Information Technology Reference
In-Depth Information
These are the
p
-values
for the main effects and
covariate effect.
Source
verb_cov
group
F Value Pr
>
F
DF
Type III SS
Mean Square
1
2
123.8980812
13.0293005
123.8980812
6.5146503
87.42
4.60
<.0001
0.0176
math_dv
LSMEAN
LSMEAN
Number
group
1
2
3
1
2
3
10.2018446
11.1945217
11.6869670
Least Squares Means for effect group
Pr
>
⏐
t
⏐
for HO: LSMean(i)
=
LSMean(j)
Dependent Variable: math_dv
These are the
p
-values for the
pairwise comparisons.
i/j
1
2
3
1
2
3
0.1612
0.0161
0.9558
0.1612
0.0161
0.9558
Figure 16.36
The adjusted pairwise comparisons.
To make the analysis complete, we will specify our pairwise mean
comparisons as well, rather than waiting to view the omnibus analysis
results and then go back and run our post-ANOVA mean comparisons.
Under the
Comparisons
panel, set
Show p-values for differences
to
All
pairwise differences
and set
Adjustment method for comparisons
to
Bonferroni.
Click
Run
to perform the analysis.
16.12.2 THE OUTPUT OF THE COVARIANCE ANALYSIS
The output of the ANOVA is shown in Figure 16.36. This matches the
results produced by SPSS.
16.13 COMMUNICATING THE RESULTS
A one-way between-subjects covariance design was used to assess the effec-
tiveness of three different instructional programs to teach math. Thirty-six
eighth grade children, twelve children per group, were exposed to either a
traditional, freeform, or computer-based instructional program at the end of
which they were tested on math word problems. The dependent variable was
the number of math problems solved during the hour of testing. To statistically
control for the math scores the degree to which verbal proficiency might affect
performance, a test of verbal skill was administered to the children earlier in
