ei and performance

7
Industrial and Organizational Psychology, 3 (2010), 159–164. Copyright © 2010 Society for Industrial and Organizational Psychology. 1754-9426/10 COMMENTARIES Emotional Intelligence and Job Performance: The Importance of Emotion Regulation and Emotional Labor Context DANIEL A. NEWMAN AND DANA L. JOSEPH University of Illinois at Urbana-Champaign CAROLYN MACCANN University of Sydney Cherniss (2010) described three issues that need to be addressed en route to considering emotional intelligence (EI) a useful construct for personnel psychol- ogy: (a) empirical evidence that EI pre- dicts job performance, (b) distinguishing between models of EI and models of emotional and social competence (ESC), and (c) some unresolved EI measurement problems. A recent meta-analysis ( Joseph & Newman, 2010) illuminates the first of these topics while considering the second and therefore provides empirical tests for many of Cherniss’s conceptual ideas. Regarding the third issue, we also point to recent devel- opments in EI measurement based on situ- ational judgment testing (SJT; MacCann & Roberts, 2008; Orchard et al., 2009). In this commentary, we focus particu- larly on two key concepts we believe are necessary to understand the relationship Correspondence concerning this article should be addressed to Daniel A. Newman. E-mail: [email protected] Address: Department of Psychology, University of Illinois at Urbana-Champaign, 603 East Daniel Street, Champaign, IL 61820 Daniel A. Newman and Dana L. Joseph, Depart- ment of Psychology, University of Illinois at Urbana- Champaign; Carolyn MacCann, School of Psychology, University of Sydney. between EI and job performance: emotion regulation and emotional labor. First, we emphasize the value of distinguishing between the different facets of EI (i.e., emotion perception, emotion facilitation, emotion understanding, and emotion reg- ulation, as in Mayer, Salovey, & Caruso’s [2000], four-branch EI model). Doing so reveals the critical importance of the emo- tion regulation facet as a mediator of other EI facet effects on job performance. Second, we demonstrate the vital role of context in the EI–job performance rela- tionship, illustrated by stronger EI crite- rion validity findings for high emotional labor jobs. EI, ESC, and Job Performance: Meta-Analytic Evidence of Discriminant and Criterion Validity In a recent meta-analysis relating EI to Big Five personality, cognitive ability, and job performance ( Joseph & Newman, 2010), two varieties of EI were distinguished, echoing Cherniss’s division between EI as an aptitude (EI) and EI as a set of related competencies or behaviors (ESC). From this point on, we will refer to 159

Upload: eyemanne-adnne

Post on 04-Oct-2015

212 views

Category:

Documents


0 download

DESCRIPTION

emotional

TRANSCRIPT

  • Industrial and Organizational Psychology, 3 (2010), 159164.Copyright 2010 Society for Industrial and Organizational Psychology. 1754-9426/10

    COMMENTARIES

    Emotional Intelligence and JobPerformance: The Importanceof Emotion Regulation andEmotional Labor Context

    DANIEL A. NEWMAN AND DANA L. JOSEPHUniversity of Illinois at Urbana-Champaign

    CAROLYN MACCANNUniversity of Sydney

    Cherniss (2010) described three issuesthat need to be addressed en route toconsidering emotional intelligence (EI) auseful construct for personnel psychol-ogy: (a) empirical evidence that EI pre-dicts job performance, (b) distinguishingbetween models of EI and models ofemotional and social competence (ESC),and (c) some unresolved EI measurementproblems. A recent meta-analysis ( Joseph &Newman, 2010) illuminates the first of thesetopics while considering the second andtherefore provides empirical tests for manyof Chernisss conceptual ideas. Regardingthe third issue, we also point to recent devel-opments in EI measurement based on situ-ational judgment testing (SJT; MacCann &Roberts, 2008; Orchard et al., 2009).

    In this commentary, we focus particu-larly on two key concepts we believe arenecessary to understand the relationship

    Correspondence concerning this article should beaddressed to Daniel A. Newman.E-mail: [email protected]

    Address: Department of Psychology, University ofIllinois at Urbana-Champaign, 603 East Daniel Street,Champaign, IL 61820

    Daniel A. Newman and Dana L. Joseph, Depart-ment of Psychology, University of Illinois at Urbana-Champaign; Carolyn MacCann, School of Psychology,University of Sydney.

    between EI and job performance: emotionregulation and emotional labor. First, weemphasize the value of distinguishingbetween the different facets of EI (i.e.,emotion perception, emotion facilitation,emotion understanding, and emotion reg-ulation, as in Mayer, Salovey, & Carusos[2000], four-branch EI model). Doing soreveals the critical importance of the emo-tion regulation facet as a mediator ofother EI facet effects on job performance.Second, we demonstrate the vital role ofcontext in the EIjob performance rela-tionship, illustrated by stronger EI crite-rion validity findings for high emotionallabor jobs.

