Zeidner et al. 2004.pdf

30
See discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/227794541 Emotional Intelligence in the Workplace: A Critical Review ARTICLE in APPLIED PSYCHOLOGY · JUNE 2004 Impact Factor: 1.52 · DOI: 10.1111/j.1464-0597.2004.00176.x CITATIONS 147 READS 7,933 3 AUTHORS, INCLUDING: Gerald Matthews University of Central Florida 275 PUBLICATIONS 7,153 CITATIONS SEE PROFILE Richard D Roberts Professional Examination Se… 140 PUBLICATIONS 3,998 CITATIONS SEE PROFILE Available from: Gerald Matthews Retrieved on: 05 November 2015

Transcript of Zeidner et al. 2004.pdf

Page 1: Zeidner et al. 2004.pdf

Seediscussions,stats,andauthorprofilesforthispublicationat:http://www.researchgate.net/publication/227794541

EmotionalIntelligenceintheWorkplace:ACriticalReview

ARTICLEinAPPLIEDPSYCHOLOGY·JUNE2004

ImpactFactor:1.52·DOI:10.1111/j.1464-0597.2004.00176.x

CITATIONS

147

READS

7,933

3AUTHORS,INCLUDING:

GeraldMatthews

UniversityofCentralFlorida

275PUBLICATIONS7,153CITATIONS

SEEPROFILE

RichardDRoberts

ProfessionalExaminationSe…

140PUBLICATIONS3,998CITATIONS

SEEPROFILE

Availablefrom:GeraldMatthews

Retrievedon:05November2015

Page 2: Zeidner et al. 2004.pdf

APPLIED PSYCHOLOGY: AN INTERNATIONAL REVIEW, 2004,

53

(3), 371–399

© International Association for Applied Psychology, 2004. Published by Blackwell Publishing,9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.

-Blackwell Publishing LtdOxford, UKAPPSApplied Psychology: an International Review0269-994X© International Association for Applied Psychology, 2004July 20045331000Original ArticlesEI IN THE WORKPLACEZEIDNER ET AL.

Emotional Intelligence in the Workplace: A Critical Review

Moshe Zeidner*

University of Haifa, Israel

Gerald Matthews

University of Cincinnati, USA

Richard D. Roberts

University of Sydney, Australia

Cet article est une revue critique des théories et résultats empiriques favorablesà l’intelligence émotionelle (I.E.) et à son prétendu rôle dans l’environnementprofessionnel. On s’intéresse au statut supposé de l’I.E. dans la performanceau travail, la satisfaction et l’évaluation de la carrière et des compétences(surtout dans la domaine de la sélection et de l’orientation). Globalement,cette revue de questions prouve que les recherches récentes ont fait de grandspas dans la comprehénsion de l’utilité de l’I.E. au travail. Les preuves strict-ement scientifiques sont cependant insuffisantes, la littérature accordant uneconfiance excessive aux avis d’experts, aux anecdotes, aux études de cas et auxenquêtes privées non publiées. On propose, à la fin de l’article, quelquesdirectives pratiques pour favoriser le développement et l’utilisation de mesuresde l’I.E. dans les situations professionnelles.

This paper critically reviews conceptualisations and empirical evidence insupport of emotional intelligence (EI) and its claimed role in the occupationalenvironment. Consideration is given to the purported status of EI in occupa-tional and career assessment (with particular emphasis on personnel selec-tion and placement), job performance, and satisfaction. Overall, this reviewdemonstrates that recent research has made important strides towardsunderstanding the usefulness of EI in the workplace. However, the ratio ofhyperbole to hard evidence is high, with over-reliance in the literature onexpert opinion, anecdote, case studies, and unpublished proprietary surveys. Thereview concludes by providing a number of practical guidelines for the devel-opment and implementation of EI measures within occupational settings.

* Address for correspondence: Moshe Zeidner, Center for Interdisciplinary Research of Emo-tions, University of Haifa, Mt Carmel, 31905, Israel. Email: [email protected]

Page 3: Zeidner et al. 2004.pdf

372

ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

INTRODUCTION

Emotional intelligence (EI) is a relatively new and growing area of behavioralresearch, having caught the imagination of the general public, the commercialworld, and the scientific community. The concept resonates with a currentzeitgeist emphasising the importance of self-awareness and understanding,redressing a perceived imbalance between intellect and emotion in thelife of the collective Western mind. Emotional intelligence also connectswith several cutting-edge areas of psychological science, including theneuroscience of emotion, self-regulation theory, studies of metacognition,and the search for human cognitive abilities beyond “traditional” academicintelligence.

Although Thorndike (1921), Guilford (1956), and later, Gardner’s (1983)research into social intelligence hints at the importance of emotions tointellectual functioning, the term EI was not brought into mainstreampsychology until the 1990s (Mayer, DiPaolo, & Salovey, 1990; Salovey &Mayer, 1990). Currently, Mayer, Salovey, and colleagues argue that EIincorporates a set of conceptually related psychological processes invol-ving the processing of affective information (see Mayer & Geher, 1996; Mayer& Salovey, 1997; Salovey & Mayer, 1990, 1994). These processes includethe appraisal and expression of emotions, assimilation of emotions inthoughts, understanding emotion, and the regulation and management ofemotions.

For a concept that up until recently had received short shrift, the impressionthat the study of EI is a pivotal area of contemporary psychology appearsdifficult to dispute. Thus, EI has been touted as a panacea for modernbusiness and the essential but often neglected ingredient of nursing, legal,medical, and engineering practices (see Zeidner, Matthews, & Roberts,2001). In some commentators’ eyes, EI even provides the medium by whicheducational reform can and finally will reach its full potential, across prim-ary, secondary, and tertiary levels of schooling (see Zeidner, Roberts, &Matthews, 2002, for a critical review).

The current paper provides a critical analysis of the claimed role ofemotional intelligence in the occupational environment. Following a briefoverview of the conceptualisation and measurement of EI, considerationis given to an emerging literature that promotes the assessment, training,and the individual’s utilisation of emotional intelligence in the workplace.Throughout, an attempt is made to bring to the reader’s attention the scant,and sometimes highly controversial, empirical evidence used to supportthe importance of EI in the workplace. This approach naturally indicatesavenues that future research might profitably explore. The paper concludesby presenting some practical guidelines for the development of EI measuresfor occupational selection purposes.

Page 4: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE

373

© International Association for Applied Psychology, 2004.

EMOTIONAL INTELLIGENCE: CONCEPTS AND MEASUREMENT

Popular interest in EI has, at times, tended to obscure definitional clarity(Matthews, Zeidner, & Roberts, 2002). The emerging literature on EI containsdisparate terminology, including not only

emotional intelligence

(Goleman,1995; Salovey & Mayer, 1990), but also

emotional literacy

(Cooper &Sawaf, 1997),

emotional quotient

(Cooper, 1997), and

personal intelligences

(Gardner, 1983). To further complicate the situation, the sub-components ofEI are variously referred to as “branches” (Mayer, Caruso, & Salovey,2000), “factors” (Bar-On, 1997), or “competencies” (Boyatzis, 1982).

Definitions of EI

No matter what its hue, the aforementioned proponents all lay claim to thefact that their concept constitutes a generalised, far-reaching intelligencecovering an array of emotional functions. Unfortunately, thus used, theterm too often appears all encompassing and protean, such that EI is leftbereft of conceptual meaning. For example, the populist, though widelyinfluential account offered by Goleman (1995) appears to define EI byexclusion: as

any

desirable feature of personal character not representedby cognitive intelligence. More recently, Goleman (1998, 2001, see alsoBoyatzis, Goleman, & Rhee, 2000) suggests that two domain facets definethe competencies associated with EI: (a)

ability

—awareness versus manage-ment of emotion, and (b)

target

—whether competence relates to self versusothers. The Cartesian product of these two facets (i.e. ability by target) yieldsthe following four components: (a) awareness of emotions in self; (b) aware-ness of emotions in others; (c) management of emotions in self; and (d)management of emotions in others. However, although this analysis sug-gests some fields of inquiry, it does not identify a unifying common elementto the different components. Furthermore, this conceptualisation does nottell us how to distinguish EI from other, distinct abilities and personalitytraits that may influence recognition and regulation of emotions (e.g. traitanxiety, coping dispositions).

Perhaps the most widely accepted scientific definition of EI is “the abilityto monitor one’s own and others’ emotions, to discriminate among them,and to use the information to guide one’s thinking and actions” (Salovey& Mayer, 1990, p. 189). This definition identifies emotional informationprocessing as a necessary precursor of emotional regulation, and as we haveargued elsewhere, probably constitutes the most workable contemporarydefinition of EI (see Matthews et al., 2002). By contrast, another leadingresearcher (Bar-On, 1997) characterises EI as “an array of non-cognitivecapabilities, competencies, and skills that influence one’s ability to succeed

Page 5: Zeidner et al. 2004.pdf

374

ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

in coping with environmental demands and pressures” (p. 16). This broaderdefinition makes no direct reference to the acquisition, retrieval, and instantia-tion (through appropriate behaviors) of emotional information. It appears toexclude cognitive skills that might contribute to emotion management,although, confusingly, Bar-On also lists apparently cognitive abilities suchas problem solving and reality testing as components of EI. Conversely,Bar-On’s definition places more emphasis on adaptation to environmentaldemands.

