Analysis of Repeated Measures Design
/****************************************************************/
/* S A S S A M P L E L I B R A R Y */
/* */
/* NAME: GLMREPT */
/* TITLE: Analysis of Repeated Measures Design */
/* PRODUCT: STAT */
/* SYSTEM: ALL */
/* KEYS: analysis of variance, repeated measures analysis */
/* PROCS: GLM */
/* DATA: */
/* */
/* SUPPORT: UPDATE: */
/* REF: */
/* MISC: */
/* */
/****************************************************************/
*-----------------------------------------------------------------+
| |
| Multivariate Analysis of a Repeated Measures Design |
| |
| Each subject receives 3 treatments, t1-t3, in random order. A|
| data step is used to code an intercept and treatment contrasts. |
| Subjects are also classified as male/female. GLM is used to test|
| treatment, sex, and treatment*sex effects. |
| |
+-----------------------------------------------------------------;
title 'Multivariate Analysis of a Repeated Measures Design';
data;
input sex $ t1-t3;
intercep = 1;
mean = (t1 + t2 + t3) / 3;
t12 = t1 - t2;
t13 = t1 - t3;
dummy = (sex = 'M') - (sex = 'F');
datalines;
F 1 2 4
F 2 3 1
F 2 1 3
F 1 3 2
M 1 2 4
M 2 1 5
M 2 1 3
;
proc print;
run;
*--------------------------old way-------------------------;
proc glm;
model t12 t13 = intercep dummy/noint; manova h=intercep;
title2 'Test of Treatment Effect';
run; quit;
proc glm;
class sex;
model mean = sex;
title2 'Test of Sex Effect';
run; quit;
proc glm;
class sex;
model t12 t13 = sex; manova h=sex;
title2 'Test of Treatment by Sex Interaction';
run; quit;
*--------------------------new way-------------------------;
proc glm;
class sex;
model t1-t3=sex;
repeated treatmnt 3 contrast(1);
title2 'Repeated Measures Analysis using REPEATED Statement';
run; quit;