6 8 7 4
;data b; set a;
drop t1-t3;
time=1; score=t1; output;
time=2; score=t2; output;
time=3; score=t3; output;
proc glm data=b;
class subj time;
model score= subj time ;
run;Data is usually in the format of one case per line, which has a case number or some other
subject identification followed by the response variable score for each of the
measurement occasions (see data after the cards statement). To create a single response
variable score for use in PROC GLM a data step is used to rearrange the data. This new
variable score is the reading accuracy score at each measurement occasion (time 1, time 2
and time 3 in Table 8.11). To associate each score with its measurement occasion a
variable time is created. The PROC GLM procedure analyses the response variable score
with subject and time as the independent variables. For an explanation of the MODEL
statement refer to the description given in the section-on computer analysis for an
unrelated One-way ANOVA. SAS output for this analysis is shown in Figure 8.18.
General Linear Models Procedure
Class Level Information
(^) Class Levels Values
(^) Subj 5 1 2 3 4 5
(^) Time 3 1 2 3
Number of observations in data set=15
General Linear Models Procedure
Dependent Variable: SCORE
Source DF Sum of Squares^ Mean Square F Value Pr>F
6 140.933333 23.488889 6.38 0.0100
Error 8 29.466667 3.683333
Corrected
Total 14 170.400000
R-Square C.V. Root MSE SCORE Mean
0.827074 26.65557 1.91920 7.20000
Source DF Type I SS Mean Square F Value Pr>F
Subj 4 33.733333 8.433333 2.29 0.1481
Time 2 107.200000 53.600000 14.55 0.0022
Source DF Type III SS Mean Square F Value Pr>F
Subj 4 33.733333 8.433333 2.29 0.1481
Time 2 107.200000 53.600000 14.55 0.0022
Inferences involving continuous data 331