The biology of leadership The relation between leadership, psychopathy and hormones



Download 436.06 Kb.
Page5/7
Date conversion21.02.2016
Size436.06 Kb.
1   2   3   4   5   6   7

4. Results


The respondents were divided into two groups, leaders and non-leaders. The scores of the two groups on the 9 factors of leadership and 3 factors of psychopathy were first analyzed by an independent t-test or a Mann-Whitney test, with a dummy for leadership as a grouping variable and the 12 factors as test variables. The tests on the 9 leadership factors are to test Hypothesis 1 and the tests on the 3 psychopathy factors are to test Hypothesis 2.

The cognition factors are the first factors tested, which can be viewed in table 4. The first test is run to check if the mean scores of leaders on intelligence are higher than the mean scores of the control group on intelligence. A one tailed Mann-Whitney test is used, because the factor is not normally distributed. The leader group has a mean score of 5.26 (mdn = 5.25) on the intelligence factor, whereas the control group had a mean score of 4.65 (mdn = 5.00). The scores of the leaders proved to be significantly higher than the scores of the control group, (U = 574.00, p < 0.01). This result supports the findings of Kirkpatrick (1991), Judge (2004), French and Raven (1959) and Lord et al. (1986). The second test run, is a one tailed Mann-Whitney test on creative thinking, to test if leaders score significantly higher on creative thinking. The leader group scored a mean of 5.00 (mdn = 5.00), whereas the control group scored a mean of 4.36 (mdn = 4.50). Leaders proved to think significantly more creative, than the control group (U = 576.50, p < 0.01). This result supports the findings of Mumford et al. (1991; 2000; 2002), Chusmir (1986), DeVeau (1976) and Sinetar (1985), which connect creative thinking to leadership. The scores of the cognition dimension support Hypothesis 1.

The social cognition factors are the second group of factors that are tested, as shown in table 4. A two-tailed Mann-Whitney test is used, to check if the scores of leaders on agreeableness differ significantly from the score of the control group. The leaders scored a mean of 4.93 (mdn = 5.00), whereas the control group has a mean score of 5,5 (mdn = 5.67). Leaders proved to score significantly lower on agreeableness, than the control group (U = 569.50, p < 0.05). This could have been expected, because leadership emergence is negatively related to agreeableness (Judge et al., 2002). The scores of the leaders and the control group differ significantly, therefore the findings of the Mann-Whitney test on agreeableness support Hypothesis 1. Leaders (M = 5.29, SE = 0.148) did not score significantly higher on the leadership factor empathy than the control group (M = 5.30, SE = 0.17, t(82) = 0.241, p = 0.445) according to an independent one-tailed t-test. If empathy is a leadership factor, the scores should have differed significantly. This result is in contrast to the theory of Mayer and Caruso (2002). An explanation could be, that empathy is only positively correlated with transformational leaders and not with transactional leaders (Bass, 1985). Social cognition is a less important factor in transactional leadership, according to the literature. Leaders (mdn = 5.00) did not score significantly higher on charm compared to the control group (mdn = 5.00, U = 799.50, p = 0.297), based on a one-tailed Mann-Whitney test. The same explanation goes for charm, charm is also a trait that is important in transformational leadership and less important in transactional leadership. The social cognition factors partially support Hypothesis 1, because agreeableness did show a significant difference in score between the two groups. On the other hand, leader scores on empathy and charm are not found to differ significantly from the control group.

