Vous êtes sur la page 1sur 13

Personality and Individual Differences 33 (2002) 11011113 www.elsevier.

com/locate/paid

Emotional intelligence and leadership in adolescents


Danielle Charbonneau*, Adelheid A.M. Nicol
Department of Military Psychology and Leadership, Royal Military College of Canada, PO Box 17000, Station Forces, Kingston, ON, Canada K7K 7B4 Received 8 June 2001; received in revised form 13 November 2001; accepted 9 December 2001

Abstract We tested the validity of two measures of emotional intelligence (EI) and we investigated the relation between EI and leadership in 191 adolescents (M=14.33 years) attending a 3-week military training camp. A scale by Schutte et al. [Personality and Individual Dierences 25 (1998) 167] assessed primarily the intrapersonal aspect of EI, whereas selected items from the Weisinger [Emotional intelligence at work (1998) Dan Francisco, CA: Jossey-Bass] scale measured primarily the interpersonal aspect. Participants were also rated by their peers and junior leaders on the Weisinger items. Leadership was assessed using a peer nomination system for task-goal and socio-emotional orientation [Schneider, Ehrhart, & Holcombe (in press) Leadership in adolescence: comparing peer and teacher perspectives and correlates, Leadership Quarterly]. Both measures, but especially the Schutte et al. scale, correlated with social desirability, suggesting problems of discriminant validity. Scores on the Schutte et al. scale did not correlate with any peer nominations, indicating questionable convergent validity. In contrast, scores on the Weisinger scale (self-report) correlated with peer nominations of socio-emotional leadership and task-goal leadership. However, the lack of correlation between the self-rated and the other-rated versions of the Weisinger scale is a concern. # 2002 Published by Elsevier Science Ltd.
Keywords: Emotional intelligence; Leadership; Adolescents

1. Emotional intelligence Although the construct of emotional intelligence (EI) is still under development, research has demonstrated the usefulness of the construct of EI in college students and adults. For example, research has demonstrated the positive relations between EI and rst year college grades (Schutte et al., 1998), life satisfaction and reduced depressive symptomatology (Martinez-Pons, 1997), and
* Corresponding author. Fax: +1-613-541-6822. E-mail address: charbonneau-d@rmc.ca (D. Charbonneau).
0191-8869/02/$ - see front matter # 2002 Published by Elsevier Science Ltd. PII: S0191-8869(01)00216-1

1102

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

eective leadership (Barling, Slater, & Kelloway, 2000). The current study tested the validity of two instruments measuring EI in adolescents participating in a summer military training camp. Because these adolescents were expected to demonstrate leadership-like qualities, we investigated also the relation between EI and leadership in this age group. The study of EI includes at least two kinds of approaches. Work by Mayer and Salovey (1997) typies the rst approach that consists of identifying the abilities comprised in this type of intelligence. They have identied four basic abilities: to perceive, appraise, and express emotion; to intentionally use emotions to inuence thinking; to understand the complexity and the transient nature of emotions; and, lastly, to use emotions to promote emotional and intellectual growth. The second approach uses a broader denition of EI and focuses on what EI predicts, such as success in a broad sense across dierent situations. For instance, someone high in EI would motivate themselves, would understand emotions in others, and would handle relationships successfully (Goleman, 1995). Petrides and Furnham (2000) refer to these approaches as information-processing and trait EI, respectively. The current study focuses on the trait aspect of EI, which can be measured by self-report inventories. Weisinger (1998) wrote about the application of EI in a work environment. He distinguished between what he called the intrapersonal and interpersonal aspects of EI. He dened the intrapersonal aspect as being able to help oneself by using ones emotions. To do this, one must be aware of personal emotions and moods, be able to manage them and to use them to motivate oneself. In contrast, the interpersonal aspect focuses on helping others. To help others, one must identify and respond to others moods and emotions, guide these emotions toward a positive outcome, and help others help themselves. That is, Weisinger treated separately the awareness, management, and utilization of emotions in oneself and the awareness, management, and utilization of the emotions of others. 1.1. The study of EI in adolescents Although much literature exists on the development of emotional competence (e.g. Saarni, 1999) and prosocial behaviors (e.g. Eisenberg & Miller, 1987), the development of emotional intelligence is less well understood. Saarni explains that the construct of emotional competence is broader than that of EI because it includes empathy, moral character, a focus on emotions in a social context, and a sense of self-ecacy. EI in adolescents has been less well investigated than EI in college students and adults. One resulting diculty in assessing EI in adolescents is that most instruments were designed for use with post-adolescent samples. Consequently, our selection of measures and the basic validation of these instruments represent important aspects of this study and warrant some explanation. Martinez-Pons (1998) is among the few who have investigated EI in adolescents. He studied parents inuence on their adolescents EI. He used the Trait Meta-Mood Scale (Salovey, Mayer, Goldman, Turvey, & Palfai, 1995) to measure the EI, but he claimed that this measure assesses the self-oriented aspect of EI. Indeed, he argued that EI is comprised of two aspects: the self-oriented and other-oriented, which are reminiscent of Weisingers (1998) intrapersonal and interpersonal aspects, respectively. Martinez-Pons concluded that the other-oriented aspect was a better predictor of the quality of adolescents social interactions whereas the self-oriented aspect correlated