Models of EI

Mayer and colleagues distinguish between (1) “mental ability models”,focusing on aptitude for processing affective information, and (2) “mixedmodels” that conceptualise EI as a diverse construct, including aspects ofpersonality as well as the ability to perceive, assimilate, understand, andmanage emotions. These “mixed models” include motivational factors andaffective dispositions (e.g. self-concept, assertiveness, empathy; see Bar-On,1997; Goleman, 1995). These conceptual disagreements are mirrored by amajor disjunction in measurement paradigm. Those who conceptualise EIas a fairly well-defined set of emotion-processing skills (e.g. Mayer, Caruso,& Salovey, 1999, 2000) aim to assess EI through objective, performancetests. Conversely, those who view EI as encompassing multiple aspects ofpersonal functioning (e.g. Bar-On, 1997; Boyatzis et al., 2000; Goleman,1995) aim to measure EI through self-report protocols. By and large, thesetests are designed to assess beliefs and perceptions about an individual’scompetencies in specific domains (Salovey, Woolery, & Mayer, 2001).

Table 1 summarises some of the cardinal differences among mixed andability models of EI along a number of dimensions, such as conceptualcontext, focus, dimensionality, measurement procedures, and their psycho-metric properties. The manifest differences, contained in this table, shouldalert the reader to a particularly problematic feature associated with currenttheories of EI—whatever is being measured within “mixed models”, it isunlikely the same type of EI as that assessed by “mental ability models”.We take up this notion still further in the passages that follow.

A number of problems and serious omissions currently plague the researchon EI conducted under the “mixed-model” banner, which employs self-reportmethodologies (see Davies, Stankov, & Roberts, 1998; Matthews et al., 2002;Roberts, Zeidner, & Matthews, 2001). According to Barrett, Miguel, Tan, andHurd (2001), self-report measures of ability suffer from low reliability, low orno criterion-related validity, limited construct validity, and are easily faked. Itis also questionable whether items asking students to self-appraise intellectualability (e.g. “I am an extremely intelligent student”) would make for a validmeasure of any intelligence. Moreover, tests of EI that assess non-cognitive

Page 6: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE

375

© International Association for Applied Psychology, 2004.

TABLE 1Comparison of Mixed vs. Ability Models of Emotional Intelligence

Dimension Model of Emotional Intelligence

Mixed Models Ability Models

Conception of EI EI viewed as melange of competencies and general dispositions for adaptive

personal functioning

and coping with environmental demands. The construct encompasses multiple aspects of emotional and personal knowledge and personal functioning that are rather loosely related to emotion, including: motivation, personality traits, temperament, character, and social skills

EI is viewed as a well-defined and conceptually related set of cognitive

abilities

for the processing of emotional information and regulating emotion adaptively

Psychological focus Affective CognitiveTheoretical model Personality/Psychological adjustment Intelligence/PerformanceTypical facets Self-awareness, self-motivation,

self-regulation, empathy, social skills, assertiveness, stress tolerance, impulse control, coping with stress, reality testing, social problem solving, etc.

Emotion identification, understanding of emotions, assimilation of emotion in thought and use of emotions to enhance thought, emotion regulation

Number of competencies

Anywhere from 4 to 2 dozen abilities. These can be grouped into 4 core areas: self-awareness, self-regulation/management, social awareness, relationship management and social skills (Cherniss & Goleman, 2001)

4 major branches: identification, understanding, usage, and self-regulation (Salovey et al., 2000)

Morphological structure

Nonhierarchical—“oligarchic” organisation

Hierarchical model—from basic psychological processes to higher more psychologically integrated processes

Key proponents Goleman (1995), Bar-On (1997) Mayer et al. (2000a)Measurement approaches

Quasi-personality

(self-report, Likert-type scales)

Competency

(performance type items such as identification of emotions in pictures, identifying progressions and blends of emotions, solving problems, etc.)

Examples of scales Bar-On’s EQ-i, Schutte’s EI scale, Boyatzis and Goleman’s Emotional Competence Inventory, Cooper’s EQ Map

Mayer, Caruso, & Salovey’s MEIS, MSCEIT

Scoring of scales No veridical scoring criteria. Scores obtained by linear sum of Likert-type scale response categories scored in direction of high EI

Consensus, Expert, and Target scoring protocols, with presumable veridical or “objective” scoring criteria

Page 7: Zeidner et al. 2004.pdf

376

ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

Factor structure Little empirical data. General factor found for individual published scales, but little evidence to support claims of multiple factors (cf. Petrides & Furnham, 2000)

Inconsistent with 4-branch model. Exploratory factor analytic data consistent with 3 factor models of perception, understanding, regulation (Mayer, Caruso, & Salovey, 2000; Roberts et al., 2001)

Reliability of scales Satisfactory (Bar-On, 1997; Dawda & Hart, 2000)

Low to Moderate (Roberts et al., 2001); inconsistency among scoring procedures and low subtest reliabilities

Susceptibility of items to response sets

Inconsistent data; some evidence for extreme item endorsement (Dawda & Hart, 2000)

Not relevant

Convergent validity (vis-a-vis ability)

Very low—negligible correlations with IQ (Bar-On, 2000; Derksen et al., 2002)

Moderate correlations of about .30 with ability (Mayer et al., 2000; Roberts et al., 2001)

Divergent validity (vis-a-vis personality)

Low discriminant validity vis-a-vis personality measures, particularly N

Good discriminant validity, with low correlations with “Big 5” personality facets (Roberts et al., 2001)

Predictive validity Good, but may reflect confounding with personality (Janovics & Christiansen, 2001)

Good, but may reflect confounding with ability (Janovics & Christiansen, 2001)

Dimension Model of Emotional Intelligence

Mixed Models Ability Models

TABLE 1Continued

traits (e.g. assertiveness, optimism, impulse control) seem to be tappingdimensions of individual differences that relate to established personality con-structs rather than to “intelligence” (Matthews et al., 2002).

In view of the foregoing problems associated with the use of self-reportmeasures, Mayer, Salovey, and colleagues have advocated the development ofobjective, performance-based ability indicators of EI (see e.g. Mayer, Carusoet al., 1999, 2000; Mayer & Salovey, 1997; Mayer, Salovey et al., 2000a, 2000b).Consequently, task-based measures engage participants in exercises designedto assess abilities supporting emotionally intelligent behavior. The ability-based mode of assessment, and its underlying four-branch conceptual modelof EI, has gained currency largely because it appears to be performance-

Page 8: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE

377

© International Association for Applied Psychology, 2004.

oriented and empirically based. Unfortunately, there is considerable difficultyin determining objectively correct responses to stimuli involving emotionalcontent, and in applying truly veridical criteria in scoring tasks of emotionalability (Roberts et al., 2001). Proponents of EI as a type of cognitive abilityhave promoted alternative scoring procedures in order to discriminateright from wrong answers on performance-based measures of EI (consensual,expert, target; see Mayer, Caruso et al., 2000). While still in their infancyand requiring stringent empirical studies to ascertain certain shortcomingsand alternatives the rather novel approach adopted to measurement in abilitymodels, along with positive results to be discussed shortly, suggests that theymay be the focus of research on EI for some time.

Emotional Competencies

Another approach, sharing more in common with “mixed models” but movingbeyond a rigid conceptualisation of EI, advocates differentiation betweenemotional intelligence (a dispositional aptitude) and emotional

competencies

(learned capabilities) (Boyatzis, 1982; Goleman, 2001). Based on a host of casestudies, anecdotal accounts, and evaluation studies, Goleman (1998) con-cludes that the major qualities differentiating successful from unsuccessfulexecutives are the competencies underlying (or presumably nested within)EI. Failing executives, apparently, have poorer emotional control, despitestrengths in cognitive abilities and technical expertise.

Under this formulation, EI encompasses such characteristics as motives,traits, and aspects of one’s self-image. In short, EI designates the potentialto become skilled at learning certain emotional responses. By contrast,emotional

competencies

are learned capabilities, based on EI, that result inoutstanding performance at work (Goleman, 2001). Akin to the distinc-tion between

fluid

and

crystallised

ability (cf. Matthews et al., 2002), EI (asa fluid ability) does not guarantee that individuals will actually manifestcompetent behaviors at the workplace. That is, there is no guarantee thatthe individual has been exposed to essential environmental experiences orlearning situations and practices necessary to acquire specific emotional com-petencies or skills (e.g. assertiveness, service orientation, initiative). WhereasEI may determine a person’s potential for learning practical job-relatedemotional and social skills, the level of emotional competencies (as a crystal-lised ability) manifested by that person shows how much of that potentialshe or he has actually realised. It is emotional competence then that aids thelearning of job-related skills and which translates EI into on-the-job capab-ilities. For example, in order to be able to actually empathise with another’splight, one needs to have learned the specific empathic skills that translateinto caring and compassionate pastoral counseling, bedside-nursing, or effectivepsychotherapy (cf. Cherniss & Goleman, 2001).

Page 9: Zeidner et al. 2004.pdf

378

ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

Within this general framework, a large array of competencies have beenclaimed to be critical for success in occupational settings (see e.g. Boyatziset al., 2000; Cooper & Sawaf, 1997; Weisinger, 1998). For example, Goleman(1998) lists 25 different competencies necessary for effective performance invarious occupational contexts. Thus, confidentiality is touted as importantfor loan officers and priests, while trust and empathy appear vital for psycho-therapists, social workers, and marriage counselors. Among the specificcompetencies claimed to be of critical importance in a variety of occupationalsettings are the following:

Emotional self-awareness.

