random effect. SUBJECT_ID may be a
random effect
too. I really can't tell from way over here.)
When you do this, do your residuals look normal? Independent?
Equally-distributed?
How much missing data do you have, if you don't 'balance'?
> >>81 PROC GLM DATA = DATASET ;
> >>82 CLASS PRODUCT_CODE SUBJECT_ID ;
> >>83 BY CELL TIME ;
> >>84 MODEL RESPONSE = PRODUCT_CODE SUBJECT_ID ;
> >>85 LSMEANS PRODUCT_CODE / PDIFF ;
> >>86
> >>87 TITLE "TEST";
> >>88 RUN ;
HTH,
David
--
David L. Cassell
mathematical statistician
Design Pathways
3115 NW Norwood Pl.
Corvallis OR 97330
_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today - it's FREE!
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/
| 
