Will the Real Emotional Intelligence Please Stand Up?
On Deconstructing the Emotional Intelligence Debate
Catherine S. Daus
Southern Illinois University Edwardsville
Neal M. Ashkanasy
The University of Queensland
Neal Ashkanasy and I (Catherine Daus) had the distinct pleasure of debating
Ed Locke and Frank Landy on the construct of emotional intelligence at this past SIOP Annual Conference in Orlando. I, unintelligently (emotionally or otherwise!), agreed to substitute for Peter Salovey who had professional commitments that precluded him from attending. Apart from the discomfort of being in Frank Landys direct line of fire (to which anyone who has had the privilege could certainly attest) and the sheer intimidation factor of standing and speaking to an audience of standing-room-only capacity (Neal calculated about 350+ people), I feel the experience was enlightening and valuable to me, as I learned what concerns academics in our field have with the construct of emotional intelligence. However, it was also quite frustrating as I feel that some academics in our field have a rather limited exposure and a very narrow-minded, unsubstantiated view of the construct.
The primary arguments proffered by Landy and Locke can be captured in three broad points:
1. Emotional intelligence is little more than a loose conglomeration of extant personality traits.
2. Emotional intelligence does not meet psychometric standards.
3. Emotional intelligence has no clear measurement rubricit changes all the time.
Our responses to these three points are integrally linked, particularly when we connect them to our starting point, which is an ability model of emotional intelligence, exemplified by the research of Mayer and colleagues (Mayer, Caruso, & Salovey, 1999; Mayer, Salovey, & Caruso, 2002; Mayer, Salovey, Caruso, & Sitaraneos, 2001, 2003). Thus, our key point is:
I-O psychologists should view emotional intelligence as based in ability and behavior, rather than in self-reports of attitudes, preferences, and/or values.
Let us first begin by making one point crystal clearwe do not endorse a Goleman (1995) or Bar-On (1997) type of approach to studying emotional intelligence in the workplace. These models may indeed be useful for organizational development and interventions, but they are much too broad in scope, and do not appear to markedly differ from traditional personality models or competency models. In addition, these two approaches have generated assessment devices that are based upon self-report, yielding self- and other-perceptions of these traits rather than an estimate of a persons actual emotional ability.
In fact, we did not disagree with Locke and Landys points regarding the questionable nature of these models, and we also feel that to an extent, they have done much more harm than good regarding establishing emotional intelligence as a legitimate, empirical construct with incremental validity potential. It is unusual, and sometimes disturbing, to see scientists and researchers uncritically cite a popular trade book alongside peer-reviewed articles.
Rather, we suggest that the only approach that makes sense is one that meets these criteria: (a) uses a skill-based or behavioral, preferably nonself-report method of measurement; (b) focuses narrowly and specifically on emotional skills and abilities only; (c) has demonstrated construct distinctiveness; and (d) has demonstrably good psychometric properties.
The Mayer et al. approach meets these criteria.
We begin with a definition of emotional intelligence. Our definition of emotional intelligence is Mayer et al.s (1999) definition:
Emotional intelligence refers to an ability to recognize the meanings of emotions and their relationships, and to reason and problem-solve on the basis of them. Emotional intelligence is involved in the capacity to perceive emotions, assimilate emotion-related feelings, understand the information of those emotions, and manage them (p. 267).
Thus, ability approaches to emotional intelligence conceive of the construct as the ability to solve emotional problems; it is developed in childhood and can be strengthened over time.
According to Mayer et al. (2003), Emotional intelligence involves problem solving with and about emotions (p. 1). Such measures require performance tests to assess the construct. For example, in their most recent measure, the Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT; 2002), the subscales and tests require the participant to view a series of faces and report how much of each of six emotions is present, answer questions about emotional scenarios and responses (e.g., judge how much joy one might experience while planning a party), and solve emotional problems (e.g., decide what response is appropriate when a friend calls you upset over losing her job).
