Information Technology Reference
In-Depth Information
Pairwise Comparisons
Measure: MEASURE_1
Mean
Difference
(I-J)
95% Confidence Interval for
Difference
a
Sig.
a
(I) prepost
1
(J)prepost
2
3
4
5
1
3
4
5
1
2
4
5
1
2
3
5
1
2
3
4
Std. Error
Lower Bound
Upper Bound
.125
.479
1.000
-1.807
2.057
2.750*
.590
.023
.372
5.128
3.875*
.895
.034
.268
7.482
4.000*
.845
.021
.595
7.405
2
-.125
.479
1.000
-2.057
1.807
2.625*
.625
.040
.107
5.143
3.750*
.750
.016
.728
6.772
3.875*
.693
.008
1.084
6.666
3
-2.750*
.590
.023
-5.128
-.372
-2.625*
.625
.040
-5.143
-.107
1.125
.581
.938
-1.214
3.464
1.250
.366
.112
-.225
2.725
4
-3.875*
.895
.034
-7.482
-.268
-3.750*
.750
.016
-6.772
-.728
-1.125
.581
.938
-3.464
1.214
.125
.398
1.000
-1.479
1.729
5
-4.000*
.845
.021
-7.405
-.595
-3.875*
.693
.008
-6.666
-1.084
-1.250
.366
.112
-2.725
.225
-.125
.398
1.000
-1.729
1.479
Based on estimated marginal means
*.The mean difference is significant at the .05 level.
a.
Adjustment for multiple comparisons: Bonferroni.
Figure 10.17
Pairwise comparisons of means.
codes used by SPSS in the pairwise comparisons are presented in the first
column of the output.
The
Pairwise Comparisons
table is shown in Figure 10.17. We have
seen this sort of table from our simple effects analyses in Chapters 8 and 9.
There are five major rows in the table because each of the five levels of the
repeated measure is given its own focus. The conditions are designated by
their value codes rather than by any descriptive name or label.
We will talk you through the first major row, which focuses on the
first level
pre1
.Thevaluecodeof
1
in the upper left-hand corner under
the column
(I) prepost
stands for
pre1
, the first level of the prepost
within-subjects factor. The value codes under the column
(J) prepost
stand for the other levels of
prepost
(e.g., Level 2 is
pre2
,Level3is
post1,
and so on). The difference between the means of
pre1
and
pre2
(
pre1
-
pre2
) is 0.125 (9
.
125
−
9
.
00
=
0
.
125) and is not statistically significant
(
p
000 as shown in the
Sig
column). The difference between
pre1
and
post1
(Levels 1 and 3 using value codes) is 2
=
1
.
.
750 (9
.
125
−
6
.
375
=
2
.
750)
and is statistically significant. In like manner (
p
=
.
023),
pre1
differs
significantly from
post2
and
post3
.
Overall, the two pretest means do not differ significantly from each
other and the last two posttest means do not differ from each other. The
other differences are statistically significant.