The third comparison of means is done on the motivation factors. The next one-tailed Mann-Whitney test is the test on taking charge. Leaders scored a mean of 4.89 (mdn = 5.00) and the control group scored a mean of 3.90 (mdn = 4.00). Leaders proved to score significantly higher on taking charge, than the control group (U = 271.00, p < 0.001). Which is in line with the leadership theory (Winter, 1987; Zaccaro 2004; Chusmir, 1986) and very logical, because taking charge is one of the key tasks of a leader. Leaders (M = 2.37, SE = 0.10) do not score significantly different on the need for achievement compared to the control group (M = 2.47, SE = 0.15, t(82) = 0.536, p = 0.296), according to an independent two tailed t-test, but the need for achievement is significantly related to age (r = -0.281, p < 0.01), which could have caused this insignificant difference in means. Leaders (mdn = 5.00, M = 5.14) do score significantly lower on the need for affiliation than the control group (mdn = 6.00, M = 5.54, U = 640.00 p = 0.035), based on a Mann-Whitney test. Need for affiliation scores of leaders should be low according to Zaccaro (2004) and Chusmir (1986). Need for affiliation could be a driver of empathy scores and agreeableness, because of the significant positive correlations with empathy (r = 0.356, p < 0.01) and agreeableness (r = 0.256, p < 0.05). The scores of leaders (mdn = 3.00, M = 3.02) on risk taking do not differ significantly from the score of the control group (mdn = 2.75, M = 3.00, U = 777.00, p = 0.609), based on a one-tailed Mann-Whitney test. This test result is not expected and in contrast to the theories of Zaccaro (2004), Winter (1987) and in contrast to the findings of Chusmir (1986). Need for affiliation should be low according to Zaccaro (2004) and Chusmir (1986), in combination with a high need for achievement and a high score on taking charge. Leaders should be more entrepreneurial, than the control group and therefore score higher on risk taking (Javadian, 2003; Lawrence et al. 2008). The scores on the motivation factors are partially supporting Hypothesis 1. Taking charge scores of leaders are significantly higher, which supports the hypothesis. Need for achievement scores and risk taking scores did not show any differences, which does not support the hypothesis. Need for affiliation scores of leaders are found to be significantly lower, which also supports Hypothesis 1.



Variable

Group

M

SE

T

U

p

Intelligence

Leader

5.2264







599.500

0.022

 

Control

4.6563

 

 

 

 

Creative thinking

Leader

4.9717







600.500

0.024

 

Control

4.3672

 

 

 

 

Agreeableness

Leader

4.9308







578.000

0.013

 

Control

5.5000

 

 

 

 

Charm

Leader

5.0755







810.500

0.732

 

Control

4.9583

 

 

 

 

Empathy

Leader

5.2453

0.17622

0.223




0.824

 

Control

5.2969

0.14530

 

 

 

Taking charge

Leader

2.3774







278.000

0.000

 

Control

2.4750

 

 

 

 

Need for achievement

Leader

4.8868

0.15418

0.539




0.591

 

Control

3.8958

0.10567

 

 

 

Need for affiliation

Leader

5.1415







670.500

0.100

 

Control

5.5469

 

 

 

 

Risk taking

Leader

3.0283







807.500

0.711

 

Control

3.0000

 

 

 

 

Boldness

Leader

3.2036

0.08169

-4.148




0.000

 

Control

2.8207

0.05213

 

 

 

Disinhibition

Leader

1.5349







706.500

0.198

 

Control

1.6578

 

 

 

 

Meanness

Leader

1.8763

0.06517

-0.082




0.935

 

Control

1.8698

0.04767

 

 

 

Table 4: Descriptives of group scores on the 12 factors

The leader group scored a mean of 3.21 on boldness (SE = 0.05), whereas the control group scored a mean of 2.82 (SE = 0.08), as shown in table 4. Leaders proved to score significantly higher on boldness, than the control group (t(82) = -4.142, p < 0.001), according to an independent one-tailed t-test. This confirms the theory of Hare (1991) and Boddy (2010). The leader group (mdn = 1.48) did not score significantly higher on disinhibition compared to the control group (mdn = 1.62, U = 695.50, p = 0.11), according to a one-tailed Mann-Whitney test. Neither do leaders (M = 1.88, SE = 0.05) score different from the control group (M = 1.86, SE = 0.07) on meanness (t(82) = -0.159, p = 0.437), based on an independent one-tailed t-test. This is not in line with the theories of Hare (1991), Babiak and Hare (2006), Boddy (2010). An explanation for the lack of relation between meanness and leadership could be that meanness measures delinquency and violence, which is proven to be low for successful psychopaths, therefore it could be that the relation between leadership and meanness is not significant. There is minimal confirmation for Hypothesis 2, the only factor where leaders scored significantly higher, was the factor boldness. The other two factors do not show significant differences between the mean scores of leaders and the control group.



