Information Technology Reference
In-Depth Information
Figure 9.26
We have now specified Bonferroni corrected pairwise comparisons for the
main effect of political preference and for the simple effects of the three-way interaction.
Because there were only two levels of
kidshome
,weknowfromour
descriptive statistics that families with children in the home (coded as
1) were less satisfied than families without children in the home (coded
as 2). The main effect of
voter
was not significant, indicating that these
two overall means can be treated as being comparable. Political pref-
erence, however, has three levels, requiring us to perform a multiple
comparison procedure to determine which pairs of means were reliably
different.
We wish to perform comparisons of the means for (a) the main effect
of
politics
and (b) the means of the groups for the three-way interaction;
that is, we wish to perform tests of simple effects. As we did for the two-
way design discussed in Chapter 8, we will use a Bonferroni corrected
multiple comparisons test for both of these so that we can accomplish the
specifications in one step.
Configure the analysis as described above for the omnibus analy-
sis. Then select the
Least Squares
portion of the
Post Hoc Tests
tab.
This brings you to the screen shown in Figure 9.26. For
Class effects
to use
,set
politics
and
kidshome
voter
to
True
;thiswillgen-
erate the post hoc tests for the main effect of
politics
and the simple
effects tests for the interaction, respectively. For
Comparisons
,set
Show
p-values for differences
to
All pairwise differences
and set
Adjustment
method for comparison
to
Bonferroni
.Thenclick
Run
to perform the
comparisons.
∗
politics
∗