Date:         Sat, 17 Feb 1996 00:04:02 -0600
Reply-To:     "Serigne F. Diop" <sfdiop01@HOMER.LOUISVILLE.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         "Serigne F. Diop" <sfdiop01@HOMER.LOUISVILLE.EDU>
Organization: SFD. Inc
Subject:      Repeated measures with SAS PROC MIXED?

Hello
     I am trying to use PROC MIXED to analyze data from a repeated
measures design.  I have VAR1 as between-subjects factors, and VAR2 and
VAR3 as within-subject factors. VAR2 and VAR3 are repeated factors. In
other words I collected data from each level of VAR3 within each level of
VAR2, themselves within each level of VAR1.  In my view VAR3 is repeated
within VAR2, which is reapeated within VAR1.
I read in a SAS manual that PROC MIXED is a generalization of PROC GLM.
They did not go into details about the difference and advantages of using
one over the other.  So I was trying to compare the two.
The problem I am having is that I do not know how to specify my model with
PROC MIXED.  With PROC GLM this is how my model looks like:
           PROC GLM DATA=RPMM2;
           CLASS VAR1;
           MODEL Y1-Y12=ZLEVEL/NOUNI CLM;
           REPEATED VAR2 3 (1 2 3), VAR3 4(1 2 3 4)/PRINTE;
Now I am trying to use PROC MIXED which syntax I qam not very familiar
of.  After rearranging the data in univariate form, I tried the following
to no vail:
           PROC MIXED DATA=RPMM2;
           CLASS VAR1 VAR2 VAR3;
           MODEL Y=VAR1|VAR2|VAR3;
           REPEATED VAR2 3 (1 2 3), VAR3 4(1 2 3 4);
 
Can anybody suggest what I am doing wrong?  What are the differences
between the two procedures?
 
Thanks
 
..........