Information Technology Reference
In-Depth Information
Linear Models
The GLM Procedure
Least Squares Means
Adjustment for Multiple Comparisons: Bonferroni
lonely
LSMEAN
LSMEAN
Number
gender
reside
1
1
1
2
2
2
1
2
3
1
2
3
15.0000000
38.0000000
42.0000000
16.0000000
28.0000000
10.0000000
1
2
3
4
5
6
Least Squares Means for effect gender * reside
Pr
>
⏐
t
⏐
for H0: LSMean(i)
=
LSMEAN(j)
The difference between
group 1 (urban female)
and group 6 (rural male) is
not statistically
significantly different.
Dependent Variable: lonely
i/j
1
2
3
4
5
6
1
2
3
4
5
6
0.0365
0.2328
0.0192
0.0688
<
.0001
<
.0001
1.0000
1.0000
<
.0001
1.0000
<
.0001
<
.0001
<
.0001
<
.0001
1.0000
<
.0001
1.0000
<
.0001
<
.0001
1.0000
The difference between
group 5 (small city male)
and group 6 (rural male) is
statistically significantly
different.
0.0365
0.2328
0.0192
0.0688
0.0014
1.0000
<
.0001
<
.0001
1.0000
0.0014
Figure 8.32
Tests of simple effects.
assigned by the
Linear Model
procedure to each level of
reside
.The
lower display uses the code numbers in presenting a matrix showing the
probability values associated with each mean difference. For example,
the mean difference of the groups coded 1 and 3 (the large city and rural
groups) has a value of
p
=
.
0022 and that mean difference is therefore
statistically significant.
8.16.2 SIMPLE EFFECTS ANALYSES
Figure 8.32 displays the pairwise comparisons of the means of the inter-
action structured in the same way as we just described for the main
effect. The upper table gives code numbers to the groups and the
lower table shows the
p
values associated with the pairwise compar-
isons. Recall that females were coded as 1 and males as 2, and that
large cities, small towns, and rural communities were coded as 1, 2, and
3, respectively. Thus, the group coded as 2 represents gender
=
1and
reside
2, that is, females living in small towns; the group coded as 5
represents gender
=
=
2 and reside
=
2, that is, males living in small towns.