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

1103

with task-orientation and reects a concern for mastery and personal improvement. In this paper, the terms self-oriented and intrapersonal are used interchangeably, as are the terms other-oriented and interpersonal. 1.2. Selection and basic validation of instruments Of the available measures of EI, we opted for the scale developed by Schutte et al. (1998) for four reasons. First, the development of the items has a theoretical foundation because it was based on an earlier version of an emotional intelligence model (Salovey & Mayer, 1990). Second, Schutte et al. showed that this scale has sound psychometric properties in adults and Abraham (1999) recommends its use for this reason. Third, it is readily available. Finally, the content of the scale led us to believe that it might be suitable for use with adolescents. The two aspects of EI, intrapersonal and interpersonal, are represented in the scale although the majority of items describe intrapersonal EI. Therefore, we consider this scale as representing predominantly the self-oriented aspect of EI. This assumption is consistent with Schutte et al.s claim that the scale is unifactorial, at least in adults (for a dierent opinion, see Petrides & Furnham, 2000). A second instrument was included to assess specically the other-oriented aspect of EI because we suspected that the quality of ones social interactions may be related to leadership, the process of intentionally inuencing others (Yukl, 1998). Martinez-Pons (1998) developed three questions to measure this aspect of EI, but we wanted an instrument that captured the various dimensions of this construct. We selected the interpersonal dimension of Weisingers (1998) EI scale. Convergent and discriminant validity, two aspects of construct validity, were examined in this study. Convergent validity is shown when a scale correlates with a measure of a theoretically related construct (DeVellis, 1991). Empathy includes the appraisal and expression of others emotions, particularly distress, and, as such, is related to the other-oriented aspect of EI (Ciarrochi, Chan, & Caputi, 2000; Goleman, 1995, 1998). Empathy should also correlate with the Schutte et al. (1998) scale. Indeed, Salovey and Mayer (1990) claim that empathy may be an important feature of emotionally intelligent behaviour, as empathy includes the ability to comprehend anothers feelings and to re-experience them oneself (p. 194). To measure another aspect of convergent validity, self-reports and other-reports were taken with Weisingers (1998) interpersonal scale, thereby investigating the possibility that some aspects of EI may be assessed by others. A multi-rater approach was not taken with the Schutte et al. (1998) scale because information on how a person deals with his/her emotions may not be readily observable, unless we know the individual very well. Campbell and Fiske (1959) stated that a test is not valid if it correlates too highly with another test that is not theoretically related to it. This is known as discriminant validity. In the present study, the two measures of EI were expected to correlate poorly with a desirability scale. One other validity analysis was conducted. We wanted to replicate some gender dierences. A gender dierence on the empathy scale has been reported in adolescents, with female adolescents scoring higher than male adolescents (Davis & Franzoi, 1991). Also, adult women scored more highly than adult men did on the Schutte et al. (1998) scale. Therefore, we expected a gender dierence using the same scale with an adolescent sample.

