26
Developing and measuring the emotional intelligence of leaders Kevin S. Groves and Mary Pat McEnrue Department of Management, College of Business & Economics, California State University, Los Angeles, California, USA, and Winny Shen Department of Psychology, College of Natural & Social Sciences, California State University, Los Angeles, California, USA Abstract Purpose – The purpose of this study is to empirically test whether it is possible to deliberately develop emotional intelligence (EI) as conceptualized in the Mayer and Salovey model. Design/methodology/approach – This empirical study utilized a sample of 135 fully-employed business students in a treatment/control group research design in which treatment group participants underwent an intensive 11-week EI training program. Additional samples of 270 and 130 fully employed business students were utilized to develop an EI measure appropriate for EI development. Findings – The results indicate that EI can be deliberately developed; the treatment group demonstrated statistically significant overall EI gains and across each EI dimension, while the control group did not show any significant pre-/post-test differences. Practical implications – In addition to illustrating EI training best practices, a new EI measure is described that is appropriate for leadership development. Research limitations/implications – Research implications are discussed for the role of EI training in leadership development programs and fertile research directions for EI training. Originality/value – Emotional intelligence training has emerged into a popular and lucrative field, but empirical evidence on the deliberate development of EI has been substantially more elusive. This study provides an empirical EI training study that overcomes the conceptual and methodological limitations of extant research on the EI development process. Keywords Emotional intelligence, Training, Leadership development Paper type Research paper Introduction Management scholars, educators, and development practitioners have evidenced marked interest in understanding emotions in the workplace for more than a decade (e.g. Daus and Ashkanasy, 2005; Salovey and Mayer, 1990). Research has examined a range of diverse issues such as the role of affective experience in work motivation (Seo et al., 2004), the impact of emotion on purchasing decisions (Lerner et al., 2004), and the challenge of handling emotional interaction among members of multicultural teams (Von Glinow et al., 2004). More recently, empirical and theoretical studies on emotions have expanded to include topics particularly relevant to management development practitioners, such as the empirical relationship between leader emotional intelligence (EI) and transformational leadership (e.g. Barbuto and Burbach, 2006; Brown and Moshavi, 2005; Leban and Zulauf, 2004), the impact of EI on organizational change (Huy, 2002), and the link between EI and job satisfaction (Sy et al., 2006), life The current issue and full text archive of this journal is available at www.emeraldinsight.com/0262-1711.htm Emotional intelligence of leaders 225 Received 10 September 2006 Accepted 19 November 2006 Journal of Management Development Vol. 27 No. 2, 2008 pp. 225-250 q Emerald Group Publishing Limited 0262-1711 DOI 10.1108/02621710810849353

Developing and measuring the emotional intelligence of leaders

  • Upload
    winny

  • View
    214

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Developing and measuring the emotional intelligence of leaders

Developing and measuring theemotional intelligence of leaders

Kevin S. Groves and Mary Pat McEnrueDepartment of Management, College of Business & Economics,California State University, Los Angeles, California, USA, and

Winny ShenDepartment of Psychology, College of Natural & Social Sciences,

California State University, Los Angeles, California, USA

Abstract

Purpose – The purpose of this study is to empirically test whether it is possible to deliberatelydevelop emotional intelligence (EI) as conceptualized in the Mayer and Salovey model.

Design/methodology/approach – This empirical study utilized a sample of 135 fully-employedbusiness students in a treatment/control group research design in which treatment group participantsunderwent an intensive 11-week EI training program. Additional samples of 270 and 130 fullyemployed business students were utilized to develop an EI measure appropriate for EI development.

Findings – The results indicate that EI can be deliberately developed; the treatment groupdemonstrated statistically significant overall EI gains and across each EI dimension, while the controlgroup did not show any significant pre-/post-test differences.

Practical implications – In addition to illustrating EI training best practices, a new EI measure isdescribed that is appropriate for leadership development.

Research limitations/implications – Research implications are discussed for the role of EItraining in leadership development programs and fertile research directions for EI training.

Originality/value – Emotional intelligence training has emerged into a popular and lucrative field,but empirical evidence on the deliberate development of EI has been substantially more elusive. Thisstudy provides an empirical EI training study that overcomes the conceptual and methodologicallimitations of extant research on the EI development process.

Keywords Emotional intelligence, Training, Leadership development

Paper type Research paper

IntroductionManagement scholars, educators, and development practitioners have evidencedmarked interest in understanding emotions in the workplace for more than a decade(e.g. Daus and Ashkanasy, 2005; Salovey and Mayer, 1990). Research has examined arange of diverse issues such as the role of affective experience in work motivation (Seoet al., 2004), the impact of emotion on purchasing decisions (Lerner et al., 2004), and thechallenge of handling emotional interaction among members of multicultural teams(Von Glinow et al., 2004). More recently, empirical and theoretical studies on emotionshave expanded to include topics particularly relevant to management developmentpractitioners, such as the empirical relationship between leader emotional intelligence(EI) and transformational leadership (e.g. Barbuto and Burbach, 2006; Brown andMoshavi, 2005; Leban and Zulauf, 2004), the impact of EI on organizational change(Huy, 2002), and the link between EI and job satisfaction (Sy et al., 2006), life

The current issue and full text archive of this journal is available at

www.emeraldinsight.com/0262-1711.htm

Emotionalintelligence of

leaders

225

Received 10 September 2006Accepted 19 November 2006

Journal of Management DevelopmentVol. 27 No. 2, 2008

pp. 225-250q Emerald Group Publishing Limited

0262-1711DOI 10.1108/02621710810849353

Page 2: Developing and measuring the emotional intelligence of leaders

satisfaction (Law et al., 2004), job performance (e.g. Cote and Miners, 2006; Semadaret al., 2006), and organizational citizenship behavior (Cote and Miners, 2006).

Given mounting evidence that EI predicts a range of positive outcomes, there havebeen repeated calls in the research literature for empirical studies that establishwhether or not EI can be trained (e.g. Lopes et al., 2004; Gowing et al., 2006; Kerr et al.,2006; Dulewicz and Higgs, 2004). Indeed, professionals from the academic, practitioner,and management education communities have clamored for high-quality empiricalresearch that addresses whether it is possible and how to develop the emotionalintelligence of leaders, particularly among those responsible for effectingorganizational change. Unfortunately, most information on this issue is anecdotal innature and widely dispersed across a host of academic books, professional websites,consulting literature, and trade magazines. However, several researchers haveconducted empirical studies to test whether it is possible to develop EI amongmanagers and prospective leaders. In fact, EI development among managers has beenaddressed in a dozen published and unpublished empirical studies, which arethoroughly reviewed by McEnrue et al. (2006a).

A critical analysis of these EI development studies indicates that they contain somelimitation that calls into question the results reported. In particular, the studies either:

. lack a control group;

. use an EI measure of unknown psychometric quality;

. fail to control for demographic factors such as age, gender or work experiencethat may account for the results obtained;

. measure training effects at markedly different time periods for control andtreatment groups (e.g. nine months versus immediately after training);

. provide no check on social desirability as an alternative explanation for resultsreported;

. rely on a conceptually suspect EI model;

. involve training that does not focus specifically on emotional intelligence;

. use diagnostic and training criteria that appear devoid of emotional content; or

. provide no information about the training process at all.

In fact, extant empirical research typically suffers simultaneously from more than oneof these limitations. Thus, while developing managers’ leadership capabilities throughEI training is an attractive proposition to management development professionals anda dozen studies have been conducted across various disciplines, extant research doesnot provide solid evidence concerning whether and/or how to accomplish thisobjective.

The purpose of this study is three-fold. First, based on our evaluation of theavailable EI models and measures according to their leadership development utilityand psychometric properties, we set out to develop an instrument ideally suited for amanagement training program based on the Mayer and Salovey (1997) EI model. Next,we set out to study whether or not a deliberate EI training program, aided by aconceptually valid and practical EI measure, will enhance training participants’emotional intelligence. Finally, we attempt to disentangle the active components of anEI training program through detailed reporting of an 11-week treatment intervention.

JMD27,2

226

Page 3: Developing and measuring the emotional intelligence of leaders

This article is organized as follows. First, we briefly review the nature of extantresearch on EI training. We discuss important upfront decisions that impactdevelopmental outcomes, such as the selection of an EI conceptual model and measure.We then describe an EI training program and research study expressly designed toaddress the threats to internal and external validity apparent in existing research. Aspart of this effort, we developed and herein describe an EI measure specifically createdfor leadership development purposes. Finally, we report the results of the research weconducted, discuss the implications for leadership development programs, and offerseveral suggestions for future EI training research.

This article adds value to the existing body of research on EI development amongmanagers in several ways. First, we identify and systematically address many of theproblems in early research on EI development thereby enhancing confidence in theresults we report. Additionally, we thoroughly describe the training that we carried outso that it will be possible for others to ascertain exactly what we did and replicate ormodify it to better understand the dynamics of EI development. The study offersempirical evidence that management and leadership development professionals canuse to decide whether and how to engage in EI training within their organization.Lastly, we outline a set of recommendations that researchers can pursue in the future toadvance the field.

EI training: model and measure considerationsPart of the difficulty in answering whether or not EI can be developed amongmanagers lies in the nature of EI. Those unfamiliar with theory and research in thefield are frequently unaware that EI research emanates from four different models ofthe phenomenon each having different assumptions, dimensions, and measures. For acomprehensive review of the major models of EI, the various measures of each, andtheir respective utility for training and development purposes, please refer to McEnrueand Groves (2006).

