Psychology 815:Computer Lab #10
Analysis of Variance: repeated measures and hierarchical analysis
November 15, 1996


Objectives:

     1) Analysis of Variance (repeated measures)

     2) Analysis of Variance (replication within cells

     3) Analysis of Variance (split-plot)

     4) Analysis of Variance (hierarchical analysis)




objective 1: Basic repeated-measure ANOVA

This is the basic repeated measure ANOVA. The model statement specifies the dependent variable that you are looking at over time. Notice that there is no independent variable (after the equal sign) in this design.

proc glm data=SASUSER.data815;
model cb33 cb34 cb35 = ;
repeated time 3 ( 1 2 3 ) ;
run;

The "repeated" statement is SASCODE indicating that this is a repeated measure ANOVA. "Time" is a name that you give to the repeated factor. You can call it whatever you wish. The "3" indicates that there are three levels for the repeated measure (e.g., time 1, time 2, and time 3) "(1 2 3) assigns the names for each of the levels. You can substitute any name you wish.


objective 2: replication within cells

Here is the input statement: input id sex age replicat time1 time2;

Here's the program:

Proc GLM data=weights2;
class id;
model time1 time2 = id;
repeated time 2 (1 2);
random id /test;
run;
 
 

 1   1   10    1    100   155

 2   1   13    1    122   150

 3   1   15    1    132   155

 4   1   14    1    125   180

 5   1   18    1    150   160

 6   1   15    1    140   190

 7   1   16    1    130   135

 8   1   18    1    142   145

 9   1   17    1    115   175

10   1   19    1    118   190

11   2   12    1     55    67

12   2   13    1     85    75

13   2   18    1     45    85

14   2   15    1     52    80

15   2   14    1     38    90

16   2   16    1     47    75

17   2   18    1     80    85

18   2   19    1     45   100

19   2   11    1     35   105

20   2   15    1     25    50

 1   1   10    2    100   100

 2   1   13    2    122   122

 3   1   15    2    132   132

 4   1   14    2    125   125

 5   1   18    2    150   150

 6   1   15    2    140   140

 7   1   16    2    130   130

 8   1   18    2    142   142

 9   1   17    2    115   115

10   1   19    2    118   118

11   2   12    2     55    55

12   2   13    2     85    85

13   2   18    2     45    45

14   2   15    2     52    52

15   2   14    2     38    38

16   2   16    2     47    47

17   2   18    2     80    80

18   2   19    2     45    45

19   2   11    2     35    35

20   2   15    2     25    25

objective 3: Split-plot with experimental/control groups

This is the same as Objective 1, except that you are looking at the dependent variable by sex.

proc glm data=SASUSER.data815;
class SEXcb;
model cb33 cb34 cb35 = SEXcb ;
repeated time 3 ( 1 2 3 ) ;
run;


objective 4: Hierarchical ANOVA

This is NOT a repeated measure design. It is a nested ANOVA. Unfortunately, I do not have meaningful data for you to use right now. However, go ahead and run it with the nonsense variables below. It won't make any sense, but at least you'll get a feel for the format.

Note that like a regular ANOVA, the class statement indicates the independent variables. Also, like a regular ANOVA, the model statement indicates the dependent variables to the left of the equal sign and the independent variables are to the right of the equal sign.

Of course, in this case, we have a nested design. In this case the variable FILOUTBY is nested in AGE.

proc glm data=SASUSER.DATA815;
class FILOUTBY AGE;
model TOTLPROB = FILOUTBY (AGE) ;
run;