1104

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

1.3. EI and leadership in adolescents The association between EI and leadership behaviours has been demonstrated in adults (Barling et al., 2000). The authors demonstrated that managers who scored high on the Bar-on (1997) EI inventory were perceived by their subordinates as displaying more transformational leadership behaviours. Transformational leaders share their ideals with their subordinates and act as role models, inspire them by articulating an appealing vision, stimulate them intellectually to think in innovative ways, and show care and concern for each of them (Bass, 1998). Megerian and Sosik (1996) described theoretical mechanisms that explain the link between EI and transformational leadership. For instance, they postulated that leaders need to show empathy towards their subordinates in order to be eective in two-way communication, in delegating responsibilities, and in attending to individual dierences in needs. Also, one characteristic of eective leadership concerns the ability to interpret what others have said and to identify the emotion associated with it (Locke, 1991). To our knowledge, the relation between EI and leadership has not been tested in adolescents. The study of early signs of leadership in adolescents, although sparse, suggests that adolescents can display leadership behaviours similar to those of adults. For instance, Zacharatos, Barling, and Kelloway (2000) have shown that adolescents who display transformational leadership behaviours engender their teammates satisfaction with their leadership style and willingness to exert additional eort in a sports context. Schneider and his collaborators have launched a research program that investigates the prediction, understanding, and durability of leadership behaviours in adolescents but have not looked at EI (Schneider, Ehrhart, & Holcombe, in press; Schneider, Paul, White, & Holcombe, 1999). To examine the relation between EI and leadership, we followed a peer nomination method similar to that used by Schneider et al. (in press). In this method, adolescents are asked to make nominations for two leadership styles, namely task-goal leadership (ability to complete a task and achieve a goal) and socio-emotional leadership (ability to make people cooperate with each other and to be supportive; e.g. Yukl, 1998). These two aspects of leadership were chosen because of their resemblance, albeit distant, with Martinez-Pons (1998) criteria of task-orientation and quality of social interactions that were dierentially associated with the two dimensions of EI. We expected that individuals high in other-oriented EI would be nominated more frequently as socio-emotional leaders, because of their high social functioning (Martinez-Pons, 1998). We predicted that scores on the Schutte et al. (1998) scale would correlate with nominations for both the socio-emotional and the task-oriented leadership, because this instrument contains items representing both aspects of EI.

2. Method 2.1. Participants Participants were adolescents enrolled in a 3-week summer camp providing military skills training. During camp, participants resided together in military facilities. They were divided into groups of ve to 21 members, depending on their trade (e.g. boson, gunnery, music, and sailing). Scheduled activities were geared towards the development and practice of trade skills and other

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

1105

military skills, under the supervision of junior leaders. Junior leaders were older adolescents who underwent the summer training in previous years. These group leaders were responsible for the participants discipline and wellbeing. Two 3-week sessions occurred during the summer and participants were asked to complete the study package during their last week at camp. No signicant dierence between participants enrolled in these two sessions was observed on any of the variables measured here, so the two samples were combined. A total of 235 adolescents participated in the study. After elimination of participants with incomplete data sets, 191 participants remained (81%). Of these participants, 110 were male and 81 were female and 107 were involved in the rst session and 84 took part in the second session. The mean age of participants was 14.33 years (SD=1.06; range=1218 years). When the seven participants belonging to the extreme age groups were excluded from the analyses (three 12 year olds, three 17 year olds, and one 18 year old), the intercorrelations were generally higher than when the analyses were conducted on the entire sample. In this paper, we present the results for the entire sample. 2.2. Materials A complete data set consisted of self-report questionnaires on intrapersonal and interpersonal EI, desirability, empathy, one to four peer ratings of interpersonal EI, and one to two junior leader ratings of interpersonal EI. In addition, participants in the second session were asked to nominate three peers within their trade group for popularity, friendship and leadership. 2.2.1. Intrapersonal emotional intelligence Participants EI was assessed using the 33-item EI scale developed by Schutte et al. (1998). A 5-point Likert-type scale was provided, where 1=Strongly disagree and 5=Strongly agree. Schutte et al. reported good internal consistency ( =0.90) and good testretest reliability (r=0.78) for the scale when administered to adults. Furthermore, the instrument demonstrated good predictive and discriminant validity. The scale contains items of intrapersonal EI (e.g. I am aware of the non-verbal messages I send to others) and of interpersonal EI (e.g. I help other people feel better when they are down). 2.2.2. Interpersonal emotional intelligence Participants evaluations of their own and their peers interpersonal EI were measured with the 23 items representing this dimension of EI from the scale developed by Weisinger (1998). Items describe the use of EI with others (e.g. Help a group to deal with their emotions; Accurately tell people how they are feeling; Communicate your feelings eectively; Recognize when others are distressed). Weisinger argued that four items describe both types of EI, and they were included in the scale (e.g. Stay calm when you are the target of anger from others). Because of the age group and the work environment context for which the items were developed, some items had to be reworded to be suitable for adolescents. For instance, Mediate conicts between others was reworded to Help resolve conicts with others. Also, items were modied for self and peers assessment. For example, Help others manage their emotions became (name provided) Helps others manage their emotions. The modied instrument was given to two adolescents, independent of the participants, for review of the wording. Answers were given on