    EI, ESC, and Job Performance:Meta-Analytic Evidenceof Discriminant andCriterion Validity

    In a recent meta-analysis relating EI to BigFive personality, cognitive ability, and jobperformance ( Joseph & Newman, 2010),two varieties of EI were distinguished,echoing Chernisss division between EIas an aptitude (EI) and EI as a set ofrelated competencies or behaviors (ESC).From this point on, we will refer to

    159

  • 160 D.A. Newman, D.L. Joseph, and C. MacCann

    performance-based tests founded on thefour-branch ability model as EI and to[usually self-report] rating scales based onmixed-model definitions of EI as ESC.The relationship between EI and ESCis small but not trivial (rcorrected = .26;Joseph & Newman, 2010; cf. Van Rooy,Viswesvaran, & Pluta, 2005), suggesting thetwo constructs are related but not the same.Theoretically, such a result is consistentwith Chernisss proposed developmentalprocess where EI may support or enable thedevelopment of emotional competencies(ESCs). However, it is also consistent withthe converse idea that ESC may leadto the development of EI (e.g., positiveemotional temperaments [ESC] lead tomore positive interactions with othersand hence more learning opportunitiesto develop emotional abilities [EI; seethe developmental investment model ofZeidner, Matthews, Roberts, & MacCann,2003]). A dynamic process of reciprocalEIESC causal influence is also plausible.

    To the point, Chernisss distinctionbetween two different varieties of emotion-related characteristics is a welcome clari-fication for the EI field. However, the ideathat the characteristic variously known astrait EI, mixed-model EI, personality EI, andnow ESC constitutes a competence or set ofcompetencies (rather than, for example, apersonality trait or set of personality traits)remains a testable and falsifiable hypothesis(e.g., see de Raad, 2005; Joseph & New-man, in press).

    Cherniss also reiterated the common pro-posal that EI is an ability concept, whereasESC is not (e.g., Mayer, Roberts, & Barsade,2008). This proposal is supported by meta-analytic evidence (Joseph & Newman, inpress a), as EI shows a stronger relationshipwith cognitive ability than does ESC (meta-analytic r = .22 vs. .09), and EI measuresare less redundant with Big Five personality(multiple R = .34) than are ESC mea-sures (multiple R = .68). The large over-lap between ESC and personality impliesthat rating-scale [self-report] mixed-modelassessments of EI (i.e., ESC) might be better

    labeled as personality or temperament thanas a type of competency.

    Next, regarding the EIjob perfor-mance relationship, the overall correlationbetween strictly ability-based EI measuresand supervisor ratings of employee jobperformance is r = .16. (The overall incre-mental validity of ability-based EI beyondBig Five personality dimensions and cogni-tive ability for predicting job performance iszero.) In contrast, the correlation betweenESC and job performance is a largerr = .32, confirming Chernisss hypothesisthat certain ESC competencies should bestronger predictors of certain outcomes thanEI (p. 123). As for explaining why ESC pre-dicts job performance more strongly than EIdoes on average, two explanations appearmost plausible: (a) ESC measures captureachievement motivation and self-efficacy,in addition to Big Five content; and (b) self-report ESC measures include items thatreflect self-assessments of past performance,and as such these ESC instruments can con-taminate the emotional competence mea-sure with their criterion. Despite the appar-ent criterion validity of ESC, we endorse thecontention (see Daus & Ashkanasy, 2005)that ability-based EI measures (i.e., theMayerSaloveyCaruso model) have fargreater construct validity than do ESC mea-sures. That is, it is not entirely clear what mixof constructs ESC measures are designed toassess, and the definition of ESC as anycharacteristic of the person that leads to orcauses effective or superior performance(Boyatsis, 1982; cited in Cherniss, 2010) istautological. This state of affairs creates ascientistpractitioner dilemma in choosingwhether to use EI versus ESC because ESCmeasures have stronger criterion validity,whereas EI measures have a much strongertheoretical basis.