This competence includes identification of emotionand understanding how emotions are related to one’s goal, thoughts, behaviors,and accomplishments (Goleman, 1998; Weisinger, 1998).

Regulation of emotions in the self.

This competence involves intentionallyeliciting and sustaining pleasant and unpleasant emotions when consideredappropriate, effectively channeling negative affect, and restraining negativeemotional outbursts and impulses (Boyatzis, 1982; Goleman, 1998).

Social awareness of emotions and empathy

, which includes awareness ofothers’ feelings, needs, and concerns, understanding and sympathising withothers’ emotions, and responding to others’ unspoken feelings (Goleman,1998; Huy, 1999; cf. Salovey & Mayer, 1990; Williams & Sternberg, 1988).

1

Regulating emotions in others

. This competence incorporates influencingothers, effectively communicating with others, and managing conflicts(Weisinger, 1998).

Motivational tendencies

, which include such components as internalstrivings, attributions, and need for achievement (Bar-On, 2000; Boyaztiset al., 2000; Cooper & Sawaf, 1997; Goleman, 1998; Weisinger, 1998).

Character

, which includes trust and integrity (Cooper & Sawaf, 1997;Goleman, 1998; Weisinger, 1998).

The preceding framework is not, of course, without its critics. Thus, tosome researchers, “competencies” is a confusing and ambiguous concept(see Barrett et al., 2001). Indeed, how specific competencies are relatedto the more overarching concept of EI is uncertain. Furthermore, it ispresently unclear to what extent a number of specific competencies maybe nested within each of these facets. Thus, certain competencies such asimpulse control, achievement motivation, and adaptability are subsumedunder regulation of emotions in self, whereas conflict resolution, teamwork,visionary leadership, and communication skills are nested within manage-ment of emotions in others (cf. Goleman, 2001). Whether placing all suchconcepts under the EI banner confuses, rather than clarifies, the role of

1

Mayer, Caruso and Salovey (2000) see this construct as a “shadow variable”—one thatmimics EI in several respects, but that seems conceptually and ontologically distinct.

Page 10: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE

379

© International Association for Applied Psychology, 2004.

emotional competencies in the workplace would seem a contentiouspoint. Because the field of EI remains new, many of the aforementionedconcepts—which have been studied in organisational psychology for sometime (often with mixed results)—are in fact better understood than this fledglingconcept. We are cautious, however, of being definitive in making a finaljudgment of this approach. On one hand, the process of reconceptualisingeach of the preceding concepts as forms of EI (or competencies) maylead to obfuscation. On the other hand, dealing with distinct but possiblyinterrelated competencies may be more tractable for research and practicalpurposes.

EI AND OCCUPATIONAL ASSESSMENT, SELECTION, AND PLACEMENT

Recently, the use of EI measures for career selection and placement pur-poses has begun to gather momentum in many organisations in the Westernworld. Thus, more and more companies are realising that EI skills maybe a vital component of any organisation’s management philosophy (andsubsequent success). A survey of benchmark practices among major cor-porations found that four out of five companies are now trying to promoteEI in their organisations. The concept of EI is thought to be useful whenevaluating ongoing functioning and the well-being of employees at criticalstages of their careers (i.e. selection, training, placement, and promotion).As one group of writers has argued: “If the driving force of intelligence intwentieth century business has been IQ, then . . . in the dawning twenty-firstcentury it will be EQ” (Cooper & Sawaf, 1997, p. xxvii).

Overview

Gowing (2001) traces the roots of EI in organisational settings to classicmanagement theory and practice. Indeed, many of the strategies used in earlyassessment centers evaluated non-cognitive abilities akin to EI (e.g. socialawareness, understanding others, communication). These abilities were foundto be predictive of successful performance in managerial positions in manycorporations. Furthermore, over three decades of psychological assessmentresearch has vindicated the importance of taking social and emotionalcompetencies into consideration when attempting to predict occupationaleffectiveness (e.g. Boyatzis, 1982; Campbell, Dunnette, Lawler, & Weich,1970; Howard & Bray, 1988; Kotter, 1982). In a now classic study, Kotter(1982) identified a number of personal characteristics discriminating morefrom less successful general managers, including such social-emotionalcompetencies as optimism, communication and relationship skills, and needfor achievement. Furthermore, research by Boyatzis (1982) has identified a

Page 11: Zeidner et al. 2004.pdf

380

ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

number of social competencies (i.e. socialised power, self-esteem, positive-ness) that appear predictive of future managerial success.

Based on their survey of the intervention literature in the domain ofmanagement, Cherniss and Goleman (2001) conclude that interventionstargeted at EI-based competencies are effective and tend to enhance suchdesired outcomes as self-awareness and rapport. Thus, they conclude: “Takentogether, all these interventions demonstrate that it is possible for adults todevelop EI competencies” (p. 214). These authors go on to offer methodsfor developing specific EI domains (e.g. developing social skills via modeling).Furthermore, in order to maximise the effectiveness of these programs,Cherniss and Goleman (2001) suggest a number of useful guidelines(e.g. creating an encouraging and supportive environment for intervention;using models of desired skills; inoculating against setback and providingfollow up support). These guidelines appear to have had some success.

Several unsubstantiated claims have appeared in the popular literatureand the media about the significant role of EI in the workplace. Thus,EI has been claimed to validly predict a variety of successful behaviors atwork, at a level exceeding that of intelligence (see Cooper & Sawaf, 1997;Goleman, 1998; Hay Group, 2000; Weisinger, 1998). In the

Time

articlewhich helped popularise EI, Gibbs (1995) wrote, “In the corporate world . . .IQ gets you hired but EQ gets you promoted” (p. 59). Watkin (2000) suggests,without empirical support: “Use of EI for recruitment decisions leads to90-percentile success rates.” He goes on to claim that “what distinguishes topperformers in every field, in every industry sector, is not high IQ or technicalexpertise, it is EI” (p. 91). Similarly, Goleman (1995) has claimed, fromresearch on over 500 organisations by the Hay Group, that EI (rather thanIQ) accounts for over 85 per cent of outstanding performance in top leaders.Of note, however, Goleman is unable to cite empirical data supporting anycausal link between EI and any of its supposed, positive effects.

The Predictive Validity of EI

Much of the current interest focusing on EI in organisational settings stemsfrom a desire to explain differential attainment of occupational success,which cannot adequately be accounted for by IQ alone. However, assess-ment of EI is only cost-effective to the extent that it provides informationadditional to that provided by measurement of established ability and per-sonality constructs. Thus, EI measures must demonstrate not just criterionand predictive validity, but also discriminant or incremental validity, withrespect to existing tests. Establishing predictive validity is made more diffi-cult by the lack of convergence between different types of EI test; Bar-On’sself-report scale, the EQ-i is only modestly correlated (at .46) with theMayer-Salovey MEIS ability test, for example (Bar-On, 2000). Self-report

Page 12: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE

381

© International Association for Applied Psychology, 2004.

and quasi-objective tests are also differentiated by their correlations withother constructs. The MEIS is modestly correlated with general intelligence,for example, but self-reports are typically independent of intelligence (seeMatthews et al., 2002, for a review of data). Self-report scales, but notability scales, are highly correlated with existing personality questionnaires.Studies of the EQ-i (e.g. Dawda & Hart, 2000; Newsome, Day, & Catano,2000) have found that most of the variance in this instrument can be attrib-uted to the well-known Five Factor Model of personality. EQ appears to belargely low neuroticism, with smaller contributions from extraversion, agree-ableness, and conscientiousness. Hence, there is a problem with divergentvalidity; it has not been established that questionnaire measures of EI addmuch to orthodox personality assessments.

Overall, conventional intelligence tests do a very reasonable job ofpredicting occupational criteria (especially when compared to personalitymeasures, as we shall demonstrate shortly) (see Hunter & Schmidt, 1996).General ability predicts anywhere from about 10 per cent to 30 per centof the criterion variance in job performance, leaving about 90 per cent to70 per cent of the variance in success unaccounted for (see e.g. Jensen, 1980,1998). A review of the literature by Hunter and Hunter (1984) suggests thatcognitive abilities have a mean validity for training success of about .55for all known job families. In addition, studies surveyed by Hunter andcolleagues show that ability tests are valid across all jobs in predicting jobproficiency. The validity coefficients vary by both outcome criteria (higherfor job training and lower for job performance) and job complexity (higherfor greater job complexity).

According to Sjoberg (2001), one reason for the interest in the non-cognitive factors may simply be that it has proved to be very difficult toimprove, in the cognitive domain, on traditional measures of generalintelligence. As Schmidt (1994) points out:

After over 50 years of research . . . it is now evident that refinements in themeasurement of abilities and aptitudes are unlikely to contribute nontrivialincrements to validity beyond that which is produced by good measures of generalability. The areas of personality, biographical data, physical abilities, and perhapsinterests are considerably more promising in that respect. (pp. 348–349)

2

In general, personality measures are considerably less predictive ofjob performance than are ability measures. Meta-analyses of relationships

2

Whereas one may object to the use of self-report measures to assess EI (which is pur-portedly an ability and would thus require more objective performance-based measures), thisassertion does not hold with respect to personality, where self-report measures may be useful,provided that respondents are motivated to respond truthfully.

