Date:         Fri, 27 Jun 2008 16:29:26 -0700
Reply-To:     stringplayer_2@yahoo.com
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Dale McLerran <stringplayer_2@YAHOO.COM>
Subject:      Re: MIXED - COMPOUND SYMETRY
In-Reply-To:  <b8ea349c-1119-48ae-b7de-84f8ddf7958d@x19g2000prg.googlegroups.com>
Content-Type: text/plain; charset=utf-8

--- On Fri, 6/27/08, nchapinal@YAHOO.COM <nchapinal@YAHOO.COM> wrote:

> From: nchapinal@YAHOO.COM <nchapinal@YAHOO.COM>
> Subject: MIXED - COMPOUND SYMETRY
> To: SAS-L@LISTSERV.UGA.EDU
> Date: Friday, June 27, 2008, 3:36 PM
> Imagine you have a trial where you applied 2 different treatments to
> animals (one particular animal is only in one of the treatments)and
> you measure their weights on different periods (let’s say 3, and
> period is categorical)
> I always thought that if you wanted to use CS, you could write either
> of the codes below. However, the second option (random animal) doesn’t
> always give me the right Degrees of Freedom! Does anyone know why?
>
> THANKS!
>
> proc mixed ;
> class animal period treatment;
> model weight = period treatment ;
> repeated period/subject=animal type=cs;
> run;
>
> proc mixed ;
> class animal period treatment;
> model weight = period treatment ;
> random animal;
> run;


The reason that you do not get the correct df for the code with the
RANDOM statement is that you have not specified that animals are
nested within treatments.  If you modify your code as follows, then
your df should be correct.

  proc mixed ;
    class animal period treatment;
    model weight = period treatment ;
    random animal(treatment);
  run;


Dale

---------------------------------------
Dale McLerran
Fred Hutchinson Cancer Research Center
mailto: dmclerra@NO_SPAMfhcrc.org
Ph:  (206) 667-2926
Fax: (206) 667-5977
---------------------------------------