Hypothesis 3 is tested by using Spearmans’s correlation coefficients between the leadership factors and the psychopathy factors. Intelligence does not have any significant correlations with the psychopathy factors, which is to be expected because of the psychopathy literature. Psychopathy itself is not related to intelligence, the amount of delinquencies and violence is expected to be related to intelligence. As expected, creative thinking is positively correlated with boldness (r = 0.409, p < 0.001). Agreeableness is significantly negatively correlated with meanness (r = -0.338, p < 0.01). Charm is significantly positively correlated to boldness (r = 0.326, p < 0.01). Charm is not included as a trait in the questionnaire of Patrick (2010), but is a psychopathy trait according to Hare (1991). Empathy has a significant positive correlation with boldness (r = 0.325, p < 0.01). A correlation between empathy and psychopathy could be expected, because of the lack of empathy trait in psychopathy. The correlation with boldness, rather than meanness is unexpected. The lack of empathy trait is included in meanness and not in boldness. Empathy is mostly connected to the boldness traits persuasiveness (r = 0.421, p < 0.001) and dominance (r = 0.258, p < 0.05). ). Need for achievement is positively

Variable

Group

N

Mean

SE

t

df

Sig.

Cortisol 1

Control

30

14.5067

1.06543

-2.253

79

0.027

 

Leader

51

17.8157

0.9355

 

 




Cortisol 2

Control

30

17.0167

1.33164

-2.356

79

0.021

 

Leader

51

21.5922

1.26593

 

 




Cortisol 3

Control

30

8.6

0.84353

0.431

79

0.668

 

Leader

51

8.1275

0.67942

 

 




Cortisol 4

Control

28

6.1429

0.62051

-0.371

75

0.711

 

Leader

49

6.4776

0.58106

 

 




Cortisol 5

Control

30

5.48

1.41312

1.868

31

0.071

 

Leader

51

2.7941

0.26437

 

 




Testosterone 1

Control

26

345.1423

46.88574

0.333

74

0.74

 

Leader

50

325.254

35.56534

 

 




Testosterone 2

Control

27

253.937

30.21027

1.308

71

0.195

 

Leader

46

213.7043

15.58685

 

 




Testosterone 3

Control

27

186.4852

19.69404

1.389

70

0.169

 

Leader

45

155.5822

12.55613

 

 




Testosterone 4

Control

25

140.5

12.4737

0.719

64

0.475

 

Leader

41

127.2146

12.24943

 

 




Testosterone 5

Control

27

146.8333

15.81757

-0.048

64

0.962

 

Leader

39

148.0205

17.45908

 

 




C/T 1

Control

26

0.058

0.00832

-1.728

73

0.088

 

Leader

49

0.0815

0.00888

 

 




C/T 2

Control

27

0.0858

0.00959

-2.223

71

0.029

 

Leader

46

0.1216

0.01098

 

 




C/T 3

Control

27

0.0506

0.0047

-1.312

70

0.194

 

Leader

45

0.06

0.00474

 

 




C/T 4

Control

25

0.0567

0.0083

-0.248

64

0.805

 

Leader

41

0.059

0.00517

 

 




C/T 5

Control

27

0.0309

0.00625

0.596

63

0.553

 

Leader

38

0.0268

0.00374

 

 




AUCgCT

Control

32

0.017

0.00233

-1.943

83

0.055

 