Page 13: Zeidner et al. 2004.pdf

382

ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

between the “Big Five” and job performance suggest that, even whencorrections are made for statistical artifact, mean validity coefficients donot exceed .2–.3 (Barrick & Mount, 1991, 1993; Tett, Jackson, Rothstein &Reddon, 1999). However, higher correlations may be found when moderatorfactors are taken into account. For example, extraversion is modestlypredictive of success for people in management and sales, but not for thosein other professions (Barrick & Mount, 1991). In general, confirmatorystudies, that are guided by some a priori hypothesis, obtain higher validitycoefficients than purely exploratory studies (Tett et al., 1999). Criteria otherthan objective performance may be more strongly linked to personality.

These existing personality studies place some constraints on the expectedvalidity of questionnaire scales for EI. Generally, it seems unlikely that scalessuch as the EQ-i will explain large percentages of variance in performancecriteria, although scales that are less strongly correlated with the Big Fivemight potentially do so. Existing research also shows that personality traitslinked to emotion may also have both positive and negative effects dependingon context (Matthews, 1997). Neuroticism appears to relate to low EI, in thathigh N persons are moody, vulnerable to stress, and tend to cope ineffectively.However, across the board, high N is not a barrier to occupational success:Barrick and Mount (1991) found that the corrected correlation between N andjob proficiency was a paltry

.07. High N does seem to relate to performanceimpairment in highly stressful occupations such as police work, but, con-versely, high N relates to greater work effort and sales volume in insurancesalespersons (Mughal, Walsh, & Wilding, 1996). Neuroticism may sometimesact as a spur to occupational achievement. Agreeableness (A), another correlateof the EQ-i, also has a near-zero overall correlation with job performance(Barrick & Mount, 1991). However, it seems that high A may be beneficial inteamwork situations (Hough, 1992), but low A is related to superior per-formance when managers operate under high levels of individual autonomy(Barrick & Mount, 1993). Qualities of agreeableness such as empathy, altruism,and interpersonal sensitivity are central to conceptions of EI, but these qualitiesmay mitigate against effective performance in jobs requiring ruthlessness,toughness, and individual initiative. It follows that research on EI should beacutely sensitive to possible moderator factors, and, unlike conventional ability,emotional intelligence may have both positive and negative associations withperformance, depending on contextual factors.

Over the past few years, a number of studies have attempted to determinethe concurrent validity of EI in predicting job performance, either in simulatedsettings or on the job. We now examine this empirical literature.

Empirical Studies: Reported Positive Results.

In one of the first studiesof its kind, Janovics and Christiansen (2001), using an incidental sample of176 undergraduates (70% female), found that EI (as assessed by the Mayer-

Page 14: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE

383

© International Association for Applied Psychology, 2004.

Salovey-Caruso Emotional Intelligence Test—MSCEIT) was modestlycorrelated with job performance (

r

=

.22)—as assessed by supervisors’ ratingsof employees on items evaluating professional work duties. Interestingly,job performance correlated significantly with only two of the four branchesof this test: Perception (

r

=

.14) and Understanding (

r

=

.30). This result iscurious since these higher-order factors are the least cognitive of the four-branch model of EI. Nevertheless, when added to a regression equationusing cognitive ability and the Big Five factor of Conscientiousness, as cov-ariates, a general EI score from the MSCEIT added 3 per cent to the incre-mental variance of the job performance criterion. Janovics and Christiansen(2001) conclude (we might argue, contentiously): “While EI measurement isunlikely to be as useful as popular authors may suggest, empirical evidenceusing the most advanced available measure suggests that EI offers someadditional contribution beyond measures of existing constructs” (p. 6).

Dulewicz and Higgs (2000) reanalysed data from a seven-year study ofthe career progress of 58 managers in the UK and Ireland assessing threedomains of self-reported ability: EQ, IQ, and managerial competency. Emo-tional competencies were derived from a job competency inventory (e.g.perceptive listening, integrity, stress tolerance, motivating others). EI wasfound to contribute to the prediction of the job advancement criterionabove and beyond managerial EI and self-reported intellectual perform-ance, adding about 36 per cent incremental variance to the prediction oflevel of advancement over a seven-year period. Taken together self-reportedcognitive and emotional intelligence accounted for 52 per cent of the cri-terion variance. Unfortunately, this study failed to assess the full spectrumof EQ and did not cover classic facets identified with EI such as emotionawareness and emotion regulation. Moreover, the fact that intelligence levelwas self-reported renders highly problematic any claims that EQ sharedhigher correlation with the criterion than IQ.

Bachman, Stein, Campbell, and Sitarenios (2000) hypothesised that emo-tional competencies enable account officers to achieve greater success incollections. Based on a small sample of 36 account officers, a “best prac-tices” group was found to possess a level of EI significantly higher thanthat of the North American population at large. These individuals alsoperformed better than a less successful group—particularly in the area ofproblem solving skills. The “best practices” group scored higher on the EQ-iscales of Optimism and Happiness. In a second study, based on 34 accountofficers, the high cash collectors group performed better than did the lowcash collectors group on all EQ-i sub-scales, with the exception of Empathyand Impulse Control. However, no effort was made to control for IQ orpersonality factors that may, in part, account for the observed differences.

Indeed, a number of studies based on Bar-On’s EQ-i purportedly supportthe validity of EI in the workplace. In fact, the publishers of the Bar-On test

Page 15: Zeidner et al. 2004.pdf

384

ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

assert that it is a better predictor of job success than IQ, referring to a few(as yet unpublished) studies in support of this claim. For example, Bar-On(1997) cites a study conducted on a sample of 81 chronically unemployedindividuals. These individuals had unusually low EQ-i scores, with the lowestscores on Assertiveness, Reality Testing, and Happiness. Similarly, Bar-On(1997) found that individuals from the Young President’s Organization(i.e. whose membership is dependent on individuals reaching top leadershippositions in expanding companies) obtained scores on the EQ-i (on virtu-ally all sub-scales) exceeding the average by significant amounts. Accordingto Bar-On, this group’s success was dependent on an ability to be veryindependent and to assert their individuality, while being able to withstandvarious stressors occurring within the job.

The direction of causality in each of these instances raises some concerns.In particular, low EI scores among the unemployed are likely to be a con-sequence (rather than a cause) of being chronically unemployed. Similarly,those performing well in their job are likely to report high levels of emo-tional stability.

This argument notwithstanding, Bar-On (2000) reports that in a surveyof nearly 100,000 employees in 36 countries, social responsibility surfacedas one of the most important factors determining effectiveness at work.However, according to Barrett et al. (2001), the latter study is little else buta typical name-catching exercise, whereby the authors claim that socialresponsibility is important for success and because their test supposedlymeasures social responsibility, it is valid for predicting success. Bar-On,however, does not cite any predictive or concurrent studies in this chapterto support his claims. In the EQ-i technical manual (1997), Bar-On assertsthat the data “indicate a strong connection between EQ-i scores and job‘performance’, based on a self-rating scale tapping a worker’s sense of com-petence” (p. 140). This assertion is based on a study of 324 workers fromthe US and Canada, who performed the EQ-i and a (self-reported) Sense ofCompetence Questionnaire. The correlation between the tests while high(

r

=

.51), needs to be qualified by the fact that both measures are based onself-reports, presumably having considerable overlap with the Big Fivepersonality constructs, especially neuroticism, which predicts self-efficacy.Notably, no objective measure of job performance criteria, which mighthave elucidated the veracity of this claim, was collected.

Finally, Jordan, Ashkanasy, Hartel, and Hooper (2002) have demon-strated that coaching can improve the effectiveness of low EI teams so thattheir performance is functionally identical to that of high EI teams.

Empirical Evidence: Negative or Mixed Results.

In a recent review, Dulewiczand Higgs (2000) noted that while the concept of EI is purportedly basedon extensive research evidence, the organisational applications of EI “tend

Page 16: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE

385

© International Association for Applied Psychology, 2004.

to be based on derivative arguments and largely anecdotal descriptions”(p. 341). Barrett at al.’s review (2001) concurs that much of the existingevidence bearing on the role of EI in occupational success is anecdotal,impressionistic, or collected by consulting companies and not published inthe peer-reviewed literature. While proprietary data collected in organisa-tional settings may be the surface of a rich and deep research tradition, itis nevertheless of uncertain validity. When a study is submitted for publica-tion in a peer-review journal, although the process is imperfect, it doesprovide some quality control for the methods and results and conclusions(Cherniss, 2001).

Barrett at al. (2001), one of the most vociferous group of critics of the EIconstruct, have argued that the irrational exuberance surrounding EI stemsfrom the concept being inappropriately linked to past research, exaggeratedclaims, and the fact that its major proponents come from particularly pres-tigious colleges. Barrett et al. have identified a number of glaring incongru-ities between assertions made by key proponents regarding EI and the resultsof the actual research they cite. As a case in point, consider Goleman’s(1995, 1998) reference to a study of Bell Laboratory engineers in which thetop performers were reportedly more emotionally intelligent than theirpeers (although not differing in level of general intelligence). A carefulreading of the original report shows that this is pure conjecture—theBell Laboratory engineers were

never

actually tested with any instrumentdesigned to assess EI. Nevertheless, the conclusions of this study, in supportof the important role of EI in occupational studies, have been accepteduncritically. More damaging to the field, perhaps, is the fact that theseunsubstantiated claims have been recycled in numerous popular books andarticles on EI in the workplace (e.g. Cooper & Sawaf, 1997; Gibbs, 1995;Hay Group, 2000).