    A Facet-Level Process Modelof EI and Job Performance:The Cascading Model

    Our proposal is to use the four-branchhierarchical theory of Mayer, Salovey, andCaruso (2000) as a more explicit theoretical

  • Emotional intelligence and job performance 161

    basis to develop EI process models of jobperformance. The common definition of EIfrom Mayer et al. involves four branches(i.e., emotion perception, emotion facili-tation, emotion understanding, and emo-tion regulation), such that EI is definedas four distinct constructs rather than onesingle entity. In addition, the theory pro-poses a hierarchical structure among thesefour branches, where abilities in the higherbranches depend or build upon abilitiesin the lower branches. For example, anaccurate appraisal of a situations emotion-relevant dimensions may underlie effec-tive emotion regulation strategies. In thefour-branch hierarchical model, percep-tion underlies facilitation, which underliesunderstanding, which underlies emotionregulation such that emotion regulationforms the apex of the model (Mayer,Salovey, Caruso, & Sitarenios, 2001). Giventhe consensus in the scientific commu-nity that the MayerSaloveyCaruso four-branch definition constitutes a workabledefinition of EI (a consensus Chernissechoes), it is puzzling that much of thediscussion around EI in scientific and tradetexts refers to a single monolithic EI con-cept and has mostly ignored the proposedhierarchical relationships between the EIfacets.

    We believe an important step forwardfor EI research and theory will be to rec-ognize that the subfacets of EI have dis-tinct locations in the nomological network.Based on these ideas, Joseph and New-man (2010) proposed a theoretical modelof the causal relationships among EI facets,personality, cognitive ability, and job per-formance. In this cascading model of EI,shown in Figure 1, emotion perception pre-dicts emotion understanding, which pre-dicts emotion regulation, which in turnpredicts effective job performance. That is,emotion understanding fully mediates therelationship between emotion perceptionand regulation, and emotion regulation fullymediates the paths from other EI facets tojob performance. Emotion facilitation wasexcluded from this model because of itsconceptual and empirical redundancy with

    the other EI dimensions. Empirical supportfor the cascading model (Figure 1) wasderived from several meta-analyses, con-firming the proposed pattern of relationships(for model fit indices and specific details ofthe meta-analyses, see Joseph & Newman,2010).

    Given the known associations of jobperformance with cognitive ability andkey personality variables (Barrick & Mount,2000; Schmidt & Hunter, 2004), thesewere also included in the cascadingmodel (Figure 1). In addition, Conscien-tiousness relates to emotion perception,cognitive ability to emotion understand-ing, and Neuroticism to emotion regula-tion. These proposed relationships werebased on theory from the EI, emotions,and coping literatures. For instance, Con-scientious behavior requires the perceptionof social norms, which can be obtainedfrom the accurate reading of others emo-tional cues (Matsumoto et al., 2000). Also,emotion understanding is proposed to bethe most cognitively saturated of the EIfacets (Mayer et al., 2001). Finally, com-pared with emotionally stable individuals,neurotic individuals use more ineffectiveregulatory strategies such as emotionallyfocused coping, withdrawal, and wishfulthinking; and use fewer effective strate-gies such as cognitive reappraisal (Connor-Smith & Flachsbart, 2007; Gross & John,2003). In the cascading model, emotionregulation fully mediates the effect of Neu-roticism on job performance, meaning thatNeuroticism affects job performance onlythrough the process of emotion regulation.In essence, Joseph and Newmans cascad-ing model confirms that emotion regulationis the active ingredient in EI, and it medi-ates the effects of the other EI facets (as wellas Neuroticism) on job performance (seeFigure 1).

    This cascading model of EI has severalconsequences for organizational researchon EI. First, it suggests that practitionersinterested in applications of EI might getthe biggest bang for their collective buck byhoning in on emotion regulation. Second,the finding that emotion regulation is

  • 162 D.A. Newman, D.L. Joseph, and C. MacCann

    Figure 1. Cascading model of emotional intelligence based on Joseph and Newman(2010). The dotted line (direct effect of Neuroticism on job performance) is not empiricallysupportedthat is, emotional regulation fully mediates the effect of Neuroticism on jobperformance.

    the active ingredient in EI promises toconnect EI to a large and establishedbase of mainstream psychological theory(e.g., Eisenberg, 2000; Gross, 1998). Third,the development of psychometric testsmight most usefully concentrate on emotionregulation rather than on the other threedimensions of EI.