Leader

53

0.0249

0.00278

 

 




AUGiCT

Control

29

-0.0331

0.00633

1.672

78

0.099

 

Leader

51

-0.0525

0.00793

 

 




AUCgC

Control

30

4.0327

0.38448

-1.869

78

0.065

 

Leader

50

4.9525

0.30344

 

 




AUCiC

Control

30

-10.474

0.93116

1.863

78

0.066

 

Leader

50

-12.9295

0.85406

 

 




AUCgT

Control

26

64.0008

7.40733

-0.082

66

0.935

 

Leader

42

64.6777

4.64818

 

 




AUCiT

Control

26

-281.1211

42.53819

0.04

66

0.968

 

Leader

42

-283.4644

37.98887

 

 

 

Table 5: T-test results of hormones

correlated with meanness (r = 0.459, p < 0.001), as expected. Taking charge is significantly correlated with boldness (r = 0.646, p < 0.001). The leadership factor taking charge and the psychopathy trait dominance are closely related, according to Babiak and Hare (2006). Therefore, a correlation between boldness and taking charge is expected, because dominance is a trait that is included in boldness. Need for affiliation is negatively correlated with meanness (r = -0.310, p < 0.01). Need for affiliation measures the need to be liked by co-workers, therefore it can be expected that need for affiliation is negatively correlated meanness. Taking charge is significantly correlated with boldness (r = 0.646, p < 0.001). The leadership factor taking charge and the psychopathy trait dominance are closely related, according to Babiak. Risk taking is expected to be correlated to disinhibition, because of the boredom proneness and impulsivity traits. The bivariate correlation analysis supports this expectation (r = 0.337, p < 0.01). Risk taking is also correlated to meanness (r = 0.236, p < 0.05), which can be expected, because excitement seeking is a trait that is included in meanness. A great share of the findings support Hypothesis 3. Charm, creative thinking, and taking charge are correlated with boldness. Need for achievement, agreeableness, need for affiliation and risk taking are correlated with meanness. Empathy is not correlated with meanness, but unexpectedly with boldness. The finding that risk taking is correlated with disinhibition does also support Hypothesis 3.



The leader (M = 17.82, SE = 0.94) group has significantly higher cortisol level at the moment of waking up (cortisol 1) compared to the control group (M = 14.50, SE = 1.07, t(79) = -2.253, p < 0.05). Which is opposite of what is expected, based on the theory of Terburg (2011). The results of cortisol 2, have a similar significant result, leaders (M = 21.59, SE = 1.27) showed a significantly lower score on cortisol compared to the control group (M = 17.02, SE = 1.33, t(79) = -2.356, p < 0.05). The third, fourth and fifth measurement do not show a significant difference between groups, neither do the measurements of testosterone, as can be seen in table 5. Testosterone scores are expected to be significantly higher, than the scores of the control group (Terburg, 2011). The second C/T scores of leaders (M = 0.12, SE = 0.01) do differ significantly from the control group (M = 0.09, SE = 0.01, t(71) = -2.23, p < 0.05), again the effect is opposite of what is expected. The scores on the area under the curve of cortisol with respect to the ground of leaders (M = 4,95, SE = 0.38) differ marginally significant from the non-leader group (M = 4.03, SE = 0,38, t(78) = -1,87, p = 0.065). The difference between the AUCi scores of cortisol (AUCiC) scores of the two groups also differ marginally significant. Leaders score a mean of -12,93 (SE = 0.85), whereas the control group scores a mean of -10,47 (SE = 0.93, t(78) = 1.86, p = 0.066) on AUCiC. The same goes for AUCg of the C/T ratio (AUCgCT). The AUCgCT level of leaders (M = 0.25, SE = 0.01) differs marginally significant from the score of the control group (M = 0.17, SE = 0.01, t(83) = -1,94, p = 0.055). Again, the difference in the AUC values are exactly opposite of the expected direction.

