Information Technology Reference
In-Depth Information
Figure 7.30
The
Enter User Code
window with the planned comparisons specified.
We have already discussed the rationale for the coefficients in these
three planned contrasts. The structure of the code is as follows:
Contrast
is a keyword in SAS to indicate that we are comparing
certain means.
The characters in single parentheses are “literal strings”; they will be
reproduced in the output exactly as we typed them, and will serve
as labels for us to help us read the output.
The word
group
is required to identify the effect that we are using
as the focus of the contrasts.
The coefficients define the particular contrast we are making.
Each line ends in a semicolon.
After entering the code, click
OK
. This returns you to the
User Code
window where your new code can now be seen once you scroll down to the
end of the file (see Figure 7.31). Click
OK
to return to the
Least Squares
window, and click
Run
to perform the analysis.
7.32 OUTPUT FOR PLANNED COMPARISONS IN SAS
The output from the three contrast analyses is presented the final three
rows of the output in Figure 7.32. As you can see, each is identified by
the label we typed in quotation marks and each is evaluated by its own
F
ratio. All three yielded statistically significant effects. Note that the
F
ratio
displayed by SAS is translatable to the
t
value displayed by SPSS because
t
2
F
. For example, for the first contrast, squaring the
t
value of 9.958
displayedbySPSSproducesthe
F
ratio of 99.16 shown in the SAS output.
Thus, we may conclude the following:
=
The combined four groups having studied for the SAT performed
better than the group that did not prepare.
The students preparing four months (Group 3) performed better on
the SAT than those students preparing for two months (Group 2).
The students preparing for relatively longer periods of time (six and
eight months combined) performed better on the SAT than students