Information Technology Reference
In-Depth Information
Project Type
Leadership Style
Simple Projects
Complex Projects
Overall Mean
Authoritarian
6.00
1.00
3.50
Democratic
6.40
6.80
6.60
Laissez Faire
3.60
2.80
3.20
Overall Mean
5.33
3.53
Figure 13.2
Cell means and marginal means for the data in our numerical example.
Figure 13.2 presents the cell means as well as the row and column
means. This is a convenient way to quickly summarize the results and
makes the factorial structure of the design clear. The main effect of lead-
ership style deals with comparing the row means to each other, the main
effect of project type deals with comparing the column means to each
other, and the interaction effect deals with the patterns of the cell means.
13.3 EFFECTS OF INTEREST
The summary table showing the results of the analysis is presented in
Table 13.1. The effects in which we are interested are analogous to the
other two-way designs that we have covered in Chapters 8 and 11: two
main effects and the two-way interaction. As can be seen, all three of these
effects are statistically significant.
The eta squared values are shown in the last column of the sum-
mary table. Although it is customary to use the total variance as the
basis for computing this statistic in between-subjects and within-subjects
designs, in mixed designs it is common to keep the between-subjects and
Table 13.1.
Summary table for a simple mixed factorial design
Source
2
SS
df
MS
F
η
Between subjects
85.87
14
28.35
∗
A
(leadership style)
70.87
2
35.44
.83
S
/
A
15.00
12
1.25
Within subjects
67.50
97.20
∗
B
(project type)
24.30
1
24.30
.36
80.40
∗
A
×
B
(Leadership
×
Project)
40.20
2
20.10
.60
B
×
S
/
A
3.00
12
0.25
Total
153.37
29
∗
p
<
.05.
2
A
Note:
η
=
SS
A
/
SS
S
/
A
.
B
=
SS
B
/
SS
B
×
S
/
A
.
η
2
A
η
B
=
SS
A
×
B
/
SS
B
×
S
/
A
.
×