The importance of the decision to select an EI model for training purposes cannot beoverstated since it should dictate the skills being developed, method and duration ofdevelopment, and the type of measure employed. Unfortunately, a review of theliterature demonstrates that this is often not the case. EI training and developmentprograms are often based upon EI models comprised of mostly personality traits,which are by definition not amenable to change (McEnrue et al., 2006a).

Daus and Ashkanasy (2005) have asserted that such models “may indeed be usefulfor organizational development and interventions, but they are much too broad inscope, and do not appear to markedly differ from traditional personality models orcompetency models” (p. 69). They suggest that although practitioners and researchersmay derive multiple benefits from such models as consulting tools, they should “notconfuse them with emotional intelligence” (p. 443). Thus, training practitionersattempting to develop EI among managers would do well to select a model that clearlydelineates EI skills and abilities that can be reliably assessed before and after theintervention.

For purposes of the present study, we chose to design the training program basedupon the Mayer and Salovey (1997) EI model. Mayer and Salovey define emotionalintelligence as “the ability to perceive accurately, appraise, and express emotion; theability to access and/or generate feelings when they facilitate thought; the ability to

Emotionalintelligence of

leaders

227

Page 4: Developing and measuring the emotional intelligence of leaders

understand emotion and emotional knowledge; and the ability to regulate emotions topromote emotional and intellectual growth” (p. 5). We concur with other researchers(e.g. Ashkanasy and Daus, 2005; Daus and Ashkanasy, 2005; Brackett and Mayer,2003) that this model of EI holds advantage in comparison to others because it has beenmost successfully differentiated from traits (e.g. personality), social desirability, andcognitive intelligence (e.g. Palmer et al., 2005; Daus and Ashkanasy, 2005; Day andCarroll, 2004; O’Conner and Little, 2003; Salovey et al., 2003).

The primary measure based upon the Mayer and Salovey (1997) model is theMayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT; Mayer et al., 2003).Overall, extant research suggests that the MSCEIT demonstrates strong psychometricproperties, including construct, convergent, discriminant, and predictive validitiescompared to competing EI measures (Daus and Ashkanasy, 2005; McEnrue andGroves, 2006; Day and Carroll, 2004; O’Conner and Little, 2003; Brackett and Mayer,2003). Given the present study’s primary focus on EI development, we specificallyexamined the measure for its utility for management development applications. On thebasis of this analysis, we concluded that the MSCEIT’s face and content validity posedifficulties for EI training activities (McEnrue and Groves, 2006). In particular, traineesmay question the relevance of test items that focus on sensations and those that requirerespondents to identify varying degrees of emotion conveyed by landscape andabstract designs. Items such as these and others may provoke management trainees toquestion whether developing EI is a relevant endeavor, which forces developmentpractitioners to spend more time up-front establishing the importance and centrality ofEI to leadership performance. Perhaps more important for training applications, theaction implications of employee responses on the MSCEIT are not apparent from manyof the items, particularly those which measure the process of using emotions tofacilitate thinking. Thus, it may be necessary for development practitioners to providesubstantial ongoing assistance as employees work to enhance their EI. Leadershipdevelopment professionals who emphasize self-directed learning (SDL), which callsupon employees to assume significant responsibility for their own learning anddevelopment, may find, as a consequence, that use of the MSCEIT may not yield largepositive effects (Ellinger, 2004).

Hence, we decided to construct a tool specifically designed for training applicationsthat would tap each of the four dimensions of EI in the Mayer and Salovey (1997)model: perception and appraisal of emotions; facilitating thinking with emotions;understanding complex emotions; and the regulation and management of emotions.We used this tool to assess the EI of trainees before and after participating in an11-week leadership development program focused on EI. Before doing so, weestablished the validity of the test for assessing EI. In designing the training, weexamined management and leadership development methods previously used byothers (e.g. Day, 2001) and focused particularly on identifying best practices fromextant EI training studies.

EI training: research findingsThe EI training and development literature is littered with studies that attest tosuccessfully developing EI among employees and managers (e.g. Murray et al., 2005;Boyatzis et al., 2002). Upon close inspection of this literature, however, there appear tobe very few published reports of well-designed, psychometrically rigorous EI training

JMD27,2

228

Page 5: Developing and measuring the emotional intelligence of leaders

studies. Extant empirical EI training studies are limited by one or more research designcharacteristics. For example, several studies omit key information concerning thetraining treatment and/or fail to formally test hypotheses while others rely uponanecdotal reports as the basis of their results. McEnrue et al. (2006a) were unable tocome to a conclusion about the feasibility of EI training and development becausemany of these studies were reported without crucial information (e.g. duration, trainingactivities, research design, duration, intensity) or were conducted using ex post factodata. Overall, the available EI training studies were limited by three general issues:

(1) EI conceptual and measurement concerns;

(2) limited information concerning the training treatment and short duration oftreatment; and

(3) absence of a control group and/or necessary statistical controls.

A detailed description of these limitations is provided in McEnrue et al. (2006a). Below,we briefly review the EI training studies across each of these categories.

Several EI training studies adopted EI models and measures that appear to taptraits and outcomes that are similar to yet clearly distinct from EI skills and abilities.For example, a longitudinal study by Boyatzis et al. (2002) demonstrated that MBAstudents enhanced their cognitive, self-management, and relationship managementcompetencies (e.g. goal-setting, action, initiative, leadership, etc.) as a result ofcompleting an MBA program. Similarly, Jaeger (2003) and Dulewicz et al. (2003)utilized Bar-On’s (1997) trait-based EI definition and EQ-i measure, which includepersonal characteristics such as flexibility, conscientiousness, and intuitiveness. Asothers have argued convincingly (e.g. Ashkanasy and Daus, 2005; McEnrue andGroves, 2006), these competencies and personal characteristics appear clearly distinctfrom EI skills and abilities.

In the same vein, EI training studies by Latif (2004), Clark et al. (2003), and Mausolffet al. (2006) utilized pre-/post-training EI measures for which no psychometricproperties were provided. Latif created an ad hoc instrument, Clark et al. (2003) utilizedthe EQ-map (Cooper and Sawaf, 1996), and Mausolff et al. (2006) developed an ad hocmeasure based upon Schutte et al.’s (1998) SREI. While these studies may haveprovided encouraging pre-/post-training results, it is very difficult to judge theirfindings given that they did not provide information concerning the EI measures’reliability and validity (Zeidner et al., 2004).

A second key limitation concerning extant EI training studies is a general lack ofinformation concerning the training activities and the overall nature and duration ofthe training treatment when specified. For example, studies by Murray and Lawrence(2006) and Murray and Jordan (2006) do not provide any information concerning thenature of the training in which the treatment groups participated (e.g. duration,process, content, methods and activities, etc.). Meyer et al. (2004) conducted a one-daytraining and development course that sought to enhance the intra- and interpersonalskills development of 15 administrators and dentists using a ropes and challengecourse. While improvements were noted using the MSCEIT to assess EI, none of theresults reached statistical significance. An inadequate amount of training received mayvery well account for the results reported. However, Meyer et al. (2004) noted that theirresults may be influenced by several other factors including a small sample size, a

Emotionalintelligence of

leaders

229

Page 6: Developing and measuring the emotional intelligence of leaders

ceiling effect due to high pretest scores, and/or the fact that the training interventionwas not focused specifically on enhancing EI per se.

The third general limitation of extant EI training research concerns the use of acontrol group and/or statistically controlling for variables likely to effecttreatment/control group differences. Sala’s (2002) pre-/post-workshop EI assessmentsdemonstrated significant improvements on eight out of 20 competencies of theEmotional Competency Inventory (ECI), a pre-cursor to the current ECI-2. However, henoted that the use of good demographic data (e.g. age, years of service) to control forother confounding variables would have permitted a stronger assertion about thetraining program’s results. Moreover, he cautioned, “without an adequate controlgroup it is difficult to isolate the impact of the intervention versus that of othervariables that may have contributed to higher scores” (p. 24). EI training studies byMurray et al. (2005) and Murray and Lawrence (2006) also may have been limited by alack of statistical controls for participant demographics. While there were significanttreatment and control group differences in Work Group Emotional Intelligence Profile(WEIP-6) post-test scores, these observed differences may have been due to the age andgender differences. For example, participants in the interpersonal skills training(control) group worked in lower seniority organizational levels, were significantlyolder, and included far fewer women. Lack of control for these factors is problematicsince some existing research indicates that EI may co-vary directly with age andgender (e.g. Goldenberg et al., 2006; Palmer et al., 2005 Ciarrochi et al., 2000).

In addition to the importance of controlling for gender, age, and work experience inEI training studies, extant research suggests that participant social desirability is acritical concern that should be controlled to attain strong empirical findings. In areview of empirical research on self-report EI measures, Tett et al. (2005) concludedthat “given the practical value of EI, self-reports are likely susceptible to the tendencyto portray oneself in a favorable light” (p. 9). The researchers also noted that none ofthe existing EI measures were developed with specific attention to items that are morestrongly associated with the construct of interest than with a measure of socialdesirability. Research by Hemmati et al. (2004), Saklofske et al. (2003), and Petrides andFurnham (2000) also found a positive EI-social desirability relationship and suggestedthat EI measures must be developed and tested to minimize susceptibility to socialdesirable and acquiescent responding. Most recently, Mesmer-Magnus andViswesvaran (2006) found that EI as measured by the Wong and Law (2002) scale,which is based upon the Salovey and Mayer (1997) model, yielded a significantcorrelation (r ¼ 0:44, p , 0:05) with social desirability. Overall, these findingsstrongly suggest that EI measures utilized for EI training purposes shoulddemonstrate clear divergence from social desirability in order to gain confidence inpre-/post-training results.

