Bracket, Mayer (2003, PSPB) - Incremental validity of EI.pdf

7/25/2019 Bracket, Mayer (2003, PSPB) - Incremental validity of EI.pdf

1/12

10.1177/0146167203254596 ARTICLEPERSONALITYANDSOCIALPSYCHOLOGYBULLETINBrackett,Mayer/MEASURESOFEMOTIONALINTELLIGENCE

Convergent, Discriminant, and Incremental Validityof Competing Measures of Emotional Intelligence

Marc A. BrackettJohn D. MayerUniversity of New Hampshire

Thisstudyinvestigated the convergent, discriminant, and incre-mental validity of one ability test of emotional intelligence(EI)the Mayer-Salovey-Caruso-Emotional Intelligence Test

(MSCEIT)and two self-report measures of EIthe EmotionalQuotient Inventory (EQ-i) and the self-report EI test (SREIT).TheMSCEITshowed minimal relations to theEQ-i andSREIT,whereas the latter two measures were moderately interrelated.Among EI measures, the MSCEIT was discriminable from well-studied personality and well-being measures, whereas the EQ-iand SREIT shared considerable variance with these measures.After personality and verbal intelligence were held constant, theMSCEIT was predictive of social deviance, the EQ-i was predic-tive of alcohol use, and the SREIT was inversely related to aca-demic achievement. In general, results showed that ability EIandself-report EI are weakly related andyield different measure-ments of the same person.

Keywords:emotional intelligence; personality; behavior; emotions; va-lidity

Research on emotional intelligence (EI) hasexpanded over the lastdecade and today there are a vari-ety of tests to assess it. The three best-known tests are theMayer-Salovey-Caruso Emotional Intelligence Test(MSCEIT) (Mayer, Salovey, & Caruso, 2002a), the Emo-tional Quotient Inventory (EQ-I) (Bar-On, 1997a), andSchutte et al.s (1998) self-report EI test (SREIT). There

is a controversy, however, about what these tests actuallymeasure,what they predict, and whether the tests are dis-tinguishable from other abilities and personality attrib-utes (Hedlund & Sternberg, 2000;McCrae, 2000;Mayer,Salovey, & Caruso, 2000).

Consider theories of EI. Mayer et al.s (2000) originalperformance-based model of EI pertains to an individ-uals capacity to process and reason about emotions.These researchers distinguish their ability model fromother mixed models of EI. They assert that the term EI

has becomeunmoored from both emotion andintelli-gence because so-called mixed models combine mentalabilities (e.g., ability to perceive emotion) with self-

reported qualities such as optimism and well-being thatare clearly distinct from their mental ability approach(Mayer et al., 2000; Roberts, Zeidner, & Matthews, 2002;Salovey, Mayer, Caruso, & Lopes, 2001).

Each approach to measuring EI can influence thevalidity of the construct. For example, in intelligenceresearch, performance scales are standard because theyare based on the capacity to solve mental tasks (Carroll,1993). Self-report scales of intelligence, on the otherhand, arebased onpeoplesendorsementsofdescriptivestatements about themselves. If a persons self-concept isaccurate, then self-report data serve as an accurate mea-sure. However, most people are inaccurate reporters oftheir own abilities. Correlations between ability and self-report measures of intelligence, for instance, are gener-ally low (r= .00 to .35) (Paulhus, Lysy, & Yik, 1998).Therefore, with respect to EI, it is likely that ability andself-report models will yield different representations ofthe same person.

In the present study, one ability-based and two self-report tests ofEI areemployed. TheMSCEIT is designedtomeasure EI as a mentalability. In this conception,EI is

1

Authors Note: Thepreparation of this article wasfacilitatedby a grantfrom the National Science Foundation (Sigma Xi), a Research En-hancement Award from the University of New Hampshire, and a Sum-merFellowship awarded from theUniversity of NewHampshire to thefirstauthor. We thank our colleagues Zorana Ivcevic, Paulo Lopes, andDr. Rebecca Warner from the University of New Hampshire for theircommentsonearlierversionsof thisarticle.The articlealsobenefited agreat deal from the helpful comments of three anonymous reviewersand Paula Niedenthal. Please address correspondence to Marc A.Brackett, Department of Psychology, Yale University, P.O. Box 208205,New Haven, CT 06520; e-mail: [email protected].

