Date: Mon, 5 Mar 2001 07:33:04 -0500
Reply-To: Gerhard Hellriegel <ghellrieg@T-ONLINE.DE>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Gerhard Hellriegel <ghellrieg@T-ONLINE.DE>
Subject: Re: proc glm.please help me, i can't find the error
On Mon, 5 Mar 2001 06:34:16 GMT, Vkoala5132 <vkoala5132@AOL.COM> wrote:
>this is my log window. i have done an experiment with repeated measures.
>rewrote the data in order to do a tukey hsd test. but the program can not
>the MEANS and thus i cant complete my experiment. i have already added a
>semicolon in line 164. without it, it can not find TEST H=. as i have a due
>date for this experiment i would urgently need help. kind regards isabella.
>160 PROC GLM DATA=NEU1;
>161 CLASSES TIME SUBJ GROUP;
>162 MODEL INVEST = GROUP SUBJ(GROUP) TIME GROUP*TIME
>164 Test H=GROUP E=SUBJ(GROUP);
> /*i have added the semicolon already. Without it, the program does not
>the TEST H=*/
>165 Test H=TIME GROUP*TIME E=TIME*SUBJ(GROUP);
>166 MEANS TIME / TUKEY;
>NOTE: Means from the MEANS statement are not adjusted for other terms in
> model. For adjusted means, use the LSMEANS statement.
>WARNING: Mean separation procedures will not be carried out due to zero
> of freedom for the specified error source.
I'm no specialist for statistics or the GLM proc, but as far as I see, you
have no error. The warning relates to some definitions in your model and in
your test statement. You should have a look into the documentation or the
online help and see, how you can adjust the "degrees of freedom". Maybe
you'll understand better, what the terms there mean.