In fact, several studies examining the predictive validity of EQ in organisa-tional studies show negative results. In the study by Janovics and Christiansen(2001), two self-report measures of EI, the Trait Meta-Mood Scale (Saloveyet al., 1995) and the Schutte EQ test (Schutte, Malouff, Hall, Haggerty,Cooper, Golden, & Dornheim, 1998), were uncorrelated with assessed jobperformance. Self-reported EI showed very little convergent validity withcognitive ability, correlated weakly with performance-based measures, andfailed to demonstrate criterion-related validity.

Fox and Spector (2000) assessed the concurrent validity of three compon-ents of EI (empathy, emotion regulation, and self-presentation), affective traits(positive and negative affectivity), and general and practical intelligence,against the decision to hire (based on the simulated interview), as criterion.Whereas some of the affective and ability measures were related to interviewoutcomes, both directly and mediated by the interviewer’s response, moodregulation was not significantly related to interview outcomes. A mean

Page 17: Zeidner et al. 2004.pdf

386

ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

“Decision to Hire” index, based on a combined measure of hire andqualification ratings and two interviews, was only modestly related to proxymeasures of EI, such as self-regulation (

r

=

.17), perspective-taking (

r

=

.21),and personal distress (

r

=

.19). No effort was made to partial out the effectsof cognitive ability or personality in examining the unique effects of EI inpredicting outcomes.

Slaski (2001) studied 224 middle and senior managers from the UK’s largestsupermarket chain. Data were gathered on EI (via the EQ-i) along withbio-data and measures of distress, morale, quality of working life, andgeneral mental health. Management performance was gauged by assess-ments of immediate line managers who were asked to rate the frequency ofspecific behaviors based on a critical success factor model relating to aspectsof performance (e.g. setting objectives, planning and organising, teamwork, etc.). Whereas the total EQ-i score was moderately related to morale(r = .55), distress (r = −.57), general mental health (r = −.50), and quality ofwork satisfaction (r = .41), it was only very modestly related to managerialperformance (r = .22). Managerial performance correlated modestly withthe Interpersonal factor of the EQ-i (r = .23), but negligibly with the Inter-personal Factor (r = .01) and weakly with Stress Management (r = .15) andAdaptability (r = .18). Even these weak relationships need to be quali-fied since no correlation was partialled with general ability or personalityfactors. More puzzling is an intervention study that was subsequentlyconducted with this sample. As one might expect, compared to a controlgroup, those who underwent an EI training program scored higher on EQsix months following completion of the EQ-i—even when statistically control-ling for initial EQ scores. However, the Management Performance measureshowed no significant improvements in performance.

The Direct Effect of EI in the WorkplaceEI is claimed to affect a wide array of work behaviors, including employeecommitment, teamwork, development of talent, innovation, quality of ser-vice, and customer loyalty. According to Cooper (1997), research atteststhat people with high levels of emotional intelligence experience more careersuccess, build stronger personal relationships, lead more effectively, andenjoy better health than those with low EQ. Why is this so?

First, more emotionally intelligent individuals presumably succeed atcommunicating their ideas, goals, and intentions in interesting and assertiveways, thus making others feel better suited to the occupational environment(Goleman, 1998). Second, EI may be related to the social skills neededfor teamwork, with high EI individuals particularly adept at designing pro-jects that involve infusing products with feelings and aesthetics (Mayer &Salovey, 1997; Sjoberg, 2001). Third, organisational leaders who are high

Page 18: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE 387

© International Association for Applied Psychology, 2004.

on EI, in concert with a supportive organisational climate and the humanresources team, may affect the relationship in the work setting, which, inturn, impacts upon group and individual EI and organisational commit-ment (Cherniss, 2001). EI may also be useful for group development sincea large part of effective and smooth team work is knowing each others’strengths and weaknesses and leveraging strengths whenever possible(Bar-On, 1997). Finally, EI is claimed to influence one’s ability to succeed incoping with environmental demands and pressures, clearly an important setof behaviors to harness under stressful work conditions (Bar-On, 1997).

EI has also been claimed to be an important factor in organisationalleadership. George (2000) used the Salovey, Mayer, and Caruso four-branch model of EI as a heuristic framework for outlining the importanceof EI in effective leadership. George asserts that by accurately identifyinghow followers feel, leaders better appraise and influence followers’ emotionsso they are supportive of leaders’ goals and objectives, thus insuring ashared vision. Leaders can use intense emotions as signals to direct theirattention to issues in need of immediate attention, and can use emotions toprioritise demands. They can also better anticipate how well their followerswill react to different circumstances and changes. High EI leaders areclaimed, according to this model, to generate excitement, enthusiasm, andoptimism in the work environment and are said to be able to maintain anatmosphere of cooperation and trust through the development of highquality interpersonal relations. Leaders can also effectively instill in othersan appreciation of the importance of work activities and convey the messageto their followers that they are optimistic about their personal contributions.Nevertheless, the awareness of negative mood may foster systematic andcareful information processing and may be disadvantageous when leadersare dealing with complex problems in which errors carry high risk. Unfortu-nately, no empirical data have been provided in support of any of George’s(2000) claims. Clearly, what is required is empirical research testing theideas proposed in this paper.

A recent theoretical model proposed by Jordan, Ashkanasy, and Hartel(2002) implicates EI as a moderator variable that predicts employee emo-tional and behavioral responses to job insecurity. According to this model,employees low in EI are hypothesised to be more susceptible than employeeshigh in EI to negative emotions resulting from job insecurity. Therefore,they are more likely to behave defensively and negatively (e.g. hyper-vigilance, “copping out”, “buck passing”, avoidance), lowering affectivecommitment and increased job-related tension in response to their insecurity.These two emotional reactions then lead to negative coping (e.g. distancing,wishful thinking) and defensive decision making behaviors. By contrast,high EI employees are better able to deal emotionally with job insecurityand will be able to ameliorate the effect of job insecurity on their affective

Page 19: Zeidner et al. 2004.pdf

388 ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

commitment. This frequently leads to increased work commitment andeffort, positive coping behaviors (problem-focused), and reframing of per-ceptions of insecurity as an existing challenge. Unfortunately, no empiricaldata were provided in support of this theoretical model and its validityremains to be vindicated.

Empirical Evidence. Empirical research supporting the direct role of EIin the workplace, as the preceding account perhaps implies, is meager.Nevertheless, some research relating EI to occupational satisfaction andcommitment has been conducted. Thus, Bar-On (1997) reports a modestrelationship between total EI scores and job satisfaction in a sample of 314participants (mainly salespersons, teachers, college students, and nurses).Sub-scale scores assessing Self-Regard, Social Responsibility, and RealityTesting predicted about 20 per cent of the variance in work satisfaction.However, the nature of that link varies from occupation to occupation.

SummaryOverall, this section of our review suggests that the current excitement sur-rounding the potential benefits from the use of EI in the workplace may bepremature or even misplaced. Whereas EI appears related to performanceand affective outcomes, the evidence for performance is very limited andoften contradictory. Much of the predictive validity of questionnaire meas-ures of EI may be a product of their overlap with standard personalityfactors. Furthermore, the literature is replete with unsubstantiated general-isations, with much of the existing evidence bearing on the role of EI inoccupational success either anecdotal or impressionistic and/or based onunpublished or in-house research. Thus, a number of basic questions stillloom large: Do emotionally intelligent employees produce greater profitsfor the organisation? Does EI enhance well-being at the workplace? Are theeffects of training in EI likely to result in increases in job performanceand/or work satisfaction?

GUIDELINES FOR FUTURE DEVELOPMENT AND USAGE OF EI MEASURES IN OCCUPATIONAL SETTINGS

Thus far, it may be assumed that our review of EI has largely been negative.Let us assume, however, that the preceding problems may in the future becircumvented inside a carefully controlled (and widely disseminated) sys-tematic program of research. What then? Prior to any widespread use of EIfor occupational and career assessment, EI measures will need to be meticul-ously constructed, standardised (including norming), and validated foruse in specific occupational groups and for particular purposes (selection,

Page 20: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE 389

© International Association for Applied Psychology, 2004.

placement, promotion, and so forth). It is important also that assessment of EIbe cost-effective; the payoffs for psychological screening in pre-employmentselection to rule out emotional or social deficits may vary according to thenature of the job.

Ideally, the standards for developing EI measures to be used for selectionin an organisation should be similar to those of other selection predictors.In this respect, in defence of EI, not all of the most valid predictors aretheoretically based. They evolve from job analyses and are subsequentlyshown to be valid predictors of criteria. This process is not necessarilydesirable, but certainly represents something that is not uncommon in theliterature. Thus, future progress requires a developing synergy betweenempirically focused attempts at improving criterion (and discriminant)validity, and a stronger theoretical and psychometric basis for tests. Thefollowing then are a series of recommendations for further developing thetheoretical efficacy and psychometric adequacy of tests for EI in occupa-tional environments.

