Information Technology Reference
In-Depth Information
These are the probabilities of
the differences occurring
by chance alone for the
pairwise comparisons with
gender as the first order.
Pairwise Comparisons
Dependent Variable: range 0-60
95% Confidence Interval for
Difference
a
Mean
Difference
(I-J)
Sig
a
gender
(I) city reside size
(J) city reside size
Lower Bound
Upper Bound
Std. Error
female
large
small
rural
large
- 23.000*
3.838
.000
- 32.879
- 13.121
- 27.000*
3.838
.000
- 36.879
- 17.121
small
23.000*
3.838
.000
13.121
32.879
- 4.000
3.838
.923
- 13.879
5.879
rural
rural
large
small
small
27.000*
3.838
.000
17.121
36.879
4.000
3.838
.923
- 5.879
13.879
male
large
- 12.000*
3.838
.014
- 21.879
- 2.121
6.000
3.838
.393
- 3.879
15.879
rural
small
large
12.000*
3.838
.014
2.121
21.879
18.000*
3.838
.000
8.121
27.879
rural
rural
large
small
- 6.000
3.838
.393
- 15.879
3.879
- 18.000*
3.838
.000
- 27.879
- 8.121
Based on estimated marginal means
*. The mean difference is significant at the .05 level.
a. Adjustment for multiple comparisons: Bonferroni.
Figure 8.18
Simple effects comparisons of
reside
for each level of
gender
.
is statistically significant at the .05 level (
p
001). Just below that is the
comparison of those living in large cities and those living in rural com-
munities. The difference between these means (large
=
.
−
rural) is
−
27.00,
a difference that is also statistically significant.
As we study the table, we can see that for females there is no reliable
difference between those living in small towns and rural communities and
that for males there is no reliable difference between those living in large
cities and rural communities. All other comparisons yielded means that
were significantly different.
8.12.3 MAPPING SIMPLE EFFECTS ANALYSES TO THE GRAPH
OF THE INTERACTION
Obtaining the simple effects analyses is a necessary step in understanding
the results of a study but it is not sufficient. The next step is to conceptually
map these statistical tests back to the overall interaction effect. This can
most easily be done by viewing the graph of the interaction (drawn in
Figure 8.3) and mentally superimposing on it these simple effects out-
comes. Such processing of the output is needed to write an informative
and accurate results section.
Focusing first on the lines of the graph, we learned from the simple
effects analysis that for females there was no significant difference between
the small town and rural community respondents. Thus, we can conclude
that females are significantly lonelier in small towns and rural communi-
ties than in large cities. Males, on the other hand, are significantly lonelier
in small towns than in either large cities or rural communities.