1106

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

a 5-point Likert-type scale, where 0= Not at all and 4=Frequently or always. The psychometric properties of the scale are unknown. 2.2.3. Desirability scale The 8-item scale of desirability was taken from the Personality Research Form (PRF-E) (Jackson, 1984). This scale measures the extent to which a participant has endorsed or not endorsed self-report statements which describe socially desirable behaviours. 2.2.4. Empathy scale As a self-report measure of empathy, the 28-item Interpersonal Reactivity Index (Davis, 1980) was used. This scale contains four 7-item factors: Perspective Taking, Fantasy, Empathic concern, and Personal Distress. Perspective Taking includes the ability to understand other peoples perspective. Fantasy relates to the ability to identify oneself with ctional characters. Empathic Concern describes feelings of compassion and concern for others. Lastly, Personal Distress assesses how anxious and uncomfortable one becomes in response to anothers distress. The 5point Likert-type response scale ranged from 0=Does not describe me well to 4=Describes me very well. Davis reported internal reliabilities in the range of 0.710.77 for the scales. Davis (1983) demonstrated adequate validity for each of the four scales. This scale has been used successfully with high school students (e.g. Davis & Franzoi, 1991). 2.2.5. Popularity, friendship and leadership measures Participants were asked to identify the three people in their group that were the most popular and their closest friends (Schneider et al., in press). For the leadership assessment, a hypothetical scenario involving someone placed in charge of a committee was presented. Participants had to nominate three people who, would do the best job of getting things done and making sure your goals are reached (taskgoal leadership) and three people who would do the best job of helping the committee cooperate and work well together (socio-emotional leadership). The variable of interest was the number of times a participant had been nominated in each category. Schneider et al. provided a list of names for adolescents to chose from. We replicated this method in the rst data collection but the nominal lists we had been given for each trade group were incorrect. The incorrect lists could have biased someones choice and, consequently, we excluded this data from analyses. Of particular concern was the fact that someone may have been prevented from nominating a person in their group because the name of that person did not appear in the incorrect list of names they were given to chose from. Conversely, some individuals may not have been nominated because their name appeared on the list given to a dierent group. For the second data collection we asked participants to nominate three individuals from their group without providing them with a nominal list. 2.3. Procedure Nominal lists of participants were obtained from the personnel in charge of running the training camp. During the last week of camp, participants were taken to one of two rooms where the study and the voluntary aspect of their participation were explained to them. Participants were asked to complete the intrapersonal and interpersonal EI scales, the empathy scale, and the

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

1107

desirability scale. For the peer evaluation, participants were asked to rate their roommate (when possible) and three or four randomly selected peers, mainly from amongst their trade group, on interpersonal EI. Names of peers to be rated were written on Post-itTM ags, which were removed once the measures had been sorted and stapled together for each participant. Blank questionnaires were returned when participants did not know well the peer(s) they were asked to rate. This happened because of the incorrect lists we were provided and resulted in an unequal number of raters per participant. Participants also nominated three peers from their group for popularity, friendship, and leadership. Junior leaders were asked to complete the questionnaire for each member of their respective group at the end of camp. They were involved in both sessions and thus rated both samples.

3. Results Peers evaluations of participants interpersonal EI were aggregated, as were the junior leaders evaluations. An index of interrater agreement, rwg, was computed to evaluate the appropriateness of averaging the scores of dierent raters (James, Demaree, & Wolf, 1984, 1993). A computation was made for four dierent groups. Results indicated that the following percentages of withinparticipant raters had rwg index scores of greater than 0.7: for two peer raters, 87% (N=34); for three peer raters, 85% (N=70); for four peer raters, 90% (N=75); for two junior leader raters, 100% (N=119). The random selection of peers used for the peer assessments may have contributed to the lower index scores observed for peer ratings as compared to those of junior leader ratings. On the basis of these results, it was concluded that averaging of scores was acceptable. 3.1. Reliability Means, standard deviations, intercorrelations and internal consistencies (when applicable) for each of the measures are presented in Table 1. For the second half of the sample, standardized scores were used for within-groups peer nominations to mitigate the eect of dierent group sizes (between 14 and 22). All observed internal consistency values were above conventional cutos (DeVellis, 1991) with the exception of the Personal Distress subscale. The scale was used, but results based on the Personal Distress subscale should be interpreted with some caution. 3.2. Validity To demonstrate convergent validity, the two EI scales should correlate with each other and with empathy. As expected, results indicate that the scores on the Schutte et al. (1998) and the self-report version of the Weisinger (1998) scales are highly correlated. Both scales correlated signicantly with three empathy factors, namely Perspective Taking, Fantasy, and Empathic Concern. Personal Distress did not correlate signicantly with either EI scale. A further check on the convergent validity of the Weisinger (1998) scale revealed that the self-rating scores did not