    Regarding this third point (i.e., mea-surement of emotion regulation), severalnew assessments of emotion regulationare being developed in the format ofsituational judgment tests (SJTs). Early SJTresults show promising preliminary evi-dence for construct validity, such asrelationships with life satisfaction, aca-demic performance, and less stress, anxiety,and depression (Freudenthaler & Neubauer,2007; MacCann & Roberts, 2008). In con-tinuation of this work, MacCann, Roberts,and colleagues have noted that writtenSJT vignettes may be contaminated withreading comprehension as a source ofconstruct-irrelevant variance. They havehence developed video-based measures ofemotion regulation and emotion under-standing (Orchard et al., 2009; Roberts,MacCann, Matthews, & Zeidner, in press)with promising initial results (i.e., reliability,

    large correlations with the written SJT mea-sure and cognitive ability). Finally, we notehere that existing SJT measures of emotionregulation focus primarily on the effec-tive down regulation of negative emotionrather than on the effective up regulationof positive emotion. It may, therefore, beuseful for future measures to include sit-uational stimuli that require the effectivedisplay of positive emotion. Early researchby Grandey (2003) and Beal, Trougakos,Weiss, and Green (2006) may be instructivein conceptualizing such new measures.

    Emotional Labor: A KeyContextual Moderator

    In tracing the relationship between EIand job performance, Cherniss entreatsresearchers to pay more attention tocontext. Wong and Law (2002) highlightedemotional labor as an important contextualvariable, demonstrating that EI predictedjob performance ratings more strongly inhigh emotional labor jobs. Emotional laborhas been defined as the degree to whichworkers are expected to express positiveemotion as part of the job (Grandey, 2003).Joseph and Newman (2010) reconfirmedWong and Laws results via meta-analysis,showing that emotional labor moderated

  • Emotional intelligence and job performance 163

    the EIjob performance and ESCjobperformance relationships (we note that,whereas Wong and Law defined emotionallabor to include hiding negative emotions,Joseph and Newman defined emotionallabor as whether the job requires displayingpositive emotions). In the meta-analysis,EI predicted job performance in highemotional labor jobs (r = .22) but showedno relationship to job performance forlow emotional labor jobs (r = .00). Further,the incremental validity coefficient for EI(controlling Big Five and cognitive ability)was positive for high emotional labor jobsbut negative for low emotional labor jobs,meaning that the use of EI measures toselect applicants into low emotional laborjobs may actually harm job performance.EI facet-level results (i.e., distinguishingemotion regulation from the other EIfacets) reaffirmed this conclusion, showingthat emotion regulation positively predictsperformance only for jobs that requirepositive emotional displays. These findingsfurther emphasize Chernisss notion of theimportance of considering context in thestudy of EI.

    Summary

    We have highlighted four points regard-ing research and practice on the EI con-struct. First, new meta-analytic data showthat EI predicts supervisor-rated job perfor-mance, but overall incremental validity isnil. Second, ESC measures are empiricallydistinct from EI, they overlap greatly withBig Five personality, and they may pre-dict job performance more strongly than EI,which creates a sciencepractice dilemmabetween good construct specification (EI)versus maximal criterion validity (ESC).Third, the EI facets of emotion percep-tion, emotion understanding, and emotionregulation relate differentially to job per-formance, cognitive ability, and personal-ity, such that they fit a cascading patternin which emotion regulation is the laststep preceding job performance. We there-fore recommend organizational researchersand practitioners focus more on emotion

    regulation mechanisms and measurement.Fourth, meta-analytic data suggest EI pos-itively predicts job performance in highemotional labor settings (but not in lowemotional labor settings) when emotionallabor is defined as role-prescribed up regu-lation of positive emotion.

    ReferencesBarrick, M. R., & Mount, M. K. (2000). Select on

    conscientiousness and emotional stability. InE. A. Locke (Ed.), The Blackwell Handbook ofprinciples of organizational behavior (pp. 1528).Malden, MA: Blackwell Publishing.

    Beal, D. J., Trougakos, J., Weiss, H. M., & Green, S.(2006). Episodic processes in emotional labor:Perceptions of affective delivery and regulationstrategies. Journal of Applied Psychology, 91,10531065.

    Boyatzis, R. E. (1982). The competent manager: Amodel for effective performance. New York: JohnWiley and Sons.

    Cherniss, C. (2010). Emotional intelligence: Towardclarification of a concept. Industrial and Organiza-tional Psychology, 3, 110126.

