Character for Leadership
54
members were identified as those whose responses fell within the lowest third,
middle third, and upper third percentiles, respectively, for each character scale. The
traditional ANOVA is based on the assumptions that the
population from which the
data are taken is both normally distributed and is homoscedastic. The data in this
study meet the criterion of equal variances according to Levene’s test. Table 6
presents this information.
Table 6: Levene’s Test for Equality of Variances
Levene statistic
p
H
1
2.90
.06
H
2
1.19
.31
H
3
1.62
.20
H
4
.24
.79
H
5
.55
.58
H
6
1.10
.34
H
7
.14
.87
However, due to the noncontinuous nature of the independent variable data,
a nonparametric test must be used to examine differences between means.
Therefore, the data for this study were analyzed using the nonparametric equivalent
of
the traditional ANOVA, the Kruskal–Wallis test. Data were analyzed for
hypotheses 1 - 6 by Kruskal–Wallis to determine whether or not there were
significant differences in character groups (low, medium, and high levels) for each
leadership scale. Data for hypothesis 7 were also analyzed for differences between
groups due to the division of data for the independent variable,
current ministry
leadership involvement, into three discrete groups. Kruskal–Wallis was also used to
analyze data for hypothesis 7. Table 7 presents the results of all data analyses for
the stated hypotheses. An alpha level of .05 was used for all statistical tests.
Character for Leadership
55
Table 7: Chi-Squared Results for Hypotheses,
α
= .05
χ
2
p
H
1
Confident leadership
by self-directedness level
20.37
.00
H
2
Visionary
leadership
by
self-directedness level
17.29
.00
H
3
Follower-centered leadership by cooperativeness level
7.97
.02
H
4
Visionary leadership by cooperativeness level
18.40
.00
H
5
Mature character by self-transcendence level
2.87
.24
H
6
Visionary leadership by self-transcendence level
.22
.95
H
7
Visionary leadership
by current ministry level
5.45
.07
Hypothesis 1 posited a difference in how the VLT construct of confident
leadership was enacted between groups based on measured levels for the character
trait self-directedness. The difference between participants demonstrating low,
medium, and high levels of self-directedness was found statistically significant for
confident leadership,
χ
2
(2,
Ν
= 101)
= 20.37,
p
= .00.
This hypothesis was
supported, allowing rejection of the null hypothesis at that level. Hypothesis 2
posited a difference in how visionary leadership [the total score on Sashkin et al.’s
(1997) TLP] was enacted between groups based on measured levels for the
character trait self-directedness. The difference between participants demonstrating
low, medium, and high levels of self-directedness was found statistically significant
for visionary leadership,
χ
2
(2,
Ν
= 101)
= 17.29,
p
= .00. This hypothesis was also
supported, allowing rejection of the null hypothesis at that level. Hypothesis 3
proposed a difference in how the VLT construct of follower-centered leadership
was enacted between groups based on measured levels for the character trait
cooperativeness. The difference between participants demonstrating low, medium,
and high levels of cooperativeness was found statistically significant for follower-
centered leadership,
χ
2
(2,
Ν
= 101)
= 7.97,
p
= .02. This hypothesis was supported,
allowing rejection of the null hypothesis at that level. Hypothesis 4
proposed a
difference in how visionary leadership was enacted between groups based on
measured levels for the character trait cooperativeness. The difference between