Date:         Sat, 6 Mar 1999 00:21:59 PST
Reply-To:     Hayrettin OKUT <hokut@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Hayrettin OKUT <hokut@HOTMAIL.COM>
Subject:      Re: Crossover Analyses
Comments: To: afjb@IX.NETCOM.COM
Content-type: text/plain

Use PROC GLM for fixed effects. A sample from SAS/STAT is as folowing.
This sample is for Grizzle's two pariod crossover design. For random and
mixed, however, use PROC MIXED.
As you know there are many tpyes of Changeover (Crossover) such as
Optimal Crossover
Extra-Period-Change-Over
Tied-Double-Change-Over
Balanced Two-Period Change-Over
Balaam's Two-period Deslgn with t2 Experimental Units
and so on. If you have SAS programs for these designs I would be
appriciated to have them.
 
Dr. Hayrettin OKUT
Yuzuncu Yil Universitesi
Fen Bilimleri Enst. Mud.65080 Van-Turkey
Phone  +90 2251097
fax +90 2251123
e.mail:   hokut@hotmail.com
hokut@yyu.edu.tr
hokut@ahtamara.yyu.edu.tr
 
 
 
DATA ONE;
   INPUT PERIOD TRT $ SUBJECT RESPONSE;
   IF _N_<=12 THEN SEQ=1; IF _N_>12 THEN SEQ=2;
   CARDS;
1 A 11 .2
1 A 12 0
1 A 13 -.8
1 A 14 .6
1 A 15 .3
1 A 16 1.5
2 B 11 1.
2 B 12 -.7
2 B 13 .2
2 B 14 1.1
2 B 15 .4
2 B 16 1.2
1 B 21 1.3
1 B 22 -2.3
1 B 23 0
1 B 24 -.8
1 B 25 -.4
1 B 26 -2.9
1 B 27 -1.9
1 B 28 -2.9
2 A 21 .9
2 A 22 1.0
2 A 23 .6
2 A 24 -.3
2 A 25 -1
2 A 26 1.7
2 A 27 -.3
2 A 28 .9
PROC PRINT;
PROC GLM; CLASSES TRT SEQ SUBJECT;
  MODEL RESPONSE=TRT SEQ SUBJECT(SEQ);
  TEST H=SEQ E=SUBJECT(SEQ);
  RUN;
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com