A SOLID THEORETICAL RATIONALE FOR THE USE OF EI IN ORGANISATIONAL ASSESSMENT

A vocational (or career-relevant) EI measure will ideally be one with demon-strated theoretical and empirical relevance to a particular occupationalcontext. Unfortunately, the predominant view of EI as an underlying emo-tional competence has not been clearly established, and there are other,equally viable, conceptions of what is actually measured by EI tests (Zeidneret al., 2001). Indeed, with so many disparate definitions of EI, the phenom-enon being dealt with may be entirely different, although the name remainsthe same (see also Zeidner et al., 2001). In general, EI and its componentsshould be differentiated from related constructs in the same conceptualdomain, such as wisdom, practical intelligence, emotional adaptiveness,emotional knowledge, social intelligence, and ego resiliency (Izard, 2001).The schism between ability- and mixed-model approaches to EI is especiallyproblematic.

Other important conceptual questions have been largely ignored: Is EI abasic competence that develops early in life, or a set of acquired skills anditems of knowledge? In the field of cognitive intelligence, Ackerman (1996)has demonstrated that intellectual knowledge (Gk) is distinct from fluid(Gf) and crystallised (Gc) intelligences. Perhaps EI should be assessed asacquired knowledge and, like Gk, assessed through test items tappingspecific content areas for knowledge. Informal “tacit knowledge” might alsobe assessed (Sternberg & Grigorenko, 2000). Is EI expressed primarily asexplicit, “declarative” skills, or as “implicit” procedural skills that aredifficult to express verbally? A procedural conception of EI suggests that

Page 21: Zeidner et al. 2004.pdf

390 ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

tests relying on overt opinions and judgments should be replaced by per-formance tests that might, for example, assess speed of response to suitableemotional stimuli. Does EI even make sense as a purely individual attribute,or should it be conceptualised as a fit between person (P) and environment(E)? A P–E fit conception suggests assessment of EI in terms of the degreeof match between personal and organisational characteristics.

These questions call for a program of psychometric research that aimsexplicitly to operationalise and discriminate different conceptions of EI.A science of EI requires specifying the definition, number, type, and rangeof primary emotional abilities within a formal psychometric model. Thusfar, this disciplined scientific approach to understanding EI has not beenrealised, although Mayer et al.’s (2000a, 2000b) four-branch model is an inter-esting beginning.

Issues of Utility: Matching the Test to the JobAn essential step in constructing EI instruments to meet organisationalrequirements is to identify precisely the specific contexts, needs, and purposesfor which that EI test is being developed. Without sounding trite, differentjobs call for varying levels of social and emotional involvement and activity.Disparate occupations also require different types of interpersonal inter-action. In some jobs (e.g. nursing) one interacts emotionally with othersduring most of their time on the job. Inside such professions, there is a realneed to have frequent interchanges with clients at an emotional level.Incumbents within these jobs not only need to talk with others face-to-faceand exhibit positive, prosaic behavior (e.g. receptionist), but also assess thereactions of others, and attempt to influence others’ emotions and motives(e.g. insurance agent). Some jobs require matching one’s own behavior tothe needs of others (e.g. psychotherapist), creatively influencing others byengaging their emotions, and transforming one’s own emotions and alsothose of others. In other jobs (e.g. mathematician) one interacts with peoplea smaller percentage of time, such that the need to be able to recognise andmanipulate others’ feelings is relatively unimportant, but one may need tomanage personal frustrations.

The preceding account suggests that a systematic emotional task analysisneeds to be conducted in order to match the different facets of EI to the“criterion space” defined by the demands of different kinds of occupation.The selection of the relevant emotional competencies to be assessedneeds to be matched with the relevant career components. For example,an analysis of the criminal justice system may suggest that police officersneed to be able to identify and regulate their aversive emotions. Thus, ameasure of emotional regulation should be developed and included in anassessment battery that the researcher might devise for police officers.

Page 22: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE 391

© International Association for Applied Psychology, 2004.

This measure, in turn, should be validated against behavioral criteria forregulating emotions at work, such as frequency of angry verbal behaviors.It is an open question whether multifaceted general EI measures will beadequate for this purpose. An alternative approach would be to developcontextualised tests oriented towards the particular emotional challenges ofspecific jobs. For example, an EI test for police officers might include testsdefined by job-relevant items, probing, for example, reactions to confronta-tions, to the paperwork intrinsic to modern policing, and to dealing withethnic minorities.

Just as traditional job analysis is increasingly being supplemented bycognitive task analysis, so too we may eventually need different levels ofanalysis for the emotional requirements of jobs. At present, practitionersmay need to rely on a relatively superficial dissection of emotional require-ments. However, as the theory of emotional competence becomes more fullyarticulated, more theory-driven analysis of emotional tasks at work maybecome possible.

ValidationChoosing Appropriate Research Designs. The process of validating an

EI measure requires convincing, empirical evidence that a measure of EIpredicts career success or other important on-the-job criteria. The mostbasic task for validation research is to show that EI measures reliably dif-ferentiate between low- and high-performing groups on particular work-related criteria. Such studies should focus on predicting success both acrossjobs and within jobs, identifying the occupations for which EI is more andless important (e.g. social workers vs. financial analysts). The use of EIcomponent sub-tests also needs to be validated using large-scale, trait-performance validation designs. It is highly plausible that effective perform-ance in different occupations involves different patterns of emotional (orsocial) characteristics.

Throughout we have emphasised the importance of discriminant validitywith respect to existing ability and personality constructs. What EI mightpredict over and above IQ is still an open question. Nevertheless, as onereviewer noted, one may take issue with the notion that EI needs only topredict variance above and beyond ability. Thus, EI may (a) predict differ-ent criterion behaviors than those predicted by cognitive ability or (b)reduce the negative impact of selection based on ability measures alone forspecific social categories (ethnic, social class, gender). In other words, wemight find that (a) it is necessary to broaden the criterion space, or (b)systematic research is needed that demonstrates EI is somehow a less-biasedmeasure than IQ (of which we have doubts, see Matthews et al., 2002). Suchissues aside, EI is only one factor, along with abilities, interests, motivation,

Page 23: Zeidner et al. 2004.pdf

392 ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

and personality traits, that encompass sets of individual difference variablesthat are part of a person’s career profile (Lowman, 1991). EI measuresmight be used together with other variables in the predictor stock in amultiple regression prediction equation of relevant job behaviors, or used ina non-compensatory “multiple-hurdle” framework. In this case, a sequentialmodel is adopted for integration of multiple measures used in any selectionbattery that assesses, in turn, job-relevant abilities, occupational issues, andappropriate measures of EI. Under such a framework, if a person has boththe ability and interest patterns associated with a particular occupationalcluster, the EI factors may be assessed for goodness of fit. While being timeconsuming and expensive, this process will most likely result in more accur-ate assessment.

Uncertainty over the causal role of EI in job success requires longitudinaldesigns tracking the interplay between EI and attainment. High EI may bean eventual consequence of working in fields that involve the problems ofothers, but professional success may be guided by other variables, such asspecific skills and competencies. For example, physicians and judgesmay both score high on EI, yet a person who scores high on EI will notnecessarily make a good doctor or judge. On the other hand, scoring lowon EI (e.g. low emotional regulation) may constitute grounds for exclusionfrom certain occupations (e.g. social work, police work, clinicians, andteachers), provided it can be demonstrated that low EI is meaningfullyassociated with unacceptable performance in these occupations. An import-ant task for future research is establishing cut-off points that may be usedfor exclusion.

Choosing Adequate Criterion Measures. The impetus of proponents ofEI in the workplace should be on testing the validity of EI in predicting awide array of meaningful criteria. As a first step, it would seem importantto look for the variance explained by EI with regard to conventional criteria(supervisor’s ratings of performance, objective criteria such as sales, absentee-ism, etc.) and whether EI remains predictive with IQ and personalityfactors statistically controlled. It is not clear whether these criteria shouldbe recast somewhat to reflect the importance of emotional factors in theworkplace. In any case, the criteria against which EI predictors in occupa-tional selection and placement are validated should be valid, reliable, anduncontaminated. Questionnaire measures, in particular, may be subject tocriterion contamination: i.e. the criterion measure itself has been based, atleast in part, on predictor measures (see Cohen & Swerdlik, 1999). Forexample, the EQ-i (Bar-On, 1997) includes scales for general mood, whichmight be seen as a criterion rather than a predictor. Choice of criterionmeasures requires an understanding of the relevance of the criterion to theorganisation, as we next discuss.

Page 24: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE 393

© International Association for Applied Psychology, 2004.

Assessment in the Service of the Organisation

Assessment of EI should meet the needs of the organisation: it is essentialto distinguish the well-being of the organisation from the well-being of theemployee. In the context of P–E fit, Schneider, Kristof-Brown, Goldstein,and Smith (1997) identify a “dark side” to good fit. High levels of individualsatisfaction may lead to inflexibility and low adaptability. At the organisa-tional level, cooperation and harmony are beneficial in the short run, but,over the longer term, lead to institutional complacency and failure torecognise the need for change. Strategic long-term decision-making may beespecially vulnerable to such dangers. Schneider et al. recommend hiringdecision-makers for diversity of values, competencies, and inclinations.The resulting conflicts and turmoil may, in the long term, support theorganisation’s capacity to adapt to changing events.

