Information Technology Reference
In-Depth Information
The results from the
Contrast Tests
are shown in Figure 7.26. Our
contrasts are evaluated by
t
tests. One very nice feature of this analysis
is that you can select the tests that apply when equal variances can and
cannot be assumed. As you can see from the display, all three contrasts
resulted in statistically significant outcomes.
7.30 COMMUNICATING THE RESULTS
OF THE CONTRASTS ANALYSES
In reporting the results of these user-defined contrasts, we would probably
construct a small table (call it Table 1 for the sake of the write-up) of means
and standard deviations rather than reporting them in the text. Given that,
one way to report the results is as follows:
The amount of preparation for the SAT in which students engaged appeared
to
significantly
affect
their
performance
on
the
test,
F
(4, 30)
=
43
.
47,
2
85. The means and standard deviations of the groups are pre-
sented in Table 1. Three sets of a priori contrasts, all orthogonal to each
other, were performed. The results of these yielded the following outcomes.
First, the students not studying at all performed more poorly on the SAT
than all of those who did study,
t
(30)
p
<.
05,
η
=
.
05. Second, those study-
ing for two months performed worse than those studying for four months,
t
(30)
=−
9
.
96,
p
<.
05. Third, students studying for a relatively shorter
period of time (two months and four months combined) performed more
poorly than those studying relatively more (six months and eight months
combined),
t
(30)
=−
4
.
04,
p
<.
=−
7
.
63,
p
<.
05.
7.31 PERFORMING USER-DEFINED CONTRASTS
(PLANNED COMPARISONS) IN SAS
Planned comparisons are performed in
SAS Enterprise Guide
by adding
a couple of lines of command code to the
Linear Models
procedure.
When we are using the point-and-click interface to specify the details
of a statistical analysis, both SPSS and SAS are translating that into a
set of instructions. SPSS calls those instructions command “syntax” and
SAS calls those instructions command “code.” Either way, they resemble
strings of words or abbreviations that these applications refer to when
performing an analysis. We will see in Chapter 8 how to add syntax to
the SPSS commands so that we may perform our simple effects analysis.
Here,weneedtoaddcodetoourSAScommandsinordertoperformour
planned comparisons.
As we have done before to perform the Tukey post hoc test and the
Dunnett comparisons, we select from the main menu
Analyze
ANOVA
➜
Linear Models
.Onthe
Task Roles
tab, specify
satscore
as the
Depen-
dent variable
and
group
as the
Classification variable
.Onthe
Model
➜