Together, the studies reviewed above indicate that deliberate EI training anddevelopment may be effective. However, the EI conceptual and measurement concerns,short duration of and lack of information concerning the training treatment, and theabsence of a control group and/or necessary statistical controls restricts confidence inEI training results reported. Thus, we set out to learn from the extant EI trainingresearch and to design a study that addresses the aforementioned limitations. Westarted by evaluating all available EI models and measures according to theirleadership development utility and psychometric properties (McEnrue and Groves,

JMD27,2

230

Page 7: Developing and measuring the emotional intelligence of leaders

2006). Utilizing this review, we set out to develop an instrument ideally suited for amanagement training program based on the Salovey and Mayer (1997) model. As notedearlier, while the MSCEIT demonstrates strong construct and predictive validity, themeasure’s low face validity and lack of items that generate action-planningimplications present serious limitations for development practitioners. Next, weexamine whether or not a deliberate EI training program, aided by a conceptually validand practical EI measure, will enhance training participants’ emotional intelligence.Finally, we attempt to disentangle the active components of an EI training programthrough detailed reporting of an 11-week treatment intervention.

MethodParticipantsA total of 535 fully employed business students at a medium-sized public university inthe Southwestern USA participated in this research. Of the total, 270 and 130 studentswere recruited to develop and to validate an EI measure for training purposes basedupon the Mayer and Salovey (1997) model. Hereafter, these are referred to as Sample 2and Sample 3, respectively. The remaining individuals (n ¼ 135), hereafter referred toas Sample 1, were either members of the control group (n ¼ 60) or treatment group(n ¼ 75) who participated in training designed to enhance EI. Sample 1 had 8.17(SD ¼ 3:88) mean years of full-time work experience with 84.2 percent having at leastone year of managerial work experience. Of those who reported their gender, 62.6percent were female and 37.8 percent were male. Their mean age was 28.83 years(SD ¼ 6:84). The ethnic composition of the sample was as follows: 31.1 percentAsian-American, 36.3 percent Hispanic/Latin-American, 17.8 percent Anglo-American,5.2 percent African-American, and 9.6 percent responding with other ethnicclassifications. Based on non-significant chi-square and t-tests for groupequivalency, Samples 2 and 3 had comparable demographic characteristics toSample 1. All respondents were provided informed consent and participation wascompletely voluntary.

EI measure developmentSample 2 (n ¼ 270) was utilized to develop and analyze the factor structure of themeasure. A total of 128 self-descriptive items across each of the four branches and 16associated dimensions of Mayer and Salovey’s (1997) model were writtenindependently by two professors with collective expertise in organizationalbehavior, test construction, and management development. Four criteria wereapplied to generate and evaluate the items:

(1) clearly representative of the respective dimension of Mayer and Salovey’smodel;

(2) void of double-negatives, double-barreled content, and other complications;

(3) void of overly desirable or undesirable content; and

(4) relevant to leadership development applications and organizational phenomena.

After reviewing the items generated and seeking peer evaluation from a third expertsource, redundant and/or poor items were merged or dropped. A total of 113 items wereretained from this process for initial validation.

Emotionalintelligence of

leaders

231

Page 8: Developing and measuring the emotional intelligence of leaders

Sample 2 (n ¼ 270) completed the 113 items and a set of demographic questions. Thedirections instructed respondents to indicate the extent to which a series of behaviorswere descriptive of them by marking responses on a Likert-type scale of 1 (“stronglydisagree”) to 7 (“strongly agree”). We conducted an exploratory factor analysis of the EImeasure using their responses. The maximum likelihood method was utilized for factorextraction, and the Varimax rotation method was used to transform the final factorsolution into a simple solution for interpretation. Given that the items were derived fromMayer and Salovey’s (1997) model of four EI dimensions, we expected a four-factorsolution to emerge from the data. Initially, the exploratory analysis produced eight factorswith an eigenvalue greater than unity. Upon further investigation of the factor loadingsacross these factors, it was clear that the first four factors with the largest eigenvaluesrepresented the four hypothesized EI dimensions. In addition, the remaining four factorsappeared to capture only random error of individual items in that each of these factorsconsisted of a maximum of one item with a loading greater than 0.40.

To improve the measure’s psychometric properties, we selected only the six itemswith the largest factor loadings from each of the first four factors. When a secondfactor analysis was conducted with these 24 items, a clear four-factor solution emerged.Table I displays the results of this factor analysis. The eigenvalues and percentage of

Items Factor 1 Factor 2 Factor 3 Factor 4

PA1 0.66 20.06 0.10 0.04PA2 0.74 0.18 0.01 0.10PA3 0.64 20.03 0.22 0.02PA4 0.75 0.05 0.15 0.11PA5 0.70 0.12 0.09 20.02PA6 0.73 0.04 0.07 0.03FT1 0.11 0.60 0.05 0.11FT2 20.08 0.63 20.03 0.28FT3 0.13 0.60 0.09 20.02FT4 20.02 0.71 0.13 0.21FT5 0.07 0.70 0.04 0.12FT6 0.10 0.68 0.03 0.07UE1 0.16 0.25 0.80 0.20UE2 0.04 0.13 0.74 20.01UE3 0.09 0.03 0.76 0.04UE4 0.18 0.10 0.72 20.02UE5 0.03 0.02 0.66 0.04UE6 20.01 20.03 0.78 0.03RE1 0.09 0.12 0.01 0.76RE2 0.03 0.07 0.17 0.80RE3 0.01 0.23 20.21 0.67RE4 0.07 0.05 0.02 0.79RE5 0.19 0.09 0.12 0.60RE6 0.03 0.11 0.04 0.58Eigenvalue 4.86 3.67 2.37 2.10Percentage of variance explained 26.02 18.26 9.89 8.77

Notes: n ¼ 270; PE, FT, UE, and RE represent perception/appraisal of emotions, facilitating thinkingwith emotions, understanding emotions, regulation/management of emotions, respectively

Table I.Factor analysis resultsfor Sample 2

JMD27,2

232

Page 9: Developing and measuring the emotional intelligence of leaders

explained variance across the four factors were 4.86 (26.02 percent), 3.67 (18.26percent), 2.37 (9.89 percent), and 2.10 (8.77 percent), respectively. Overall, thefour-factor solution explained 62.94 percent of the total variance. The mean loading ofthe 24 items on their respective EI dimensions was 0.69, while the cross loadings wereinsignificant. The four factors are labeled according to their respective items’ content,which matches the four dimensions of Mayer and Salovey’s (1997) model:perception/appraisal of emotion (PA), facilitating thinking with emotions (FT),understanding emotions (UE), and monitoring and regulation of emotions (RE). Wealso named the measure the emotional intelligence self-description inventory (EISDI).The items are provided in the Appendix 1.

In order to obtain discriminant validity estimates, a subset of Sample 2 (n ¼ 64) alsocompleted the 44-item John and Srivastava (1999) Big Five personality assessment.The alpha reliabilities for conscientiousness, openness to experience, neuroticism,extraversion, and agreeableness dimensions were 0.73, 0.77, 0.77, 0.85, and 0.74,respectively. We also used a subset of Sample 2 (n ¼ 33) to establish the test-retestreliability of the EI measure over a two-week period.

In order to estimate the EISDI’s relationship with personality dimensions, weexamined the inter-correlations among the scale scores and Big Five personalitydimensions (see Table II). The range of correlations among the EI scales andpersonality dimensions was 20.20 to 0.34 (n ¼ 68). Overall, the pattern of correlationsamong the EI factors and Big Five personality dimensions suggests that while thesemeasures share a moderate amount of variance, they are clearly not tapping identicalconstructs. The correlations among the four EISDI dimensions were all moderatelycorrelated (ranging from r ¼ 0:30 to 0.48), which suggests that they share somevariance but are not identical dimensions. Internal reliability estimates for the PA, FT,UE, and RE factors were 0.80, 0.72, 0.82, and 0.79, respectively. Furthermore, thetwo-week test-retest reliabilities for the PA, FT, UE, and RE factors were 0.79, 0.75,0.83, and 0.81, respectively (n ¼ 33). Taken together, these results provide preliminaryevidence for the structure and scale reliability of the EISDI.

Measure validationSample 3 (n ¼ 130) was utilized to obtain additional EISDI discriminant andconvergent validity estimates. Along with the EISDI, this sample completed twocomparable EI measures (Wong and Law, 2002; Brackett et al., 2005), a socialdesirability index (Crowne and Marlowe, 1960) and a personality measure (John andSrivastava, 1999). The alpha reliabilities across the four EISDI dimensions for thissample were 0.85, 0.76, 0.83, and 0.78, respectively, and the overall alpha was 0.91. TheBig Five personality dimensions – agreeableness, conscientiousness, extraversion,neuroticism, and openness to experience – also demonstrated acceptable reliabilitywith alphas of 0.72, 0.73, 0.85, 0.77, and 0.88, respectively.

