Information Technology Reference
In-Depth Information
Tests of Between-Subjects Effects
This is the between
subjects variance (
SS
S
)
representing individual
differences. SPSS does
not compute an
F
ratio
for this source.
Measure: MEASURE_1
Transformed Variable: Average
Type III Sum
of Squares
Source
df
Mean Square
F
Sig.
Intercept
Error
1946.025
30.575
1
7
1946.025
4.368
445.533
.000
Figure 10.14
Omnibus ANOVA for between-subjects effects.
×
×
is the Prepost
Sub-
jects” portion of the label for all of its within-subjects error terms. SPSS
provides the effect in parentheses because in analyses with more than one
within-subjects variable each effect will have its own error term; such
labeling will be very useful for helping us read the summary table in those
circumstances.
As can be seen from Figure 10.13, the within-subjects variable
was statistically significant under the Greenhouse-Geisser correction,
F
(1
Subjects interaction but SPSS leaves out the “
2
727. With more than two
conditions as we have here, this significant effect is our cue to perform a
multiple comparisons for the means.
.
951, 13
.
657)
=
18
.
624,
p
<.
05,
η
=
0
.
10.12.4 EVALUATING THE BETWEEN-SUBJECTS VARIANCE
The between-subjects portion of the variance is shown in Figure 10.14.
The
Intercept
focuses on the full model which we do not deal with here.
What SPSS calls
Error
is the variance representing individual differences;
it does not calculate an
F
ratio for this effect.
10.13 PERFORMING THE POST-ANOVA ANALYSIS IN SPSS
Define your variables just as you did for the omnibus analysis so that
you have a main dialog window configured as we showed in Figure 10.8.
Click the
Options
pushbutton to arrive at the
Options
dialog window.
Clear the bottom (
Display
) part of the window if anything is selected,
since that information was produced in the omnibus analysis and is pre-
sumably available to you. We will now work with the top portion of that
window. The top portion of the
Options
dialog window focuses on
Esti-
matedMarginalMeans
.Wehaveusedthisinpriorchaptersinstructuring
the simple effects tests having obtained a significant interaction. We do
something comparable here.
Using a point-and-click method to perform the paired comparisons
procedure, follow the steps shown in Figure 10.15: Click the
prepost
variable over into the
Display Means for
panel. Place a checkmark in the
box for
Comparemain effects
and select
Bonferroni
from the drop-down
menu that will become active when you check the box. This will produce
Bonferroni corrected paired comparisons to be performed just as though
we were performing a simple effects analysis. Click
Continue
to return to
the main dialog window and click
OK
to run the analysis.