PSPB,Vol. 29 No. X, Month 2003 1-DOI: 10.1177/0146167203254596 2003 by the Society for Personality and Social Psychology, Inc.


2/12

the capacity to reason in regard to emotions and thecapacity to use emotions to assist cognition (Mayer &Salovey, 1997). The EQ-i and SREIT are both mixed-model approaches to EI that areassessed with self-reportinventories.TheEQ-i measures an array of noncognitivecapabilities, competencies, and skills that influence

ones ability to succeed in coping with environmentaldemands and pressures (Bar-On, 1997b, p. 14). TheSREIT is a brief self-report scale that is based on Schutteet al.s (1998) understanding of Salovey and Mayers(1990)original model of EI,which broadly defined EI asan ability. There also exist EI scales that are designed fororganizational settings (e.g., Boyatzsis, Goleman, &Rhee, 2000). These tests, which usually require infor-mant reports, were not employed here.

To date, there are no studies comparing the MSCEIT,EQ-i, andSREIT. Dothe three tests assessthe same ordif-ferent things? Are the tests distinguishable from verbalintelligence andmeasuresof personalityandwell-being?

Do the tests predict important behavioral criteriabeyond what can be predicted by well-studied traits?After briefly describing the three EI tests, the presentarticle addresses the questions just raised.

BACKGROUND

ResearchonabilityEI startedwithacademicpsycholo-gists in the early 1990s (Mayer, DiPaolo, & Salovey 1990;Salovey & Mayer, 1990). By mid-decade, the topic waspopularized by Goleman (1995), who made new andextraordinaryclaims about the importance of EI, includ-ing that it is as powerful and at times more powerfulthan IQ (p. 34). Independent reviews of Golemans(1995,1998)popularwritingshave shownthat his claimsare unsubstantiated (Epstein, 1998; Hedlund & Stern-berg, 2000; Mayer et al., 2000; Roberts et al., 2002). Pres-ently, there arethree full-scale tests of EI in the scholarlyliterature. The MSCEIT, EQ-i, and SREIT are three suchtests for which preliminary empirical data are now avail-able. Here, we describe these tests and their generalcharacteristics andthendiscuss whateach test appears topredict.

MSCEIT

Salovey andMayer (1990) first defined EI as the abil-ity to monitor ones own and others feelings and emo-tions, to discriminate among them and to use this infor-mation to guide ones thinking and actions (p. 189). Inthat same year, they also provided the first demonstra-tion of how theconstructmaybe measured (Mayeret al.,1990). These researchers acknowledge that their initialconception of EI was partly a mixed model because itincorporated aspects of personality that might accom-panyemotional intelligence(Mayer et al.,2000, p.402).

Mayer and Salovey (1993)gradually refined their def-inition of EI and argued that it was a real intelligence.They then offered a revised, more focused definition ofEI as the ability to (a) perceive emotion, (b) integrateemotion to facilitate thought, (c) understand emotions,and (d) regulate emotions to promote personal growth

(Mayer & Salovey, 1997). The MSCEIT was designed tomeasure these four abilities.

TheMSCEIT measures perception of emotion by hav-ing people rate how much of a particular emotion isbeing expressed in either a picture of a face expressing abasic emotion or in a picture of a design or landscape.Emotional facilitation of thought is measured by askingpeople to describe emotional sensations andtheir paral-lels to other sensory modalities and by having peopleassimilate predetermined mood into their thought pro-cesses concerning a fictional character. Understandingemotions is measured by asking people how emotions

blend to form more complex emotions and how emo-tional reactions change over time. Finally, the MSCEITmeasures emotion management by having test-takerschoose effective ways to manage private emotions andthe emotions of others in hypothetical situations.