The four-dimension, 16-item Wong and Law (2002) measure (WLEIS) and Brackettet al.’s (2005) five dimension, 19-item self-rated emotional intelligence scale (SREIS)were utilized to estimate the EISDI’s convergent validity because both self-reportmeasures are concise and based upon the Mayer and Salovey (1990, 1997) model. Thealpha reliabilities across the four Wong and Law dimensions were 0.83, 0.77, 0.86, and0.85, respectively, and the overall alpha was 0.88. The alpha reliabilities across theBrackett et al. (2005) dimensions were 0.83, 0.75, 0.79, 0.80, and 0.83, respectively, and

Emotionalintelligence of

leaders

233

Page 10: Developing and measuring the emotional intelligence of leaders

Mea

nS

D1

23

45

67

89

1.P

A5.

011.

00(0

.80)

2.F

T4.

980.

900.

32*

*(0

.72)

3.U

E5.

100.

980.

48*

*0.

34*

(0.8

2)4.

RE

5.40

0.95

0.42

**

0.30

**

0.47

**

(0.7

9)5.

BF

-A6.

581.

040.

25*

0.18

*0.

26*

0.24

*(0

.74)

6.B

F-C

6.79

0.97

0.34

**

0.10

0.33

**

0.32

**

0.27

**

(0.7

6)7.

BF

-E6.

051.

160.

32*

*0.

110.

30*

0.34

**

0.18

*0.

25*

(0.8

5)8.

BF

-N4.

781.

112

0.19

*2

0.06

20.

20*

20.

132

0.23

*2

0.18

*2

0.23

*(0

.73)

9.B

F-O

6.67

0.75

0.34

**

0.07

0.32

**

0.30

**

0.14

0.32

**

0.44

**

20.

03(0

.75)

Notes:

PA

,F

T,

UE

,an

dR

Ere

pre

sen

tp

erce

pti

on/a

pp

rais

alof

emot

ion

s,fa

cili

tati

ng

thin

kin

gw

ith

emot

ion

s,u

nd

erst

and

ing

oth

ers’

emot

ion

s,an

dre

gu

lati

on/m

anag

emen

tof

emot

ion

s,re

spec

tiv

ely

.B

F-A

,B

F-C

,B

F-E

,B

F-N

,an

dB

F-O

rep

rese

nt

the

Big

Fiv

ep

erso

nal

ity

fact

ors,

i.e.

agre

eab

len

ess,

con

scie

nti

ousn

ess,

extr

over

sion

,n

euro

tici

sm,

and

open

nes

sto

exp

erie

nce

,re

spec

tiv

ely

.N

um

ber

son

the

dia

gon

alar

eco

effi

cien

tal

ph

as.

* p,

0:05

(tw

o-ta

iled

test

s);

** p

,0:

01(t

wo-

tail

edte

sts)

Table II.Descriptive statistics andcorrelation coefficientsfor Sample 2

JMD27,2

234

Page 11: Developing and measuring the emotional intelligence of leaders

the overall alpha was 0.81. Because of reported concerns that EI measures sharesignificant variance with social desirability (e.g. Tett et al., 2005), the Crowne andMarlowe (1960) social desirability index (a ¼ 0:75) was used to estimate the EISDI’srelationship with social desirability.

We conducted a confirmatory factor analysis of the EISDI responses from usingAMOS data analysis software (Arbuckle, 2003). Postulating that each item would loadonto its respective EI dimension, the measurement model fit the data reasonably well(x 2 ¼ 345:5, df ¼ 190) according to the following fit indices: GFI (0.93), AGFI (0.89),IFI (0.95), CFI (0.94), and RMSEA (0.05). Furthermore, maximum likelihood estimatesacross all four dimensions showed that each item loaded significantly onto itsrespective scale. Correlations among the four EI dimensions ranged from 0.28 to 0.46,results that are very similar to that of Sample 2.

Table III presents correlation results that indicate both the Wong and Law (2002)and Brackett et al. (2005) EI measures not only demonstrated stronger relationshipssocial desirability than did the EISDI but also with the five personality dimensionsthan did the EISDI. The social desirability-EISDI correlation was not significant (0.21)while both the WLEIS (r ¼ 0:43, p , 0:01) and SREIS (r ¼ 0:37, p , 0:01) weresignificantly associated with social desirability. The mean correlations across thepersonality dimensions for the EISDI, Wong and Law, and Brackett et al. measureswere 0.29, 0.41, and 0.37, respectively. Also, the EISDI correlated strongly with both theWong and Law (r ¼ 0:59, p , 0:01) and Brackett et al. (r ¼ 0:55, p , 0:01) measures,suggesting evidence of the measure’s convergent validity. However, the magnitude ofthese relationships was significantly greater than that of the EISDI’s relationships withpersonality dimensions, providing evidence of discriminant validity. Overall, theseresults extend support for the four-factor structure of the EISDI and provide additionalsupport for its convergent and discriminant validity.

Treatment/control group procedureSample 1 (n ¼ 135) was utilized for a treatment/control group EI training study. Asubset of Sample 1 (n ¼ 60) represented the control group and participated in anundergraduate management course comprised of common learning activities includinglecture and class discussions, case analyses, a team field research assignment, andstudent presentations. None of the control group learning activities addressed EI oremotion-related topics. The other subset (n ¼ 75) of Sample 1 participated in aleadership development program that focused on enhancing their emotionalintelligence. All members of treatment and control groups completed the EISDI anda series of demographic questions at the beginning and end of an 11-week period. Theresults of the EISDI administration for the treatment group were utilized to the assessparticipants’ EI ability across the four EI dimensions and design their subsequenttraining and development activities. To test for equivalency of the treatment andcontrol groups, t-tests and chi-square analyses were conducted across the demographicvariables (gender, work experience, racial/ethnic group), personality dimensions (Johnand Srivastava, 1999), and social desirability (Crowne and Marlowe, 1960). The resultsof these tests revealed no significant treatment and control group differences.

The treatment group training was intense, customized to the developmental goalsthat participants established, and focused on behavioral change. The aim was to notonly enhance participants’ understanding of four abilities incorporated in the Mayer

Emotionalintelligence of

leaders

235

Page 12: Developing and measuring the emotional intelligence of leaders

Mea

nS

D1

23

45

67

89

1.E

ISD

I5.

160.

81(0

.91)

2.W

LE

IS5.

640.

780.

59*

*(0

.88)

3.S

RE

IS3.

560.

500.

55*

*0.

66*

*(0

.81)

4.M

/C49

.05.

460.

210.

43*

*0.

37*

*(0

.75)

5.B

F-A

6.69

1.10

0.24

**

0.40

**

0.26

**

0.44

**

(0.7

2)6.

BF

-C6.

921.

120.

31*

*0.

46*

*0.

35*

*0.

32*

*0.

22*

*(0

.73)

7.B

F-E

5.91

1.48

0.33

**

0.31

**

0.48

**

0.19

*0.

120.

25*

*(0

.85)

8.B

F-N

4.47

1.42

20.

24*

20.

41*

*2

0.26

**

20.

41*

*2

0.35

**

20.

18*

*2

0.27

**

(0.7

7)9.

BF

-O6.

221.

150.

34*

*0.

48*

*0.

48*

*0.

100.

050.

39*

*0.

44*

*2

0.07

(0.8

8)

Notes:n¼

130;

EIS

DI,

WL

EIS

,an

dS

RE

ISre

pre

sen

tth

eem

otio

nal

inte

llig

ence

self

-des

crip

tion

inv

ento

ry,W

ong

and

Law

’s(2

002)

emot

ion

alin

tell

igen

cesc

ale,

and

Bra

cket

tetal.’

s(2

005)

self

-rat

edem

otio

nal

inte

llig

ence

scal

e,re

spec

tiv

ely

.BF

-A,B

F-C

,BF

-E,B

F-N

,an

dB

F-O

rep

rese

nt

the

Big

Fiv

ep

erso

nal

ity

fact

ors,

i.e.a

gre

eab

len

ess,

con

scie

nti

ousn

ess,

extr

over

sion

,neu

roti

cism

,an

dop

enn

ess

toex

per

ien

ce,r

esp

ecti

vel

y.N

um

ber

sin

the

dia

gon

alar

eco

effi

cien

tal

ph

as.

* p,

0:05

(tw

o-ta

iled

test

s);

** p

,0:

01(t

wo-

tail

edte

sts)

Table III.Descriptive statistics andcorrelation coefficientsfor Sample 3

JMD27,2

236

Page 13: Developing and measuring the emotional intelligence of leaders

and Salovey (1997) model but to stimulate and realize change in at least two elementsof the 16 which comprise these abilities. Central to the learning and behavioraloutcomes of the course, the first activity involved completion of the EISDI. The natureof the instrument was key in aiding participants to create and implement an effectivedevelopment plan. The EISDI permitted participants to fully understand themulti-dimensional nature of emotional intelligence, and appreciate the differencebetween emotional intelligence and other phenomena measured during the trainingsuch as personality (openness to experience, conscientiousness, extroversion,agreeableness, and neuroticism). Moreover, it provoked detailed and substantialdiscussion about the application and value of emotional intelligence at work and in theleadership process. Using the EISDI allowed the instructor to design leaderskill-building exercises to enhance emotional intelligence with explicit reference to theabilities participants had targeted for development. Appendix 2 outlines theseexercises. The method, duration, and intensity of the training process were similar tothat designed and described by Bass and Avolio (1992) to develop transformationalleadership. The EISDI played a key role in the training process here in the same waythe MLQ (Avolio et al., 1999) facilitated early transformational leadership developmentprograms.

