09.12.2012 Views

2003 IMTA Proceedings - International Military Testing Association

2003 IMTA Proceedings - International Military Testing Association

2003 IMTA Proceedings - International Military Testing Association

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

640<br />

arising from small sample size (however, if values are less than p < .05, we report the lower p<br />

value). Note that because we hypothesize a specific direction for our hypothesis tests, a onetailed<br />

test is appropriate (Cohen & Cohen, 1983). In light of recent recommendations (e.g.,<br />

Wilkinson et al., 1999), and to help better interpret the magnitude of these effects, we also report<br />

90% confidence intervals around each of our hypothesis tests. Thus, by presenting effect sizes, pvalues,<br />

and confidence intervals, readers should best be able to determine the “importance” of<br />

the effects and tests we report.<br />

A different concern, perhaps more important with smaller sample sizes, is the presence of<br />

outliers and extreme cases. To ensure the results were not biased by extreme cases, we examined<br />

the distributions of the variables in terms of skewness and kurtosis (zero represents perfectly<br />

normal distributions, skewness > ±3 and kurtosis > ±7 are indicative nonnormal distributions; see<br />

West, Finch, & Curran, 1995). None of the measures were nonnormal, with openness to new<br />

experiences showing the largest deviation from normality (skewness = -1.27; kurtosis = 4.37) but<br />

still falling well within the range of appropriate normality.<br />

Performance Ratings Across Typical and Maximum Contexts<br />

Table 2 shows the means, standard deviations, and correlations for all measures. As<br />

reflected in Table 2, the low correlation between the team performance measures across the<br />

typical and maximum performance contexts (r = .18, ns) suggests the ratings from these<br />

contexts were not interchangeable.<br />

Hypotheses<br />

Hypotheses 1 through 4 predicted that extroversion (Hypothesis 1), openness to new<br />

experiences (Hypothesis 2), and agreeableness (Hypothesis 3) would be positively related to<br />

transformational leadership, while neuroticism (Hypothesis 4) would be negatively related. As<br />

shown in the last row of Table 2, transformational leadership is positively related to extroversion<br />

(r = .31, p < .05, [.04; .53]), but negatively related to both neuroticism (r = -.39, p < .05, [-.59;<br />

-.14]) and agreeableness (r = -.29, p < .05, [-.52; -.03]). Transformational leadership is not<br />

significantly related to openness to experience, nor conscientiousness. Hence Hypotheses 1 and 4<br />

were supported while Hypotheses 2 and 3 were not. While the relationship between<br />

transformational leadership and agreeableness was significant, it was in the opposite direction as<br />

hypothesized.<br />

In line with Murphy’s (1996) recommendation that personality should be examined using<br />

a multivariate framework, we also conducted a multiple regression analysis in which<br />

transformational leadership was regressed on all of the FFM constructs. As shown in Table 3, the<br />

overall model comprising the five personality factors was significant, explaining 28% of the<br />

variance in transformational leadership ratings (F[5, 33] = 2.59, p < .05). However, only<br />

neuroticism (β = -.29, p < .10, [-.57; .-01]) and agreeableness (β = -.30, p < .10, [-.58; -.02]) were<br />

significant predictors at p < .10 (one tailed).<br />

Next, Hypothesis 5 predicted that transformational leadership would be more predictive<br />

of team performance in maximum rather than typical performance contexts. As Table 2 shows,<br />

transformational leadership was significantly related to team performance in both typical<br />

contexts (r = .32, p < .05, [.06; .54]) and maximum contexts (r = .60, p < .05, [.40; .75]). The<br />

formula proposed by Williams (1959) and Steiger (1980) was used to test for the difference<br />

between two non-independent correlations. We found these correlations to be significantly<br />

different t(36) = 1.63, p < .10 (one tailed), although the confidence intervals overlapped slightly.<br />

Thus, we concluded that the relationship between transformational leadership and team<br />

45 th Annual Conference of the <strong>International</strong> <strong>Military</strong> <strong>Testing</strong> <strong>Association</strong><br />

Pensacola, Florida, 3-6 November <strong>2003</strong>

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!