The results of a hierarchical regression analysis, with 6 different models that follow the literature, include some interesting findings, which can be found in table 6. The dummy variable for leadership is been set as the dependent variable and every model includes a new set of independent variables. The first model only includes one demographic variable, age. Other demographic variables are not suitable for a regression analysis, because these variables are nominal. Age has a significant influence on leadership, in all 6 models. The second model includes cognition, which does not make a big difference on the R2. The cognition factors have low ’s and lack in significance. The same can be said about the inclusion of the social cognition factors in model 3. However, when including the motivational factors in model 4, the factor empathy becomes significant, which was not the case in model 3. Empathy will remain significant in model 5 and 6. Taking charge is the only motivational factor in model 4, that has a significant influence on leadership. Including psychopathy factors in model 5 hardly changes anything in both the R2, nor the ’s of other variables. Including the hormone factors in model 6 does make a big difference. Taking charge stops having a significant influence, whereas cortisol 2 (B = 0.046, SE = 0.023, = 0.796, p = 0.052), testosterone 2 (B = -0.002, SE = 0.001, = -0.665, p < 0.05), C/T 2 (B = 5.853, SE = 5.675, = 0.348, p < 0.05), AUCgC (B = -0.298, SE = 0.152, = -1.196, p = 0.056) and AUCgCT (B = 39.922, SE = 19.873, = 1.289, p = 0.051) have a (marginally) significant influence on leadership. The  -values show that age has a very big influence on leadership from model 1 to 5. Other variables that have a big influence on leadership are the hormone variables. AUCgC has a of -1.196, which is the biggest influence seen in all six models. The first model, including the constant and the age variable explains already much of the variance, R2 = 0.325. Including the cognition variables results in a small jump to R2 = 0.361. Inclusion of the social cognition variables in model 3 has a slightly bigger effect R2 = 0.410. Inclusion of motivation factors in model 4 results in a positive change of R2 of 0.145, which indicates a big improvement in the amount of variance in the leadership dummy that is explained, R2 = 0.555. Introducing the psychopathy factors in model 5 makes a very small difference in the R2, R2 = 0.570. The introduction of the hormone variables is responsible for a big improvement in R2, R2 = 0.680.

Running a regression analysis with the amount of years of leadership performance as a dependent factor results in slightly different results, as can be seen in table 7. The first model introduces the demographic variable age. As can be expected, age has a much bigger influence in this regression model, then when the leadership dummy is set as a dependent variable. The of age is much higher ( = 0.733) compared to the former regression analysis ( = 0.570). The R2 of model 1 in both regression analyses, support that finding (R2 = 0.540) compared to (R2 = 0.325), it indicates that age has a bigger influence on leadership experience, than on the leadership dummy. Age is strongly related to years of experience and will have a significant influence in all six of the models. Cognition is introduced in the 2nd model, as can be seen in table 7, creative thinking has a significant influence. Model 3 introduces the social cognition variables. Agreeableness has a significant influence, while creative thinking remains significant. Both agreeableness and creative thinking lose their significance, when introducing the motivational factors in model 4, whereas taking charge does have a significant influence. The significance of taking charge disappears when introducing the psychopathy factors in model 5. Model six includes the hormone variables, AUCgCT is the only introduced variable that has a significant influence in this model, together with age, all other influences are found to be insignificant. Empathy is not significant in this regression analysis, which means that empathy is not evolved over time, empathy is a variable that an individual has or doesn’t have. The same goes for the hormone variables, they do not change with an increase in leadership experience, except for AUCgCT.



dummy regression.jpg

Table 6: Hierarchical regression analysis, the dependent variable is the leadership dummy

regression experience de goede!!.jpgTable 7: Hierarchical regression analysis, the dependent variable is years of leadership experience

A regression analysis, with the leadership dummy as dependent variable has been run with the significant independent variables of table 6, to show the effects of the independent variables without all insignificant variables. This regression analysis is displayed in table 8. The regression analysis including 6 different models, as described earlier and displayed in table 6, shows significant influences of several hormone variables. In this regression analysis, the influence of hormones is much less, than in the regression analysis performed earlier. The fact, that the influence of cortisol and cortisol related variables (C/T and AUC variables related to cortisol) are not present, is surprising. This means that the influence of age and the scores on taking charge predict leadership better, than hormones do.