ResultsFactor structureTo provide factor structure results from a second cross-validation sample, weconducted a confirmatory factor analysis of the 24-item EISDI using AMOS (Arbuckle,2003). All 135 participant responses from the pre-course administration (Time 1) of themeasure were utilized for analysis. Again postulating that each item would load ontoits respective EI dimension, the measurement model fit the data well (x 2 ¼ 323:7,df ¼ 193) according to the following fit indices: GFI (0.95), AGFI (0.90), IFI (0.94), CFI(0.93), and RMSEA (0.05). Furthermore, maximum likelihood estimates across all fourdimensions ranged from 0.48 to 0.82, suggesting that each item loaded significantlyonto its respective scale. Correlations among the four EI dimensions ranged from 0.29to 0.45, results that are very similar to those from Samples 2 and 3.

Group comparisonsIn order to estimate the comparative changes in participants’ EI skills across thetreatment and control groups, pre-/post-course t-tests using the Bonferroni statistic tocontrol for alpha inflation were conducted for the overall EISDI score and fourdimensions. Table IV presents the results from these analyses. The pre-/post-coursemean differences for the treatment group demonstrated that participants significantlyimproved their overall EI score and all four EI dimensions. All of the mean differenceswere in the expected direction, statistically significant, and ranged from 20.21(tð74Þ ¼ 21:79, p , 0:05) to 20.62 (tð74Þ ¼ 3:41, p , 0:01). By comparison, none ofthe pre-/post-course mean differences for the control group were statisticallysignificant, although each score slightly improved and two scores (FT and RE) weremarginally significant (tð59Þ ¼ 1:58, p , 0:10 and tð59Þ ¼ 1:60, p , 0:10,respectively). Further t-test comparisons demonstrated that the treatment andcontrol group mean differences across all EI scores were statistically significant atp , 0:001. Overall, these results demonstrate that treatment group participants

Emotionalintelligence of

leaders

237

Page 14: Developing and measuring the emotional intelligence of leaders

Tre

atm

ent

gro

up

(n¼

75)

Con

trol

gro

up

(n¼

60)

Pre

-cou

rse

mea

n(S

D)

Pos

t-co

urs

em

ean

(SD

)M

ean

dif

fere

nce

(SD

)t

(df¼

74)

Pre

-cou

rse

mea

n(S

D)

Pos

t-co

urs

em

ean

(SD

)M

ean

dif

fere

nce

(SD

)t

(df¼

59)

EIS

DI

5.25

(0.6

9)5.

66(0

.69)

20.

41(0

.73)

24.

77*

**

5.11

(0.9

8)5.

24(0

.91)

20.

13(0

.50)

21.

39P

A5.

05(0

.83)

5.48

(0.7

6)2

0.43

(0.7

8)2

4.68

**

*5.

07(0

.96)

5.16

(0.8

5)2

0.09

(0.6

2)2

1.13

FT

5.08

(0.8

2)5.

51(0

.76)

20.

43(0

.80)

24.

64*

**

4.84

(0.9

9)5.

03(0

.83)

20.

19(0

.84)

21.

58*

**

*

UE

5.15

(0.9

0)5.

77(0

.69)

20.

62(0

.57)

23.

41*

*5.

05(0

.78)

5.16

(0.9

5)2

0.11

(0.8

6)2

1.32

RE

5.74

(0.8

2)5.

96(0

.95)

20.

21(0

.66)

21.

79*

5.48

(0.6

9)5.

67(0

.68)

20.

19(0

.70)

21.

60*

**

*

Notes:

PA

,F

T,

UE

,an

dR

Ere

pre

sen

tp

erce

pti

on/a

pp

rais

alof

emot

ion

s,fa

cili

tati

ng

thin

kin

gw

ith

emot

ion

s,u

nd

erst

and

ing

emot

ion

s,an

dre

gu

lati

on/m

anag

emen

tof

emot

ion

s,re

spec

tiv

ely

.* p

,0:

05;

** p

,0:

01;

**

* p,

0:00

1;*

**

*p,

0:10

Table IV.Pre-/post-course groupdifferences across EIdimensions for Sample 1

JMD27,2

238

Page 15: Developing and measuring the emotional intelligence of leaders

significantly enhanced their EI abilities by comparison to the control groupparticipants.

DiscussionThis research demonstrates that it is possible to enhance the emotional intelligence ofindividuals through deliberate training. While several published and unpublishedempirical studies suggest this possibility, they each appear to contain significantlimitations in research design that call into question the results reported. Additionally,in most cases, a description of the process, content, duration and/or method used toenhance EI is missing (e.g. Boyatzis et al., 2002; Murray and Lawrence, 2006; Latif,2004), which relegates the training process to a “black box”. The present study not onlyindicates that it is possible to enhance EI through deliberate training but details thetraining design and process so that future researchers can replicate the process, test thedurability of the results achieved, investigate various elements which may account fortraining gains, and examine other factors which may increase the efficacy of attemptsto enhance EI among managers and potential managers in organizations. For example,McEnrue et al. (2006b) identify a myriad of individual (e.g. openness to experience), job(e.g. emotional labor), and organizational (e.g. degree of change) factors which maymitigate or increase the effects of EI training. The power of these factors to enhance thevalue of EI training and to boost training outcomes deserves consideration among bothresearchers and management development practitioners.

EI measurementThe results presented here suggest that the EISDI, based upon the Mayer and Salovey(1997) model, may be an effective instrument for management development.Pre-/post-test differences demonstrated that participants who completed the EISDIas a diagnostic measure for development purposes, as well as engaged in EIskill-building activities, enhanced their overall EI and individual ability scoressignificantly more than control group participants. The EISDI has several advantagesover most other self-report measures based upon Mayer and Salovey’s (1990, 1997)early and later conceptualization of the construct.

The EISDI internal reliability estimates from three independent samples andtest-retest reliability from a sub-sample indicate acceptable scale reliability andconsistency over time. Moreover, it demonstrated good convergence with existing EImeasures based upon the Mayer and Salovey model and acceptable divergence fromboth personality dimensions and social desirability relative to the WLEIS or SREIS.The finding that the EISDI demonstrated acceptable discriminant validity vis-a-vis theBig Five personality variables is highly significant given criticism from scholars thatEI is little more than a repackaging of personality characteristics (e.g. Locke, 2005).Note that both the WLEIS and SREIS were strongly correlated (p , 0:01) with socialdesirability, while the EISDI-social desirability relationship was non-significant.

The EISDI not only addresses all four dimensions of the Mayer and Salovey modelbut also appears to demonstrate greater face validity and match the meaning of theconstruct more closely. For example, neither the WLEIS nor SREIS adequatelymeasure understanding emotions and using emotions to facilitate thinking. Absence ofthe former dimension is problematic in that more and more individuals inorganizations are operating in diverse ethnic and cultural environments either

Emotionalintelligence of

leaders

239

Page 16: Developing and measuring the emotional intelligence of leaders

domestically or internationally. Interpreting the meaning of emotions when workingwith employees from high context cultures and understanding complex emotions arelikely to be important skills for managers under these conditions. Absence of the latterdimension is problematic given that this dimension would likely have significantimpact on the processes of establishing priorities, enabling creativity, and makingsound decisions within organizations (Sadler-Smith and Shefy, 2004; Csikszentmihalyi,1996).

Implications for management developmentThe EISDI appears to offer diagnostic results that permit professionals charged withenhancing employees’ EI to better reference the management development activities inwhich they engage to initial and ongoing diagnoses of EI ability, particularly incomparison with the MSCEIT. However, the EISDI is a self-report measure and istherefore subject to criticism from those who assert that individuals are a poor sourceof information concerning themselves and their abilities. Several researchers suggestthat an ability measure such as the MSCEIT is superior to self-report data (notreferenced against external opinion) by citing studies on self-reported verbal abilityversus tests of verbal ability (e.g. Mayer et al., 2003; Brackett et al., 2005). However,there is evidence that self-reported EI predicts emotionally intelligent behavior (e.g.Austin, 2004) and that the utilization of self-report measures is preferable fordevelopment applications. For example, Murray and Lawrence’s (2006) 18-monthstudy of self-report EI measures in training research demonstrated that self-reportassessments were very effective in determining changes in trainees’ EI skills andabilities, and “can be robust to respondent bias arising from factors such as transitorymood changes, the guinea pig effect, and the Hawthorne effect” (p. 4). Furthermore,Humphrey et al. (2006) assert that individuals are most often in the best position toreport on their own emotional states. Moreover, they indicate that one’s motivation toinflate EI self-ratings is no greater than for personality traits, attitudes, and otherconstructs commonly measured by self-reports.

Given that the EISDI was designed for developmental purposes, rather than formore response-inducing performance evaluation and compensation purposes, webelieve that the instrument provides accurate EI assessments for managementdevelopment practitioners. The treatment group participants engaged in ongoingdialogue with and received feedback from five sources: peers, a coach, an externalsource of support (e.g. boss, spouse, or coworker), the instructor, and themselves asthey reflected on and formally recorded their experience in the EI development processvis-a-vis the goals they established. Hence, it is reasonably unlikely that the EI ratingsthey assigned to themselves were a product of acquiescence or motivation to appearcompetent. The treatment group participants had ample opportunity to find out where,how, and when they excelled across the EI skills and abilities from the vantage point offour other sources.