Research on coping and adaptation similarly points towards the diffi-culties of deciding which coping strategies are most effective in any givensituation (Matthews & Zeidner, 2000). Although high EI is correlated withuse of coping strategies seen as desirable, such as high task-focus (Bar-On,1997), these characteristics are not automatically beneficial to the organisa-tion. Often, choices of coping produce a pattern of costs and benefits overa period of time. For example, a task-focused employee may succeed infinding a comfortable niche within the organisation that fails to maximisehis or her potential to make a contribution to it.

Hence, if using EI tests, organisations need to investigate exactly whatqualities are being selected, and how a preponderance of these qualities willinfluence the organisation over shorter and longer time periods. Certainly,selection of adaptable, emotionally aware, optimistic, and socially skilledindividuals has potential benefits. However, there are obvious dangers shouldit transpire that the organisation comes to be mainly peopled by Machiavellian,narcissistic, or superficially smooth individuals; those who can make a goodfirst impression, but may lack more substantial personal qualities.

CONCLUSIONS

Despite the important role attributed to a wide array of emotional com-petencies in the workplace, there is currently only a modicum of researchsupporting the meaningful role attributed to EI (and nested emotional com-petencies) in determining occupational success. Many of the popular claimspresented in the literature regarding the role of EI in determining worksuccess and well-being are rather misleading in that they seem to presentscientific studies supporting their claims, while in fact failing to do so. Inshort, despite some rather fantastic claims to the contrary, the guidingprinciple appears presently as “caveat emptor”.

Page 25: Zeidner et al. 2004.pdf

394 ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

Notwithstanding claims made by proponents of EI with respect to theimportant role of EI in career assessment, EI should probably not currentlybe included as part of standard job selection (or classification) batteries.Instead, EI should be used only where warranted by the job description.Accordingly, when particular emotional skills are part of the job description(e.g. empathy, conflict resolution), one might usefully assess EI, recognisingthat considerable skills in professional judgment will be required to interpretthe results with respect to organisational needs. Care must be taken in usingself-report instruments especially, because of their overlap with standard per-sonality scales. By contrast, in those jobs where adequate emotional skillsare really minimal, assessing EI is unlikely to be cost-effective.

Furthermore, EI measures should be used in occupational contexts onlyif the instruments are specifically developed, normed, and validated to thatend, and demonstrate adequate occupational relevance. Thus, in occupa-tional contexts it is probably best to avoid using some of the more prevalentbroad-brush omnibus EI measures (e.g. MEIS, EQ-i) originally designed forresearch and general assessment purposes, until such time as more valida-tion studies using occupational criteria have been published.

There is presently an urgent need for sound taxonomic research thatfocuses on determining the EI constructs that are crucial for performancein particular jobs and for identifying the relevant EI measures that bestassess these affective constructs. While EI may be shown in the future toreduce adverse impact in selection, recent research indicates that the use ofpersonality testing did not compensate for the adverse impact related tocognitive ability testing (Ryan, Ployhart, & Friedel, 1998). These authors goon to suggest that caution be exercised in presuming a reduction in adverseimpact by the addition of personality measures.

In general, the literature shows that the predictive validity of generalmental ability (i.e. “g”) is far from perfect. Thus, looking for better predic-tors with lower levels of adverse impact is well advised and ideologically,legally, and politically defensible. Even so, using EI as a predictor withoutvalidation is an erroneous and potentially damaging practice, since itappears premature to determine that measures of EI are a worthwhileselection tool. At the same time, while there is hype surrounding EI in organ-isational settings, most of it without scientific basis, over time there mayconstitute a body of research pointing to the usefulness of EI in the work-place. Indeed, future research may demonstrate that EI facets provideimportant dimensions otherwise missing from the conventional batteriesassessing ability and interests.

In sum, while the jury is still out on the utility of EI for occupational selec-tion and performance it would appear rash to dismiss the potential value andimportance of EI in all occupational settings. The fact that there are domainsof work where the handling of emotional encounters is pivotal renders EI,

Page 26: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE 395

© International Association for Applied Psychology, 2004.

even if a theoretical soup-stone (see Matthews et al., 2002), highly influentialand we believe important. Systematic, validated research studies, based onthe guidelines suggested above, will inform us if, when, and how a moreclearly defined EI can be effectively used in occupational settings.

REFERENCES

Ackerman, P.L. (1996). A theory of adult intellectual development: Process, person-ality, interests, and knowledge. Intelligence, 22, 227–257.

Bachman, J., Stein, S., Campbell, K., & Sitarenios, G. (2000). Emotional intelligencein the collection of debt. International Journal of Selection and Assessment, 8,176–182.

Bar-On, R. (1997). The Emotional Intelligence Inventory (EQ-I): Technical manual.Toronto: Multi-Health Systems.

Bar-On, R. (2000). Emotional and social intelligence: Insights from the EmotionalQuotient Inventory. In R. Bar-On & J.D.A. Parker (Eds.), The handbook ofemotional intelligence (pp. 363–388). San Francisco, CA: Jossey-Bass.

Barrett, G.V., Miguel, R.F., Tan, J.A., & Hurd, J.M. (2001). Emotional intelligence:The Madison Avenue approach to science and professional practice. Unpub-lished paper.

Barrick, M.R., & Mount, M.K. (1991). The Big Five personality dimensions and jobperformance: A meta-analysis. Personnel Psychology, 44, 1–26.

Barrick, M.R., & Mount, M.K. (1993). Autonomy as a moderator of the relation-ships between the Big Five personality dimensions and job performance. Journalof Applied Psychology, 78, 111–118.

Boyatzis, R. (1982). The competent manager. New York: Wiley & Sons.Boyatzis, R., Goleman, D., & Rhee, K. (2000). Clustering competence in Emotional

Intelligence: Insights from the emotional competence inventory. In R. Bar-On &J.D.A. Parker (Eds.), The handbook of emotional intelligence (pp. 343–362). SanFrancisco: Jossey-Bass.

Campbell, J.P., Dunnette, M.D., Lawler, I.E.E., & Weick, J.R.E. (1970). Managerialbehavior, performance, and effectiveness. New York: McGraw Hll.

Cherniss, C. (2001). Emotional intelligence and organizational effectiveness. InC. Cherniss & D. Goleman (Eds.), The emotionally intelligent workplace (pp. 3–12).San Francisco: Jossey-Bass.

Cherniss, C., & Goleman, D. (2001). Training for emotional intelligence: A model.In C. Cherniss & D. Goleman (Eds.), The emotionally intelligent workplace(pp. 209–233). San Francisco: Jossey-Bass.

Cohen, R.J., & Swerdlik, M.E. (1999). Psychological testing and assessment (4th edn.).London: Mayberry Publishing.

Cooper, R.K. (1997). Applying emotional intelligence in the workplace. Training andDevelopment, 51, 31–33.

Cooper, R.K., & Sawaf, A. (1997). Executive EQ: Emotional intelligence in leadersand organizations. New York: Grosset /Putnam.

Davies, M., Stankov, L., & Roberts, R.D. (1998). Emotional intelligence: In searchof an elusive construct. Journal of Personality and Social Psychology, 75, 989–1015.

Page 27: Zeidner et al. 2004.pdf

396 ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

Dawda, D., & Hart, S.D. (2000). Assessing emotional intelligence: Reliability andvalidity of the Bar-On Emotional Quotient Inventory (EQ-i) in university stu-dents. Personality and Individual Differences, 28, 797–812.

Derksen, J., Kramer, I., & Katzko, M. (2002). Does a self-report measure for emotionalintelligence assess something different than general intelligence? Personality andIndividual Differences, 32, 37–48.

Dulewicz, V., & Higgs, M. (2000). Emotional intelligence: A review and evaluationstudy. Journal of Managerial Psychology, 15, 341–372.

Fox, S., & Spector, P.E. (2000). Relations of emotional intelligence, practical know-ledge, general intelligence, and trait affectivity with interview outcomes: It’s notall just “G”. Journal of Organisational Behavior, 21, 203–220.

Gardner, H. (1983). Multiple intelligences. New York: Basic Books.Gardner, H. (1999). Foreward. In J. Cohen (Ed.), Educating minds and hearts. New

York: Teacher’s College Press.George, J.M. (2000). Emotions and leadership: The role of emotional intelligence.

Human Relations, 53, 1027–1055.Gibbs, N. (1995). What’s your EQ? Time, 2 October, pp. 60–68.Goleman, D. (1995). Emotional intelligence. New York: Bantam Books.Goleman, D. (1998). Working with emotional intelligence. New York: Bantam

Books.Goleman, D. (2001). An EI-based theory of performance. In C. Cherniss &

D. Goleman (Eds.), The emotionally intelligent workplace. How to select for,measure, and improve emotional intelligence in individuals, groups, and organizations(pp. 27–44). San Francisco: Jossey-Bass.

Gowing, M.K. (2001). Measures of individual emotional competencies. In C. Cherniss& D. Goleman (Eds.), The emotionally intelligent workplace (pp. 83–131). SanFrancisco: Jossey-Bass.