1108

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

correlate signicantly with peers or with junior leaders ratings. Peers ratings correlated signicantly with junior leaders ratings, although the correlation is small. In sum, both the intrapersonal and interpersonal EI scales demonstrated good convergent validity with each other and with empathy, but inter-rater correlations for the Weisinger scale were lower than expected. A high correlation between the EI scales and desirability would suggest poor discriminant validity (Campbell & Fiske, 1959). The Schutte et al. (1998) scale and the Weisinger (1998) scale correlated moderately with desirability. Hence, adolescents tended to answer items on both scales in a way that made them appear good, which may be more problematic for the Schutte et al. scale than for the Weisinger scale. Gender dierences were expected on the empathy factors and on the Schutte et al. (1998) scale. As expected, female participants scored signicantly higher than male participants on three empathy factors, namely Perspective taking, Empathic Concern, and Personal Distress (see Table 2). Although in the expected direction, the gender dierences failed to reach signicance for the Fantasy scale and for the Schutte et al. scale. Female participants scored higher both on the self-report and on the peer-rating version of the Weisinger (1998) scale, but there was no gender dierence in the junior leader ratings on this scale. 3.3. Popularity, friendship, and leadership nominations The peer-nomination data were obtained with a limited sample (N=84) and the scores were standardized within each group to eliminate the eect of unequal group sizes. Contrary to expectation, scores on the Schutte et al. (1998) scale did not correlate signicantly with any peer nominations (see Table 1). In contrast, scores on the self-rated version of the Weisinger (1998) scale correlated signicantly with peer nominations of socio-emotional leadership and task-goal leadership. Similarly, peer-ratings of interpersonal EI correlated signicantly with peer nominations
Table 1 Descriptive statistics and incorrelations for all study variables (N=191)
M 1. Empathy (total) 2. Perspective Taking 3. Fantasy 4. Empathic Concern 5. Personal Distress 6. Intrapersonal EI 7. Interpersonal EI (self) 8. Desirability 9. Interpersonal EI (peers) 10. Interpersonal EI (leaders) 11. Popularity (Z-peers)a 12. Friendship (Z-peers) 13. Leader-Task (Z-peers) 14. Leader- Social (Z-peers)
a b c

SD

10

11

12

13

2.29 2.33 2.39 2.70 1.58 3.70 2.77 3.46 2.37 2.42 0.01 0.02 0.03 0.05

0.71 [0.77]b 0.38 0.62c [0.67] 0.84 0.74c 0.25c [0.72] 0.67 0.76c 0.41c 0.38 [0.69] 0.61 0.40c 0.12 0.12 0.13 0.53 0.22c 0.38c 0.45c 0.38c 0.49 0.30c 0.41c 0.33c 0.48c 0.66 0.16 0.35c 0.16c 0.30c 0.72 0.10 0.05 0.17 0.12 0.82 0.01 0.10 0.00 0.12 0.98 0.01 0.03 0.04 0.01 0.99 0.00 0.04 0.03 0.12 1.00 0.19 0.05 0.18 0.28 0.97 0.31c 0.18 0.23c 0.34c

[0.57] 0.10 [0.90] 0.08 0.58c 0.26c 0.49c 0.10 0.14 0.17 0.04 0.02 0.01 0.16 0.12 0.04 0.05 0.13 0.04

[0.86] 0.34c [0.61] 0.09 0.15 0.14 0.18 0.22c 0.21 0.18 0.10 0.03 0.25 0.07 0.40c 0.00 0.34c 0.01 0.35c 0.39c 0.38c 0.00 0.36c 0.20

0.49c 0.12 0.44c

0.37c 0.60c

0.63c

For all peer nominations (11, 12, 13, and 14), data were available only for the second half of sample (N=84). Numbers in square brackets represent Cronbachs . P < 0.01.

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

1109

of friendship and the two types of leadership. Comparable junior leaders ratings correlated signicantly only with peer-ratings of task-goal leadership.

4. Discussion The present study included a basic validity check on two measures of EI in adolescents. We adopted Martinez-Pons (1998) conceptualization of EI. Specically, we predicted that the two aspects of EI would be dierentially related to leadership. The present study extends previous research on EI in four ways. First, it suggests that the use of Schutte et al.s (1998) scale may be problematic with adolescents. Second, Weisingers (1998) interpersonal EI scale demonstrated acceptable basic psychometric properties and may be used with caution in adolescents, at least until further validation is undertaken. Third, we demonstrated that the other-oriented aspect of EI is observable and can be assessed by others. Lastly, the results demonstrated that some aspects of EI are associated with leadership not only in adults (Barling et al., 2000), but also in adolescents. 4.1. Validity Convergent validation using an empathy scale yielded signicant moderate to high correlations with both self-rating scales of EI, as expected. Only the empathy factor of Personal Distress did not correlate with either scale. Many items on this factor are related to how one reacts in emergency situations. These reactions may be less closely related to EI and more closely related to emergency reaction training, such as rst aid, which is a learned skill. Also, the observed internal consistency for this subscale was low. Weisingers (1998) interpersonal scale was also tested for convergent validity using a multi-rater method. Consistent with the literature on self-reports (e.g. Sosik & Megerian, 1999), adolescents tended to rate themselves higher than did their peers or their junior leaders. However, the lack of
Table 2 Gender dierences for empathy and emotional intelligence Variable Boys M Empathy Perspective Taking Fantasy Empathic Concern Personal Distress Intrapersonal EI Intrapersonal EI (self) Intrapersonal EI (peers) Intrapersonal EI (junior leaders)
a b