Directions for future researchBased on a thorough review of extant empirical studies, McEnrue et al. (2006a) offeredmore than a dozen specific recommendations to improve both the nature of EI trainingand the quality of research concerning methods to enhance the EI of leaders. Thewisdom of pursuing additional research regarding EI training and related issues

JMD27,2

240

Page 17: Developing and measuring the emotional intelligence of leaders

hinges on whether it is possible to enhance the EI of managers. This studydemonstrates that it is indeed possible to enhance the EI of managers. Hence, wehighlight some of these recommendations here. First, we believe that futuremanagement development researchers would do well to examine which factors, if any,are likely to mitigate or increase the value and effectiveness of training. These includeindividual characteristics such as openness to experience and receptivity to feedback(McEnrue et al., 2006b), job factors such as emotional labor (Wong and Law, 2002;Brotheridge and Lee, 2003), and organizational characteristics such as team-basedstructures (Sy and Cote, 2004) and the degree of change (radical or incremental) (Huy,2002; Groves, 2006). This has clear practical value to management developmentpractitioners, participants, and the organizations in which they work.

Additional research to address the potential influence of participant demographicson the process and outcomes of EI training is advisable. Some existing researchindicates EI co-varies with age, gender, and management level (e.g. Goldenberg et al.,2006; Palmer et al., 2005 Ciarrochi et al., 2000). Recent research conducted by Downeyet al. (2006) using female managers examined the relationships amongtransformational leadership, intuition, and EI. They reported that female managershigh on transformational leadership also demonstrated higher levels of EI andintuition. Research to disentangle and understand the dynamics among these variableswould be useful for a clearer understanding of the by-products of EI training.

Thirdly, we encourage researchers to examine whether some elements of EI areeasier to train than others and are predictive of different behavioral and otherperformance criteria (e.g. Dulewicz and Higgs, 2004). Recent empirical studies suggestthat EI dimensions may differentially facilitate the components of transformational,change-oriented leadership including individualized consideration (Barling et al., 1996),inspirational motivation (Leban and Zulauf, 2004), idealized influence (Barbuto andBurbach, 2006), and vision articulation/communication (Groves, 2006; Huy, 2002).Similarly, Kerr et al. (2006) found that only the experiential dimensions of the MSCEIT– understanding emotions and using emotions to facilitate thinking – were predictiveof leadership effectiveness in a study of managers at a large manufacturingorganization. This finding suggests that the understanding emotions and regulatingemotions dimensions may not serve a proximal role for leadership effectiveness, unitperformance, or other key outcomes.

Finally, we reiterate our call for EI development researchers to provide a detaileddescription of their training process and activities. Recent research by Elfenbein (2006),which focused on training students to recognize emotions in cross-cultural contexts,isolates and effectively illustrates the process of enhancing a single element ofemotional intelligence – the understanding of facial expressions. The current studyoffers a detailed account of the methods used to enhance the EI of managers andpotential leaders. More information concerning the duration, method, activities,process, criteria and other features of training in future research such as that providedhere is essential in order to better understand how to enhance EI and help identifythose interventions that are most effective.

LimitationsThe findings discussed above concerning the EI training outcomes and the EISDI’sutility for management development professionals should be viewed in the context of

Emotionalintelligence of

leaders

241

Page 18: Developing and measuring the emotional intelligence of leaders

several limitations. The EISDI’s concurrent and predictive validity must also beestablished through future empirical research (e.g. Dulewicz et al., 2003). This willrequire researchers to examine a range of theoretically relevant EI ability-performanceoutcomes such as speed of response, client satisfaction, customer retention, shrinkage,volunteer hours donated, promotability, team learning, and workplace accidents inaddition to those typically examined in extant research. Consistent with prior empiricalstudies on the Mayer and Salovey (1997) model, the EISDI’s further developmentshould be focused on establishing predictive validity vis-a-vis job performance in fieldscharacterized by high emotional labor, including police officers, emergency responders,and other crisis-laden positions (e.g. Daus et al., 2004), and sales/customer servicerepresentatives (e.g. Cage et al., 2004).

As discussed in detail above, the limitations of using a self-report EI measure in thepresent study were mitigated by a treatment intervention that involved multipleongoing sources of feedback with ample opportunity for trainees to estimate theirperformance and development across the EI dimensions. However, future research onthe EISDI will need to verify that training participants’ self-ratings are consistent withidentifiable, emotionally intelligent behavior. For example, future studies may want totrain multiple observers to assess trainees’ performance on several behavioralexercises in order to cross-validate self-report post-training EISDI ratings.

Although our results demonstrate empirical support for the EISDI’s utility fordeveloping EI abilities, there are several alternative explanations for the observedresults. For example, instructor skill and/or experience, student motivation to enhanceEI abilities, the efficacy of specific skill-building activities, and regression toward themean may explain the treatment and control group differences. However, the EISDIappears to offer diagnostic results that permit professionals charged with enhancingemployees’ EI to better reference the development activities in which they engage toinitial and ongoing diagnoses of EI ability.

ConclusionsThe field of emotional intelligence training has emerged into an incredibly popular andlucrative endeavor, yet the empirical evidence presented to date casts doubt on theeffectiveness of deliberate EI training among managers. Using the highly endorsedMayer and Salovey (1997) EI model, this study has overcome several of the conceptualand methodological limitations of extant research on the EI development process. It isintended to serve as a solid foundation on which to design and carry out additionalresearch. We hope that future management development professionals will find ourwork useful as a template for future training efforts and for empirical studies on theprocess of enhancing EI.

References

Arbuckle, J. (2003), AMOS 5.0, SmallWaters, Chicago, IL.

Ashkanasy, N. and Daus, C. (2005), “Rumors of the death of emotional intelligence inorganizational behavior are vastly exaggerated”, Journal of Organizational Behavior,Vol. 26, pp. 441-52.

Austin, E. (2004), “An investigation of the relationship between trait emotional intelligence andemotional task performance”, Personality and Individual Differences, Vol. 36, pp. 1855-64.

JMD27,2

242

Page 19: Developing and measuring the emotional intelligence of leaders

Avolio, B., Bass, B. and Jung, D. (1999), “Re-examining the components of transformational andtransactional leadership using the Multifactor Leadership Questionnaire”, Journal ofOccupational and Organizational Psychology, Vol. 72 No. 4, pp. 441-62.

Barbuto, J.E. and Burbach, M.E. (2006), “The emotional intelligence of transformational leaders:a field study of elected officials”, Journal of Social Psychology, Vol. 146, pp. 51-64.

Barling, J., Weber, T. and Kelloway, K. (1996), “Effects of transformational leadership training onattitudinal and financial outcomes: a field experiment”, Journal of Applied Psychology,Vol. 81 No. 6, pp. 827-32.

Bar-On, R. (1997), Bar-On Emotional Quotient Inventory: Technical Manual, Multi-HealthSystems, Inc., Toronto.

Bass, B. and Avolio, B. (1992), “Developing transformational leadership: 1992 and beyond”,Journal of European Industrial Training, Vol. 14 No. 5, pp. 21-7.

Boyatzis, R., Stubbs, E. and Taylor, S. (2002), “Learning cognitive and emotional intelligencecompetencies through graduate management education”, Academy of ManagementLearning and Education, Vol. 1 No. 2, pp. 150-62.

Brackett, M. and Mayer, J. (2003), “Convergent, discriminant, and incremental validity ofcompeting measure of emotional intelligence”, Personality and Social Psychology Bulletin,Vol. 29, pp. 1147-58.

Brackett, M., Rivers, S., Shiffman, S., Lerner, N. and Salovey, P. (2005), “What is the best way tomeasure emotional intelligence? A case for performance measures”, unpublishedmanuscript.

Brotheridge, C. and Lee, R. (2003), “Development and validation of the emotional labour scale”,Journal of Occupational and Organizational Psychology, Vol. 76 No. 3, pp. 365-79.

Brown, F.W. and Moshavi, D. (2005), “Transformational leadership and emotional intelligence: apotential pathway for an increased understanding of interpersonal influence”, Journal ofOrganizational Behavior, Vol. 26, pp. 867-71.

Cage, T., Daus, C. and Saul, K. (2004), “An examination of emotional skill, job satisfaction, andretail performance”, paper submitted to the 19th Annual Meeting of the Society ofIndustrial and Organizational Psychology, Chicago, IL.

Caruso, D. and Salovey, P. (2004), The Emotionally Intelligent Manager, Jossey-Bass,San Francisco, CA.

Ciarrochi, J.V., Chan, A.Y.C. and Caputi, P. (2000), “A critical evaluation of the emotionalintelligence construct”, Personality and Individual Differences, Vol. 28, pp. 539-61.

Clark, S.C., Callister, R. and Wallace, R. (2003), “Undergraduate management skills courses andstudents’ emotional intelligence”, Journal of Management Education, Vol. 27 No. 1, pp. 3-23.

Cooper, R. and Sawaf, A. (1996), Executive EQ, Berkley Publishing, New York, NY.

Cote, S. and Miners, C. (2006), “Emotional intelligence, cognitive intelligence, and jobperformance”, Administrative Science Quarterly, Vol. 51, pp. 1-28.