Guilford, J.P. (1956). The structure of intellect. Psychological Bulletin, 53, 267–293.Hargreaves, A. (2000). Mixed emotions: Teachers’ perceptions of their interactions

with students. Teaching Education, 16, 811–826.Hay Group (2000). Emotional intelligence: A “soft” skill with a hard edge. <http://

ei.haygroup.com/about_ei />Howard, A., & Bray, D.W. (1988). Managerial lives in transition. New York: Guil-

ford Press.Hough, L.M. (1992). The “Big Five” personality variables—construct confusion:

Description versus prediction. Human Performance, 5, 139–155.Hunter, J.E., & Hunter, R.F. (1984). Validity and utility of alternative predictors of

job performance. Psychological Bulletin, 96, 72–98.Hunter, J.E., & Schmidt, F.L. (1996). Cumulative research knowledge and social

policy formulation: The critical role of meta-analysis. Psychology, Public Policy,and Law, 2, 324–347.

Huy, Q.N. (1999). Emotional capability, emotional intelligence, and radical change.Academy of Management Review, 24, 325–345.

Izard, C. (2001). Emotional intelligence or adaptive emotions? Emotion, 1, 249–257.Jackson, S.E., & Schuler, R.S. (1985). A meta-analysis and conceptual critique of

research on role ambiguity and role conflict in work settings. OrganizationalBehavior and Human Decision Processes, 36, 66–78.

Page 28: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE 397

© International Association for Applied Psychology, 2004.

Janovics, J., & Christiansen, N.D. (2001). Emotional intelligence at the workplace.Paper presented at the 16th Annual Conference of the Society of Industrial andOrganizational Psychology, San Diego, CA, April.

Jensen, A.R. (1980). Bias in mental testing. New York: Free Press.Jensen, A.R. (1998). The ‘g’ factor: The science of mental ability. Westport, CT:

Praeger.Jordan, P.J., Ashkanasy, N.M., & Hartel, C.E.J. (2002). Emotional intelligence as a

moderator of emotional and behavioral reactions to job insecurity. Academy ofManagement Review, 27, 361–372.

Jordan, P.J., Ashkanasy, N.M., Hartel, C.E.J., & Hooper, G.S. (2002). Workgroupemotional intelligence: Scale development and relationship to team process effec-tiveness and goal focus. Human Resource Management Review, 12, 195–214.

Kotter, J.P. (1982). The general manager. New York: Free Press.Lowman, R.L. (1991). The clinical practice of career assessment. Washington, DC:

APA.Matthews, G. (1997). Extraversion, emotion and performance: A cognitive-adaptive

model. In G. Matthews (Ed.), Cognitive science perspectives on personality andemotion (pp. 339–442). Amsterdam: Elsevier.

Matthews, G., & Deary, I.J. (1998). Personality traits. New York: Cambridge Uni-versity Press.

Matthews, G., & Zeidner, M. (2000). Emotional intelligence, adaptation to stressfulencounters, and health outcomes. In R. Bar-On & J.D.A. Parker (Eds.), The hand-book of emotional intelligence (pp. 459–489). San Francisco: Jossey-Bass.

Matthews, G., Zeidner, M., & Roberts, R. (2002). Emotional intelligence: Scienceand myth? Cambridge, MA: MIT Press.

Mayer, J.D., Caruso, D., & Salovey, P. (1999). Emotional intelligence meets tradi-tional standards for an intelligence. Intelligence, 27, 267–298.

Mayer, J.D., Caruso, D.R., & Salovey, P. (2000). Selecting a measure of emotionalintelligence: The case of ability scales. In R. Bar-On & J.D.A. Parker (Eds.), Thehandbook of emotional intelligence (pp. 320–342). San Francisco: Jossey-Bass.

Mayer, J.D., DiPaolo, M., & Salovey, P. (1990). Perceiving affective content inambiguous visual stimuli: A component of emotional intelligence. Journal of Per-sonality Assessment, 54, 772–778.

Mayer, J.D., & Geher, G. (1996). Emotional intelligence and the identification ofemotion. Intelligence, 22, 89–113.

Mayer, J.D., & Salovey, P. (1997). What is emotional intelligence? In P. Salovey &D.J. Sluyter (Eds.), Emotional development and emotional intelligence: Educationalimplications (pp. 3–31). New York: Basic Books.

Mayer, J.D., Salovey, P.R., & Caruso, D.R. (2000a). Emotional intelligence asZeitgeist, as personality, and as a mental ability. In R. Bar-On & J.D.A. Parker(Eds.), The handbook of emotional intelligence (pp. 92–117). San Francisco:Jossey-Bass.

Mayer, J.D., Salovey, P., & Caruso, D.R. (2000b). Competing models of emotionalintelligence. In R.J. Sternberg (Ed.), Handbook of human intelligence (2nd edn.,pp. 396–420). New York: Cambridge University Press.

Mughal, S., Walsh, J., & Wilding, J. (1996). Stress and work performance: The roleof trait anxiety. Personality and Individual Differences, 20, 685–691.

Page 29: Zeidner et al. 2004.pdf

398 ZEIDNER ET AL.

© International Association for Applied Psychology, 2004.

Newsome, S., Day, A.L., & Catano, V. (2000). Assessing the predictive validity ofemotional intelligence. Personality and Individual Differences, 29, 1005–1016.

Petrides, K.V., & Furnham, A. (2000). On the dimensional structure of emotionalintelligence. Personality and Individual Differences, 29, 313–320.

Roberts, R., Zeidner, M., & Matthews, G.M. (2001). Does emotional intelligencemeet traditional standards for an intelligence? Some new data and conclusions.Emotions, 1, 196–231.

Ryan, A.M., Ployhart, R.E., & Friedel, M.L. (1998). Using personality testing toreduce adverse impact: A cautionary note. Journal of Applied Psychology, 83,298–307.

Salgado, J.F. (1997). The five factor model of personality and job performance inthe European Community. Journal of Applied Psychology, 82, 30–43.

Salovey, P., Bedell, B.T., Detweiler, J.B., & Mayer, J.D. (2000). Current directionsin emotional intelligence research. In M. Lewis & J.M. Haviland-Jones (Eds.),Handbook of emotions (pp. 504–520). New York: Guilford Press.

Salovey, P., & Mayer, J.D. (1990). Emotional intelligence. Imagination, Cognitionand Personality, 9, 185–211.

Salovey, P., & Mayer, J.D. (1994). Some final thoughts about personality andintelligence. In R.J. Sternberg & P. Ruzgis (Eds.), Personality and intelligence(pp. 303–318). New York: Cambridge University Press.

Salovey, P., Mayer, J.D., Goldman, S., Turvey, C., & Palfai, T. (1995). Emotionalattention, clarity, and repair: Exploring emotional intelligence using the TraitMeta-Mood Scale. In J. Pennebaker (Ed.), Emotion, disclosure and health(pp. 125–154). Washington, DC: APA.

Salovey, P., Woolery, A., & Mayer, J.D. (2001). Emotional intelligence: Conceptu-alization and measurement. In G. Fletcher & M.S. Clark (Eds.), The Blackwellhandbook of social psychology (Volume 2: Interpersonal Processes) (pp. 279–307).Oxford: Blackwell.

Schmidt, F.L. (1994). The future of personnel selection in the US Army. In M.G.Rumsey, C.B. Walker, & J.H. Harris (Eds.), Personnel selection and classification(pp. 333–350). Hillsdale, NJ: Erlbaum.

Schneider, B., Kristof-Brown, A.L., Goldstein, H.W., & Smith, D.B. (1997). Whatis this thing called fit? In N. Anderson & P. Herriott (Eds.), International hand-book of selection and assessment (pp. 393–412). Chichester: Wiley.

Schutte, N.S., Malouff, J.M., Hall, L.E., Haggerty, D.J., Cooper, J.T., Golden, C.J.,& Dornheim, L. (1998). Development and validation of a measure of emotionalintelligence. Personality and Individual Differences, 25, 167–177.

Sjoberg, L. (2001). Emotional intelligence: A psychometric analysis. European Psy-chologist, 6, 79–95.

Slaski, N. (2001). An investigation into emotional intelligence, managerial stress andperformance in a UK supermarket chain. Unpublished paper.

Sternberg, R.J., & Grigorenko, E.L. (2000). Practical intelligence and its development.In R. Bar-On & J.D.A. Parker (Eds.), The handbook of emotional intelligence(pp. 215–243). San Francisco, CA: Jossey-Bass.

Tett, R.P., Jackson, D.N., Rothstein, M., & Reddon, J.R. (1990). Meta-analysis ofbidirectional relations in personality–job performance research. Human Perform-ance, 12, 1–29.

Page 30: Zeidner et al. 2004.pdf

EI IN THE WORKPLACE 399

© International Association for Applied Psychology, 2004.

Thorndike, E.L. (1921). Intelligence and its measurement: A symposium. Journal ofEducational Psychology, 12, 123–147, 195–216, 271–275.

Watkin, C. (2000). Developing emotional intelligence. International Journal of Selec-tion and Assessment, 2, 89–92.

Weisinger, H. (1998). Emotional intelligence at work: The untapped edge for success.San Francisco: Jossey-Bass.

Williams, W.M., & Sternberg, R.J. (1988). Group intelligence: Why some groups arebetter than others. Intelligence, 12, 351–377.

Zeidner, M., Matthews, G.M., & Roberts, R. (2001). Slow down, you move too fast:Emotional intelligence remains an “elusive” intelligence. Emotions, 1, 265–275.

Zeidner, M., Roberts, R., & Matthews, G.M. (2002). Can emotional intelligence beschooled? A critical review. Educational Psychologist, 37, 215–231.