Variable

B

SE



 

Constant

-,709

,344




*

Age

,016

,004

,345

**

Empathy

-,095

,048

-,187




Taking Charge

,264

,055

,468

***

Cortisol 2

,013

,011

,234




Testosterone 2

-,001

,001

-,187




C/T 2

-,835

1,691

-,117




AUCgCT

2,015

6,865

,073




AUCgC

-,016

,050

-,071

 

*p < 0.05, **p < 0,01, ***p < 0,001








Table 8: Regression analysis with the leadership dummy as dependent variable
A repeated measures ANOVA is run to test the between group and within group differences in the hormone curve. Cortisol 1, cortisol 2 and cortisol 3 are set as the within subjects measure and the leadership dummy is set as the between subjects variable. The one-way repeated measures ANOVA shows 3 different results. Firstly, it shows that cortisol levels differ over the first 3 measurement points (F(1.817, 143.546) = 82.67, p < 0.001) according to a within subjects analysis. Secondly, leaders have a significantly different curve from the control group (F(1.817, 143.546) = 4.493, p < 0.001), according to a within subjects analysis, comparing the curves of leaders and the control group. Thirdly, leaders start with a higher cortisol level (F(1, 79) = 4.507, p < 0.05), according to a between subjects analysis. Leaders start with a higher cortisol 1 and cortisol 2 value, but the mean cortisol 3 value ends up close to the cortisol 3 value of the control group. This finding proves again that leaders have a significantly higher cortisol 1 and cortisol 2 value, as was seen in the t-tests. The development of the hormone values differs significantly between groups too, because leaders drop more between the second and the third measurement point, as can be seen in graph 1.

A bivariate correlation analysis shows several correlations between leadership factors and hormone variables. This indention will briefly report those correlations. The most correlations have been found between cortisol and leadership factors. The first finding is that cortisol 1 is positively correlated with empathy (r = 0.285, p < 0.01) and taking charge (r = 0.228, p < 0.05). The direction of the correlation between empathy and cortisol is in the right direction, high cortisol is associated with empathy, whereas the direction of the correlation between taking charge and cortisol is opposite of what is expected. Taking charge is expected be positively correlated with testosterone and negatively correlated with cortisol. Cortisol 2 is also positively correlated with taking charge (r = 0.221, p < 0.05). Cortisol 3 is positively correlated with charm (r = 0.269, p < 0.05), this finding is opposite of what was expected, because testosterone is supposed to be positively correlated with charm. The testosterone 3 variable shows few correlations. Testosterone 3 is positively correlated with risk taking (r = 0.344, p < 0.01) and disinhibition (r = 0.233, p < 0.05), both findings are in line with what was expected. Both disinhibition and risk taking are traits that are highly related to psychopathy. In the C/T and AUC variables, we see a correlation with empathy reoccurring. C/T1 is positively correlated to need for affiliation (r = 0.232, p = 0.045) and C/T3 is positively correlated with empathy (r = 0.247, p < 0.05). AUCgCT is positively correlated with empathy (r = 0.272, p < 0.05), whereas AUCiC is negatively correlated with empathy (r = -0.275, p < 0.05). The negative correlation between AUCiC and empathy is the result of a big morning peak in the cortisol values. The extreme drop of the cortisol level at cortisol 3 makes the AUCIC negative.





Graph : Cortisol measurements
1   2   3   4   5   6   7


The database is protected by copyright ©essaydocs.org 2016
send message

    Main page