Crowne, D. and Marlowe, D. (1960), “A new scale of social desirability independent ofpsychopathology”, Journal of Consulting Psychology, Vol. 24, pp. 349-54.

Csikszentmihalyi, M. (1996), Creativity: Flow and the Psychology of Discovery and Invention,HarperCollins Publishers, New York, NY.

Daus, C. and Ashkanasy, N. (2005), “The case for an ability-based model of emotional intelligencein organizational behavior”, Journal of Organizational Behavior, Vol. 26, pp. 453-66.

Emotionalintelligence of

leaders

243

Page 20: Developing and measuring the emotional intelligence of leaders

Daus, C., Rubin, R. and Cage, T. (2004), “Police performance: do emotional skills matter?”, papersubmitted to the 19th Annual Meeting of the Society of Industrial and OrganizationalPsychology, Chicago, IL.

Day, D. (2001), “Leadership development: a review in context”, Leadership Quarterly, Vol. 11No. 4, pp. 581-613.

Day, A. and Carroll, S. (2004), “Using an ability-based measure of emotional intelligence topredict individual performance, group performance, and group citizenship behaviors”,Personality and Individual Differences, Vol. 36, pp. 1443-58.

Downey, L.A., Papageorgiou, V. and Stough, C. (2006), “Examining the relationship betweenleadership, emotional intelligence, and intuition in senior female managers”, Leadership &Organization Development Journal, Vol. 27 No. 4, pp. 250-64.

Dulewicz, V., Higgs, M. and Slaski, M. (2003), “Measuring emotional intelligence: content,construct and criterion-related validity”, Journal of Managerial Psychology, Vol. 18 No. 5,pp. 405-20.

Dulewicz, V. and Higgs, M. (2004), “Can emotional intelligence be developed?”, InternationalJournal of Human Resource Management, Vol. 15, pp. 95-111.

Elfenbein, H. (2006), “Learning in emotion judgments: training and the cross-culturalunderstanding of facial expressions”, Journal of Nonverbal Behavior, Vol. 30, pp. 21-36.

Ellinger, A. (2004), “The concept of self-directed learning and its implications for human resourcedevelopment”, Advances in Developing Human Resources, Vol. 6 No. 2, pp. 158-77.

Goldenberg, I., Matheson, K. and Mantler, J. (2006), “The assessment of emotional intelligence:a comparison of performance-based and self-report methodologies”, Journal of PersonalityAssessment, Vol. 86 No. 1, pp. 33-45.

Gowing, M., O’Leary, B. and Brienza, D. (2006), “A practitioner’s research agenda: exploringreal-world applications and issues”, in Druskat, V., Sala, F. and Mount, F. (Eds), LinkingEmotional Intelligence and Performance at Work: Current Research Evidence withIndividuals and Groups, Lawrence Erlbaum Associates, Mahwah, NJ, pp. 245-65.

Groves, K. (2006), “Leader emotional expressivity, visionary leadership, leadership effectiveness,and organizational change”, Leadership &Organization Development Journal, Vol. 27 No. 7,pp. 566-83.

Hemmati, T., Mills, J. and Kroner, D. (2004), “The validity of the Bar-On emotional intelligencequotient in an offender population”, Personality and Individual Differences, Vol. 37 No. 4,pp. 695-706.

Humphrey, R., Kellett, J. and Sleeth, R. (2006), “Empathy and the emergence of task and relationsleaders”, Leadership Quarterly, Vol. 17 No. 2, pp. 146-62.

Huy, Q. (2002), “Emotional balancing of organizational continuity and radical change: thecontribution of middle managers”, Administrative Science Quarterly, Vol. 47, pp. 31-69.

Jaeger, A. (2003), “Job competencies and the curriculum: an inquiry into emotion in graduateprofessional education”, Research in Higher Education, Vol. 44 No. 6, pp. 615-39.

John, O. and Srivastava, S. (1999), “The Big Five trait taxonomy: history, measurement, andtheoretical perspectives”, in Pervin, O. and John, L. (Eds), Handbook of Personality: Theoryand Research, Guilford Press, New York, NY, pp. 102-38.

Kerr, R., Garvin, J., Heaton, N. and Boyle, E. (2006), “Emotional intelligence and leadershipeffectiveness”, Leadership and Organizational Development Journal, Vol. 27 No. 4,pp. 265-79.

Latif, D. (2004), “Using emotional intelligence in the planning and implementation of amanagement skills course”, Pharmacy Education, Vol. 4 No. 2, pp. 81-9.

JMD27,2

244

Page 21: Developing and measuring the emotional intelligence of leaders

Law, K., Wong, C. and Song, L. (2004), “The construct and criterion validity of emotionalintelligence and its potential utility for management studies”, Journal of AppliedPsychology, Vol. 89 No. 3, pp. 484-96.

Leban, W. and Zulauf, C. (2004), “Linking emotional intelligence abilities and transformationalleadership styles”, Leadership & Organization Development Journal, Vol. 25, pp. 554-64.

Lerner, J., Small, D. and Lowenstein, G. (2004), “Heart strings and purse strings: carryover effectsof emotions on economic decisions”, Psychological Science, Vol. 15, pp. 337-41.

Locke, E. (2005), “Why emotional intelligence is an invalid concept”, Journal of OrganizationalBehavior, Vol. 26, pp. 425-31.

Lopes, P., Cote, S. and Salovey, P. (2004), “An ability model of emotional intelligence: implicationsfor assessment and training”, in Druskat, V., Sala, F. and Mount, G. (Eds), LinkingEmotional Intelligence and Performance at Work: Current Research Evidence withIndividuals and Groups, Lawrence Erlbaum Associates, Mahwah, NJ, pp. 53-80.

McEnrue, M.P. and Groves, K. (2006), “Choosing among tests of emotional intelligence: what’sthe evidence?”, Human Resource Development Quarterly, Vol. 17, pp. 9-42.

McEnrue, M.P., Groves, K. and Shen, W. (2006a), “Emotional intelligence training: evidenceregarding its efficacy for developing leaders”, Leadership Quarterly, under review.

McEnrue, M.P., Groves, K. and Shen, W. (2006b), “Emotional intelligence training: the role ofopenness to experience, self-efficacy, and receptivity to feedback in enhancing gainsachieved”, Journal of Management Development, forthcoming.

Mausolff, C., Bringman, N., Martinez, E. and Paggi, R. (2006), “Humanistic-experiential approachto training in emotional intelligence”, paper presented at the Academy of ManagementConference, Atlanta, GA.

Mayer, J. and Salovey, P. (1997), “What is emotional intelligence?”, in Salovey, P. and Sluyter, D.(Eds), Emotional Development and Emotional Intelligence: Educational Implications, BasicBooks, New York, NY, pp. 3-24.

Mayer, J., Salovey, P. and Caruso, D. (2003), “Measuring emotional intelligence with the MSCEITV2.0”, Emotion, Vol. 3, pp. 97-105.

Mesmer-Magnus, J. and Viswesvaran, C. (2006), “Social desirability: the role of over-claiming,self-esteem, and emotional intelligence”, paper presented at the Academy of ManagementConference, Atlanta, GA.

Meyer, B., Fletcher, T. and Parker, S. (2004), “Enhancing emotional intelligence in the health careenvironment: an exploratory study”, Health Care Manager, Vol. 23 No. 3, pp. 225-34.

Murray, J. and Jordan, P. (2006), “Improving emotional intelligence and performance throughemotions focused training interventions”, paper presented at the Academy of ManagementConference, Atlanta, GA.

Murray, J., Jordan, P. and Hall-Thompson, S. (2005), “Can emotional intelligence be increasedthrough training? An experimental study”, paper presented at the Academy ofManagement Conference, Honolulu, HI.

Murray, J. and Lawrence, S. (2006), “Using a self-report measure to determine whether emotionalintelligence can be trained”, paper presented at the Academy of Management Conference,Atlanta, GA.

O’Conner, R. and Little, I. (2003), “Revisiting the predictive validity of emotional intelligence:self-report versus ability-based measures”, Personality and Individual Differences, Vol. 35,pp. 1893-902.

Emotionalintelligence of

leaders

245

Page 22: Developing and measuring the emotional intelligence of leaders

Palmer, B., Gignac, G., Manocha, R. and Stough, C. (2005), “A psychometric evaluation of theMayer-Salovey-Caruso Emotional Intelligence Test Version 2.0”, Intelligence, Vol. 33,pp. 285-305.

Petrides, K. and Furnham, A. (2000), “Gender differences in measured and self-estimated traitemotional intelligence”, Sex Roles, Vol. 42 Nos 5/6, pp. 449-61.

Sadler-Smith, E. and Shefy, E. (2004), “The intuitive executive: understanding and applying ‘gutfeel’ in decision-making”, Academy of Management Executive, Vol. 18 No. 4, pp. 76-91.

Saklofske, D., Austin, E. and Minski, P. (2003), “Factor structure and validity of a trait emotionalintelligence measure”, Personality and Individual Differences, Vol. 34 No. 4, pp. 707-21.

Sala, F. (2002), Emotional Competence Inventory Technical Manual, HayGroup, Philadelphia, PA.

Salovey, P. and Mayer, J. (1990), “Emotional intelligence”, Imagination, Cognition, andPersonality, Vol. 9, pp. 185-211.