Girls SD 0.85 0.66 0.81 0.70 0.57 0.55 0.49 0.74 0.84 M 2.40 2.46 2.51 2.88 1.72 3.78 2.87 2.55 2.49 SD 0.43 0.67 0.86 0.60 0.64 0.48 0.49 0.64 0.78

df

2.22 2.22 2.30 2.57 1.48 3.64 2.69 2.23 2.38

189 183 181 182 180 189 186 182 187

1.72 2.41a 1.68 3.09b 2.70b 1.75 2.55a 3.09b 0.88

P < 0.05. P < 0.01.

1110

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

correlation between the self-rated and the other-rated versions is a concern. It is possible that items in the self-rated version were interpreted as how the participant focused on others emotions in general. In contrast, the same items rated by peers and junior leaders may have been interpreted more specically as how the participant dealt with the raters emotions. Alternatively, it could be that a period of three weeks is too short to get to know a person well on an emotional level, in a military environment. Theoretically, the other-oriented aspect of EI should be detectable by others. However, the correlation between the peer-rated and junior leader-rated versions is low, suggesting a poor inter-rater convergent validity. Three explanations are possible. First, the random selection of peers used for the peer assessment may contribute to the lower correlation. Indeed, it is possible that the selected peers were familiar with participants to dierent degrees, thereby introducing exaggerated variability in the responses. Second, a military environment may not be conducive to interactions on an emotional level between individuals of dierent ranks. Indeed, the military may encourage individuals to relate dierently to their peers and to their superiors. Junior leaders had authority over the participants and it may have been considered inappropriate for a participant to attend to his/her junior leaders emotions. Finally, peers were younger than junior leaders and may have an EI that is less well developed than that of junior leaders. Accordingly, peers may have been less aware of others emotions. To demonstrate good discriminant validity, the EI scales should correlate minimally with the desirability measure. Desirability was especially problematic in the Schutte et al. (1998) scale, suggesting a tendency to answer items in a way that leaves a good impression. The problem may lie with the scale (e.g. items too transparent, only three items reverse coded) or with the population (e.g. younger age group than that for which the instrument was intended), or both. Hence, the Schutte et al. scale should be used with caution with adolescents. As expected from previous ndings (e.g. Davis & Franzoi, 1991), the present results indicated signicant gender dierences on three of the four empathy factors. Only the dierence on the Fantasy factor did not reach signicance and the reason for this is unknown. Female participants scored signicantly higher than male participants in the self-rating and peer-rating versions, but not the junior leader-rating version of Weisingers (1998) other-oriented EI scale. However, contrary to Schutte et al.s (1998) reported gender dierence in an adult population, the gender dierence on the Schutte et al. scale did not reach signicance. Selforiented EI may be less developed in adolescents than in adults. Indeed, the average item scores on this scale (M=3.78 and M=3.64, for girls and boys, respectively) are lower than scores reported by Schutte et al. for her adult samples (M=3.97 and M=3.78, for women and men, respectively). The general pattern of the validation results indicate that Schutte et al.s (1998) scale may be problematic for use in adolescents mainly because it correlates moderately with a measure of desirability. The items may not be well suited for this age group. Alternatively, the problem may lie with the scale and not this age group. Indeed, most items on the Schutte et al. scale are transparent and worded positively. Comparison with the scales relation to social desirability in adult samples would clarify the issue. Weisingers (1998) interpersonal EI scale looks more promising, but should be more extensively validated before its use can be recommended for adolescent populations. In particular, it should be tested using other theoretically related constructs (e.g. extraversion, self-esteem, assertiveness,