    Connor-Smith, J. K., & Flachsbart, C. S. (2007). Rela-tions between personality and coping: A meta-analysis. Journal of Personality and Social Psychol-ogy, 6, 10801107.

    Daus, C. S., & Ashkanasy, N. M. (2005). The case foran ability-based model of emotional intelligence inorganizational behavior. Journal of OrganizationalBehavior, 26, 453466.

    de Raad, B. (2005). The trait-coverage of emotionalintelligence. Personality and Individual Differ-ences, 38, 673687.

    Eisenberg, N. (2000). Emotion, regulation, and moraldevelopment. Annual Review of Psychology, 51,665697.

    Freudenthaler, H. H., & Neubauer, A. C. (2007). Mea-suring emotion management abilities: Further evi-dence of the importance to distinguish betweentypical and maximum performance. Personalityand Individual Differences, 42, 15611572.

    Grandey, A. (2003). When the show must go on:Surface and deep acting as predictors of emotionalexhaustion and service delivery. Academy ofManagement Journal, 46, 8696.

    Gross, J. J. (1998). The emerging field of emotionregulation: An integrative review. Review ofGeneral Psychology, 2, 271299.

    Gross, J. J., & John, O. (2003). Individual differencesin two emotion regulation processes: Implicationsfor affect, relationships, and well-being. Journal ofPersonality and Social Psychology, 85, 348362.

    Joseph, D. L., & Newman, D. A. (2010). Emotionalintelligence: An integrative meta-analysis andcascading model. Journal of Applied Psychology,95, 5478.

    Joseph, D. L., & Newman, D. A. (2010). Discriminantvalidity of self-reported emotional intelligence:A multitrait-multisource study. Educational andPsychological Measurement. (DOI 10.1177/0013164409355700).

  • 164 D.A. Newman, D.L. Joseph, and C. MacCann

    MacCann, C., & Roberts, R. D. (2008). Assessingemotional intelligence with situational judgmenttest paradigms: Theory and data. Emotion, 8,540551.

    Matsumoto, D., LeRoux, J., Wilson-Cohn, W.,Raroque, J., Kooken, K., Ekman, P., et al. (2000).A new test to measure emotion recognition ability:Matsumoto and Ekmans Japanese and Caucasianbrief affect recognition test (JACBART). Journal ofNonverbal Behavior, 24, 179209.

    Mayer, J. D., Roberts R. D., & Barsade, S. G. (2008).Human abilities: Emotional intelligence. AnnualReview of Psychology, 59, 507536.

    Mayer, J. D., Salovey, P., & Caruso, D. R. (2000).Models of emotional intelligence. In R. J. Sternberg(Ed.), Handbook of intelligence (pp. 392420).Cambridge, UK: Cambridge University Press.

    Mayer, J. D., Salovey, P., Caruso, D. R., &Sitarenios, G. (2001). Emotional intelligence as astandard intelligence. Emotion, 1, 232242.

    Orchard, B., MacCann, C., Schulze, R., Matthews, G.,Zeidner, M., & Roberts, R. D. (2009). New direc-tions and alternative approaches to the mea-surement of emotional intelligence. In C. Stough,

    D. H. Saklofske, & J. D. A. Parker (Eds.), Assessingemotional intelligence: Theory, research, andapplications (pp. 321340). New York: Springer.

    Roberts, R. D., MacCann, C., Matthews, G., &Zeidner, M. (in press). Emotional intelligence:Towards a consensus of models, measures, andapplications. Social and Personality PsychologyCompass.

    Schmidt, F. L., & Hunter, J. (2004). General mentalability in the world of work: Occupationalattainment and job performance. Journal ofPersonality and Social Psychology, 86, 162173.

    Van Rooy, D. L., Viswesvaran, C., & Pluta, P. (2005).An evaluation of construct validity: What isthis thing called emotional intelligence? HumanPerformance, 18, 445462.

    Wong, C., & Law, K. S. (2002). The effects of leaderand follower emotional intelligence on per-formance and attitude: An exploratory study.The Leadership Quarterly, 13, 243274.

    Zeidner, M., Matthews, G., Roberts, R. D., &MacCann, C. (2003). Development of emotionalintelligence: Toward a multi-level investmentmodel. Human Development, 46, 6996.

  • Copyright of Industrial & Organizational Psychology is the property of Wiley-Blackwell and its content may

    not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

    permission. However, users may print, download, or email articles for individual use.