Salovey, P., Mayer, J. and Caruso, D. (2003), “Measuring emotional intelligence as a set of abilitieswith the Mayer-Salovey-Caruso Emotional Intelligence Test”, in Lopez, S. and Snyder, C.(Eds), Positive Psychological Assessment, American Psychological Association,Washington, DC, pp. 251-65.

Schutte, N., Malouff, J., Hall, L., Haggerty, D., Cooper, J., Golden, C. and Dornheim, L. (1998),“Development and validation of a measure of emotional intelligence”, Personality andIndividual Differences, Vol. 25, pp. 167-77.

Semadar, A., Robins, G. and Ferris, G. (2006), “Comparing the validity of multiple socialeffectiveness constructs in the prediction of managerial job performance”, Journal ofOrganizational Behavior, Vol. 27, pp. 443-61.

Seo, M., Barrett, L. and Bartunek, J. (2004), “The role of affective experience in work motivation”,Academy of Management Review, Vol. 29, pp. 423-39.

Sy, T., Tram, S. and O’Hara, A. (2006), “Relation of employee and manager emotional intelligenceto job satisfaction and performance”, Journal of Vocational Behavior, Vol. 68, pp. 461-73.

Sy, T. and Cote, S. (2004), “Emotional intelligence: a key ability to succeed in the matrixorganization”, Journal of Management Development, Vol. 23 No. 5, pp. 437-55.

Tett, R., Fox, K. and Wang, A. (2005), “Development and validation of a self-report measure ofemotional intelligence as a multidimensional trait domain”, paper presented at the AnnualConvention of the Southeastern Psychological Association, Atlanta, GA.

Von Glinow, M., Shapiro, D. and Brett, J. (2004), “Can we talk, and should we? Managingemotional conflict in multicultural teams”, Academy of Management Review, Vol. 29,pp. 578-92.

Wong, C. and Law, K. (2002), “The effects of leader and follower emotional intelligence onperformance and attitude”, The Leadership Quarterly, Vol. 13, pp. 243-74.

Zeidner, M., Matthews, G. and Roberts, R.D. (2004), “Emotional intelligence in the workplace: acritical review”, Applied Psychology: An International Review, Vol. 53 No. 3, pp. 371-99.

Appendix 1. EISDI itemsPerception and appraisal of emotions (PE)

(1) I can accurately identify a range of emotions that I feel from day to day.

(2) At work I can instantly tell when someone is frustrated with me.

(3) I can usually imagine what another person is feeling.

(4) I have no difficulty figuring out how much passion to demonstrate about an issue atwork.

JMD27,2

246

Page 23: Developing and measuring the emotional intelligence of leaders

(5) I can usually tell how someone is feeling even though his/her facial expression may

conflict with his/her body language.

(6) I have no difficulty identifying how a person really feels about an issue despite what

he/she may say.

Facilitating thinking with emotions (FE)

(1) I often prioritize my work tasks according to how strongly I feel about the importance of

each task.

(2) I often use my excitement about a work project to focus the efforts of others involved

with the project.

(3) I often use how I feel about a problem to define the attention I give to it.

(4) I listen to the feelings of other people in establishing priorities.

(5) I deliberately attempt to create a feeling conducive to effective problem solving when

meeting with clients or coworkers.

(6) In deciding to go forward with a decision, I always consider how other people may feel

about it.

Understanding emotion (UE)

(1) When a coworker of mine performs poorly on a project, I can usually recognize whether

he or she feels angry, embarrassed, guilty, or some other feeling (e.g. “wounded pride”).

(2) I can watch other people interact and recognize the feelings they hold toward each other.

(3) I am acutely aware of subtle cues at work that express how people feel (e.g. where they

sit, when they are silent, etc.).

(4) I can usually tell when a coworker’s emotional response to a situation is due to his/her

unique personality instead of his/her cultural background.

(5) I can usually detect subtle changes in the emotions of my coworkers.

(6) I can instantly recognize when a coworker’ frustrations with a project are escalating.

Regulation and management of emotion (RE)

(1) I look forward to a feeling of accomplishment whenever I start a new project.

(2) I am usually able to transmit a sense of enthusiasm about a work project to others.

(3) I notice when someone is very caring and compassionate toward others at work.

(4) I am capable of calming someone down who is angry and frustrated at work.

(5) When a coworker is feeling disappointed about his/her work performance, I make an

effort to offer encouraging words of support.

(6) Whenever painful events have occurred to people I know at work (i.e. death in family,

serious illness), I have expressed genuine concern and tried to help them feel better.

Emotionalintelligence of

leaders

247

Page 24: Developing and measuring the emotional intelligence of leaders

Appendix 2. Treatment group exercises and activities for EI developmentSelf-assessmentsParticipants completed five self-assessments that focused on emotional intelligence along withseveral other surveys that dealt with related phenomena (personality, coaching experience andskill, leadership style, etc.). The EI self-assessments included:

. A tool designed to provide scores on each of the major EI elements incorporated in theMayer and Salovey (1997) model of emotional intelligence. They completed this at thestart of the training process and at the end in order to identify sources of strength anddevelopmental needs.

. A 25-item survey designed to measure their readiness to engage in emotional intelligencedevelopment, including items tapping self-efficacy, receptivity to feedback, etc. (e.g. “I donot like to receive feedback from people” – reverse scored).

. A survey that gauged the extent to which the position they occupy requires emotionalintelligence in order to perform effectively.

. A questionnaire in which they identified critical incidents demonstrating effective andineffective examples of emotionally intelligent behavior.

. An inventory in which they indicated the major methods through which they learnedwhat they already knew, if anything, about emotional intelligence.

Self-development planParticipants created a written plan that they subsequently carried out with the aid of a coachover a ten-week period. It included the following nine components:

(1) identification of three emotional intelligence abilities they sought to enhance;

(2) an estimate of their current competence on these;

(3) a statement identifying their goals expressed in behavioral terms;

(4) a description of the action(s) they would take to meet their goals;

(5) identification of one or more sources of social support and feedback they would use;

(6) points of periodic review at which they would check on their goal progress;

(7) identification of the rewards they would apply to themselves contingent upon progressand upon completion of their plan along with rewards they would provide to others whoassisted them, where applicable;

(8) the amount of time they would devote to each action they planned to take and the totalamount of time for plan completion expressed in hours; and

(9) an outline of the major obstacles they anticipated they would encounter in carrying outtheir plan along with the methods they would use to surmount them.

CoachingEach individual was paired with a partner who served as a source of feedback and assistance.The coaching pairs met periodically throughout the course. They established initial expectations,met to clarify development goals and methods, reviewed progress and provided support to oneanother as they attempted to enhance their EI abilities.

On-site exercisesThey participated in several role-play exercises that permitted them to practice, observe, andlearn about several EI abilities (e.g. a top management team-based problem solving discussionand subsequent presentation of a turnaround plan for a company).

JMD27,2

248

Page 25: Developing and measuring the emotional intelligence of leaders

InterviewThey arranged and presented an analysis of an interview they conducted with an individualconcerning emotional intelligence. Participants either interviewed a manager concerning theneed for, application of, impact and development of four dimensions of emotional intelligence inhis/her job, or, an individual who was a non-native resident of the country concerning theprocesses of identifying, expressing, understanding using and regulating emotions across twodifferent cultural contexts.

CritiqueThey each critiqued two written EI development plans produced by other participants in termsof the components identified above. This provided them with an opportunity to demonstrate andpractice emotional intelligence in the process of conveying performance feedback to others. Theyreceived two critiques of the plan they had created from other participants in addition to acritique of it written by the instructor.

ReadingParticipants read and were tested on their comprehension of information contained in Carusoand Salovey (2004). This text outlines and provides examples of the Mayer and Salovey model.

Interim progress reportsEach coaching pair wrote three brief progress reports in which they:

. reviewed results achieved to date;

. identified and/or answered any questions their coaching partner had concerning EI;

. discussed and/or analyzed elements of the training such as the self-assessment data; and

. described the action they and their partner intended to take in order to achieve their goalsduring the forthcoming period.

Observation of film clipsParticipants watched several film clips demonstrating various EI abilities, as time permitted.These included:

. the ability to discriminate between honest/dishonest emotions (Rounders);

. the ability to understand complex emotions (Joy Luck Club);

. the impact of emotions on problem-solving and judgment (Mystic River and Ace Ventura:Pet Detective);

. the ability to monitor and judge emotions (The Cooler); and

. the ability to manage others’ emotions (Miracle).

One-on-one meeting with instructorEach participant met with the instructor mid-way through the course to review the feedback s/hehad received on each activity/exercised performed, assess and resolve any problems s/he voiced,offer encouragement and assistance concerning his/her progress and clarify what s/he needed todo on remaining tasks. In addition, they received ongoing feedback from three sources:

(1) their coach;

(2) the source of support they selected in their plan, which was usually their boss; and

(3) other participants in the training process.

Emotionalintelligence of

leaders

249

Page 26: Developing and measuring the emotional intelligence of leaders

Journal and summary of learning achievedEach participant wrote a summary statement in which they identified the learning they hadachieved in comparison with the initial objectives they established at the start of the course in theself-development plan they had created. This was in addition to keeping a weekly journal inwhich they were asked to think creatively about what they learned as a byproduct of the processthey undertook and how they could apply it. Note that most participants used their workplace asthe “practice field” for carrying out the activities in which they engaged to enhance their EI.

Corresponding authorKevin S. Groves can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] visit our web site for further details: www.emeraldinsight.com/reprints

JMD27,2

250