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

1111

interpersonal communication, interpersonal competence) to establish the distinctiveness of EI from other personal traits. As one anonymous reviewer rightly pointed out, the number of new scales that are labeled EI is increasing rapidly and it is imperative that new scales be tested to ensure they are not re-labeled versions of well-established measures. 4.2. Popularity, friendship, and leadership nominations Results are quite striking and contrary to expectations: none of the correlations between Schutte et al.s (1998) scale and peer nominations are signicant. Possible explanations include the questionable validity of the scale, a poorly dened construct, and the younger age of the participants. Furthermore, the factorial structure of the scale has been questioned (Petrides & Furnham, 2000). In contrast, individuals who rated themselves highly on Weisingers (1998) interpersonal EI scale were peer-nominated more often on leadership abilities. Similarly, peer-ratings on the interpersonal EI scale were signicantly correlated with nominations of friendship and the two leadership types, but not with popularity. Taken together, these results suggest that popularity was not associated with having high interpersonal EI. Lastly, a highly emotionally intelligent individual, as perceived by junior leaders, was more likely to be nominated by peers as being able to get the job done (taskgoal leadership) than to be nominated for resolving interpersonal conicts within a committee (socio-emotional leadership). Peers did not distinguish between the task-goal orientation and the socio-emotional orientation in leadership, but junior leaders did. On the basis of Martinez-Pons (1998) relation between otheroriented EI and social functioning, we expected the relation to be with socio-emotional leadership, which involves the use of social skills, and not with taskgoal leadership. In the context of a military skills training camp, junior leaders may have paid more attention to the fact that a task had been completed, and less to the manner in which it was completed. In contrast, peers may have paid attention to both task completion and the manner in which it was completed. Indeed, peers are directly involved in accomplishing the task whereas junior leaders may be interested mostly in its completion. The signicant correlations between interpersonal EI and peer nominations of leadership are consistent with Golemans (1998) claim of a relation between EI and leadership, at least in part. These results also extend Barling et al.s (2000) ndings of such a relation in adults. However, our results suggest that only some aspects of EI may be related to leadership. The pattern of peer nomination intercorrelations diers from that reported by Schneider et al. (in press). First, the correlation between the two types of leadership (r=0.63) is lower than that found by Schneider et al. (r=0.81) and suggests that the two types of leadership are theoretically related but distinct from one another. Second, peers tended to nominate their friends as good leaders in the current study whereas in the Schneider et al. study, they tended to nominate popular individuals for leadership positions. There are some important dierences between the current study and that of Schneider and his colleagues (in press). Peers had known each other for up to 24 months in the Schneider et al. sample, whereas peers had been observing each other for only 3 weeks in the current sample, although some participants knew each other from previous summers. This shorter contact time may have provided less time to get to know the popular individuals well enough to nominate them for a leadership position. Another dierence resides in the method. Schneider and his colleagues

1112

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

provided a list of names for peers to choose from whereas participants in the present study had to nominate peers within their groups, without the use of a list. In the absence of a list, it is possible that friends names came rst to mind when answering the questions. 4.3. Generalizability and limitations The present ndings can be regarded with condence for several reasons. First, gender dierences in empathy were replicated for the most part on a well-known empathy questionnaire (Davis, 1980). This gives credibility to the way participants responded to items. Second, we used peers and junior leaders as raters on one questionnaire, thereby avoiding mono-method bias. Third, it may be argued that summer military training camp emphasizes discipline and obedience more than the development of EI. Therefore, scores on EI may have been lower than would have been found in other contexts. This increases our condence that any signicant result would be replicable in another sample. Nonetheless, the instruments should be administered to dierent, non-military samples before rm conclusions about them can be reached. Some cautionary statements must be made regarding the ndings. First, the scales that we used were not pure on either dimension as each scale contained some items of the other dimension. We have not demonstrated that our assumption about the existence of two dierent aspects of EI was correct. Because of concerns due to a limited sample size and the possibility of problems with the validity of the Schutte et al. (1998) scale, conclusions based on a factorial examination of the present data set would be questionable. However, preliminary results (not reported here) suggest that a factorial exploration of the intrapersonal and interpersonal aspects of EI on a larger sample may be a worthwhile endeavor. Second, the peer-nomination data on popularity, friendship, and leadership were obtained with a limited sample. Finally, the measure of leadership was based on perceptions of who would be the most ecient individual in a hypothetical scenario. An instrument assessing leadership behaviours, such as transformational leadership behaviours, would be preferred as it may reect reality more closely. In conclusion, results indicated that the Schutte et al. (1998) scale may be problematic because of questionable discriminant validity, despite good convergent validity. In contrast, Weisingers (1998) scale revealed acceptable discriminant validity but yielded mixed results on convergent validity, evidenced by low inter-rater correlations. Results suggest that the interpersonal aspect of EI, which can be assessed by others, is associated with peer nominations of leadership in adolescents. Acknowledgements This study was supported by the Academic Research Program Grant GRC0000B1607 from the Royal Military College to Danielle Charbonneau. The authors thank Sarah Hill for her comments on an earlier draft of the manuscript. References
Abraham, R. (1999). Emotional intelligence in organizations: a conceptualization. Genetic, Social, and General Pyschology Monographs, 135, 209224. Barling, J., Slater, F., & Kelloway, E. K. (2000). Transformational leadership and emotional intelligence: an exploratory study. Leadership & Organization Development Journal, 21, 157161.