The MSCEIT has a factor structure congruent withthe four-part model of EI and it is both reliable and con-tent valid.Theauthors assert that theMSCEIT meets sev-eral standard criteria for a new intelligence: It isoperationalized as a set of abilities; it is objective in thatanswers on the test are either right or wrong as deter-mined by consensus or expert scoring; its scores corre-late with existing intelligences while also showing

unique variance; and scores increase with age (Mayer,Caruso, & Salovey, 1999; Mayer et al., 2002; Mayer &Geher, 1996).

EQ-i

TheEQ-i(Bar-On, 1997a)is a self-report test ofEI thatevolved outof theauthorsquestion, Why do some peo-ple have better psychological well-being than others?And, why are some people able to succeed in life overothers? (Bar-On, 1997b, p. 1). In the EQ-i manual, Bar-On (1997b) broadly defines EI as addressing

the emotional, personal, social, and survival dimensionsof intelligence, which areoften more importantfor dailyfunctioning than the more traditional cognitive aspectsof intelligence. Emotional intelligence is concernedwith understanding oneself and others, relating to peo-ple, andadapting to and copingwith the immediate sur-roundingsto be more successful in dealing with environ-mental demands. . . . In a way, to measure emotionalintelligence is to measure ones common sense andability to get along in the world. (p. 1)

2 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN


3/12

The EQ-i provides information on five composite fac-tors that are composed of 15 subscales, including (a)intrapersonalEQ,composed of emotionalself-awareness,assertiveness, self-regard, self-actualization, and inde-pendence;(b) interpersonalEQ,composed of empathy,relationship skills, and social responsibility; (c) adapt-

ability, composed of problem solving,reality testing, andflexibility; (d) stress management, composed of stresstolerance and impulse control; and (e) general mood,composed of happiness and optimism. Bar-On (2000),however, recently made a revision to his scale; he nowviews the general mood factor as a facilitatorof EI ratherthana partofit. Thus, total EQ-i scoresare now computedby only summing the first four scales. The comparabilityof the two scoring methods has not been reported.

Bar-On(1997b) haswritten that exploratory andcon-firmatory factor analytic studies indicate that a 15-factorsolution provides a good fit to the EQ-i. The subscalesalso have fairly high internal consistency. The reliability

of the total EQ-i, however, has not been reported. Giventhe diverse factors that comprise theEQ-i, it is importantto know if the scales represent a unidimensional or mul-tidimensional construct.

SREIT

The SREIT is a brief self-report measure of EI thatwasdeveloped by Schutte et al. (1998). These authors wrotea pool of 62 self-report items that were primarily basedon their reading of Salovey and Mayers (1990) earlymodel of EI, which pertained to the ability to monitorand discriminateemotions andto useemotions to guide

ones thinking and actions. For example, some items onthe SREIT measure a persons self-perceived ability tomonitor private feelings or the feelings of others.

Factor analysis of the initial 62 items resulted in thesingle-factor, 33-item SREIT, which has good internalconsistency and test-retest reliability. Petrides andFurnham(2000),however, havecriticized thepsychometricproperties of the SREIT. These researchers claim thatthe scale does not appropriately map onto Salovey andMayers (1990) model of EI and that the scale is notunidimensional. They prefer the results of their explor-atory factor analysis, which divided the SREIT into fourprovisional factors (optimism and mood regulation,

appraisal of emotions, social skills, and utilization ofemotions). Petrides and Furnham have not provideddata on the reliability or validity of these subscales.

Comparative Performance ofthe MSCEIT, EQ-i, and SREIT

What do we know about these three EI tests? Remark-ably, their intercorrelations have not been reported.However, we do know something about each test alone.For example, the MSCEIT and its predecessor test, the

Multi-Factor Emotional Intelligence Scale (MEIS), havebeen correlated with verbal intelligence, the Big Five,and self-reported empathy (Brackett, 2001; Ciarrochi,Chan, & Caputi, 2000; Mayer et al., 1999; Salovey et al.,2001).These preliminary studies show that MSCEIT andMEIS only correlate moderately with these constructs (rs

Documents

Bracket, Mayer (2003, PSPB) - Incremental validity of EI.pdf