D. Charbonneau, A.A.M. Nicol / Personality and Individual Dierences 33 (2002) 11011113

1113

Bar-on, R. (1997). Bar-on emotional Quotient Inventory: technical manual. New York: Multi-Health Systems. Bass, B. M. (1998). Transformational leadership: industrial, military, and educational impact. Mahwah, NJ: Lawrence Erlbaum Associates. Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrati-multimethod matrix. Psychological Bulletin, 56, 81105. Ciarrochi, J. V., Chan, A. Y. C., & Caputi, P. (2000). A critical evaluation of the emotional intelligence construct. Personality and Individual Dierences, 28, 539561. Davis, M. H. (1980). A multidimensional approach to individual dierences in empathy. JSAS Catalog of Selected Documents in Psychology, 10, 85. Davis, M. H. (1983). Measuring individual dierences in empathy: evidence for a multidimensional approach. Journal of Personality and Social Psychology, 44, 113126. Davis, M. H., & Franzoi, S. L. (1991). Stability and change in adolescent self-consciousness and empathy. Journal of Research in Personality, 25, 7087. DeVellis, R. F. (1991). Scale development: theory and applications. Newbury Park, CA: Sage Publications. Eisenberg, N., & Miller, P. A. (1987). The relation of empathy to prosocial and related behaviors. Psychological Bulletin, 101, 91119. Goleman, D. (1995). Emotional intelligence. New York, NY: Bantam Books. Goleman, D. (1998). What makes a leader? Harvard Business Review, NovemberDecember, 93102. Jackson, D. N. (1984). Personality Research Form Manual. Port Huron, MI: Research Psychologist Press. James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 8598. James, L. R., Demaree, R. G., & Wolf, G. (1993). Rwg: an assessment of within-group interrater agreement. Journal of Applied Psychology, 78, 306309. Locke, E. A. (1991). The essence of leadership. The four keys to leading successfully. New York, NY: Lexington Books. Martinez-Pons, M. (1997). The relation of emotional intelligence with selected areas of personal functioning. Imagination, Cognition and Personality, 17, 313. Martinez-Pons, M. (1998). Parental inducement of emotional intelligence. Imagination, Cognition and Personality, 18, 323. Mayer, J. D., & Salovey, P. (1997). What is emotional intelligence?. In P. Salovey, & D. Sluyter (Eds.), Emotional development and emotional intelligence: educational implications. New York, NY: Basic Books. Megerian, L. E., & Sosik, J. J. (1996). An aair of the heart: emotional intelligence and transformational leadership. The Journal of Leadership Studies, 3, 3148. Petrides, K. V., & Furnham, A. (2000). On the dimensional structure of emotional intelligence. Personality and Individual Dierences, 29, 313320. Saarni, C. (1999). The development of emotional competence. New York, NY: The Guilford Press. Salovey, P., & Mayer, J. D. (1990). Emotional intelligence. Imagination, Cognition and Personality, 9, 185211. Salovey, P., Mayer, J. D., Goldman, S. L., Turvey, C., & Palfai, T. (1995). Emotional attention, clarity and repair: Exploring emotional intelligence using the trait meta-mood scale. In J. W. Pennebaker (Ed.), Emotion, disclosure, health. Washington, DC: American Psychological Association. Schneider, B., Ehrhart, M. G., & Holcombe, K. M. (in press). Leadership in adolescence: comparing peer and teacher perspectives and correlates. Leadership Quarterly. Schneider, B., Paul, M. C., White, S. S., & Holcombe, K. M. (1999). Understanding high school student leaders, I: predicting teacher ratings of leader behavior. Leadership Quarterly, 10, 609636. Schutte, N. S., Malou, J. M., Hall, L. E., Haggerty, D. J., Cooper, J. T., Golden, C. J., & Dornheim, L. (1998). Development and validation of a measure of emotional intelligence. Personality and Individual Dierences, 25, 167177. Sosik, J. J., & Megerian, L. E. (1999). Understanding leader emotional intelligence and performance. Group and Organization Management, 24, 367390. Weisinger, H. (1998). Emotional intelligence at work. San Francisco, CA: Jossey-Bass Publishers. Yukl, G. (1998). Leadership in organizations. Upper Saddle River, NJ: Prentice Hall. Zacharatos, A., Barling, J., & Kelloway, E. K. (2000). Development and eects of transformational leadership in adolescents. Leadership Quarterly, 11, 211226.

Vous aimerez peut-être aussi