Many other measures, particularly the more popular, commercial ones such as the BarOn Eq-i (1997) and Goleman and colleagues measure, the Emotional Competence Inventory (Boyatzis, Goleman, & Rhee, 2000) sample a broad range of individual differences and as such, tend to show quite substantial overlap with the Big Five (see, for example, Davies, Stankov, & Roberts, 1998; Dawda & Hart, 2000). The danger of this, of course, is that we are simply seeing old wine in new wineskins and calling something by a fancy, sexy new name, emotional intelligence, when it is simply measuring some other well-established personality trait. We believe that this was at the heart of Lockes argument (and a large part of Landys) and suspect that they might agree with the following statement by noted emotion researcher, Joseph Ciarrochi, and colleagues (Ciarrochi, Chan, Caputi, & Roberts, 2001): Generally, we should not make up a test and call it EI when really it is a measure of some other, well-established personality trait (p. 28). We also couldnt agree more! The Mayer, Salovey, Caruso model has shown little to no overlap with the Big Five personality constructs (Mayer et al., 1999; Mayer et al., 2001) and small to medium overlap with other, well known personality constructs such as positive and negative affectivity (Ciarrochi, Chan, & Caputi, 2000; Ciarrochi et al., 2001). We concur with Ciarrochi and colleagues (2001) assessment: Perhaps one of the greatest strengths underlying the MEIS is its distinctiveness ( p. 31).
The original measure of Mayer, Salovey, and Caruso, the Multifactor Emotional Intelligence Scale (MEIS) was the first comprehensive ability measure of emotional intelligence. Psychometric evidence for this measure, its reliability, convergent and discriminant validity were all reasonable. In the debate, the point was raised by Landy that some of the subscales of the MEIS had low reliabilities. This is indeed the case. While we feel it is a moot point to clarify since the newer MSCEIT does not suffer from this problem, we do wish to inform the reader, lest he or she be suspicious that we are ignoring important contradictory evidence. Depending on the type of scoring chosen, the internal consistency reliabilities for the 12 subscales (tapping four branches of Perceiving, Facilitating, Understanding, and Managing emotions) ranged from .49 to .94 (consensus scoring), and .35 to .86 (expert scoring; Mayer et al., 1999). Total branch reliabilities were all higher, reasonable, and their use recommended by the authors. We feel it critical to emphasize that one or two subscales with less than desirable reliabilities, for a measure in its theoretical, psychometric, and empirical infancy, is not only quite reasonable, but also to be expected as a natural step in construct specification and measurement development. In fact, their revised measure, the MSCEIT, addressed and rectified most of the low reliabilitiesbranch score reliabilities for this measure ranged from .7691. Further, a confirmatory factor analysis provided convincing support for their proposed four-branch factor structure (Mayer et al., 2003). Advances in scoring have also been made with the MSCEIT, addressing earlier concerns with the first-generation test. It is certainly far from a perfect measure, but seemingly robust enough for serious use.
In the debate, Landy quipped that, much like the weather, if you dont like how emotional intelligence is measured, stick arounditll change. The implication, of course, is that this measure cant be trusted; that researchers in the area of emotional intelligence dont know what they want to measure or how to measure it. To counter, first, as we directly said in the debate, and as we reiterate here, any researcher serious about studying emotional intelligence must consider an ability approach, which is focused on emotional problem solving; we feel Mayer, Salovey and Carusos approach sets the gold standard for this. Second, of course measures change! This is the hallmark of empirical inquiry and measurement developmentthe essence of the hypothetico-deductive model. Indeed, we feel it quite remarkable that for a construct that is roughly only a decade or two old, there has been as much work on measurement development and refinement as exists. Im not sure that we were this far along a mere 10 years after Galton measured the craniums of visitors to a fair or Binet asked school children a few simple questions.
Bar-On, R. (1997). The Emotional Intelligence Inventory (EQ-I): Technical manual.
Toronto, Canada: Multi-Health Systems, Inc.
Boyatzis, R. E., Goleman, D., & Rhee, K. S. (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: Theory, development, assessment, and application at home, school, and in the workplace.
San Francisco, CA: Jossey-Bass.
Ciarrochi, J., Chan, A., & Caputi, P. (2000). A critical evaluation of the emotional intelligence construct.
Personality and Individual Differences, 28, 539561.
Ciarrochi, J., Chan, A., Caputi, P., & Roberts, R. (2001). Measuring emotional intelligence. In
Ciarrochi, Forgas, & Mayer (Eds.), Emotional Intelligence in everyday life: A scientific inquiry. Philadelphia, PA: Psychology Press.
Davies, M., Stankov, L., & Roberts, R. D. (1998). Emotional intelligence: In search of an elusive construct.
Journal of Personality and Social Psychology, 75, 9891015.
Dawda, D. & Hart, S. D. (2000). Assessing emotional intelligence: Reliability and validity of the BarOn Emotional Quotient Inventory (EQ-I) in university students.
Personality and Individual Differences, 28, 797812.
Goleman, D. (1995). Emotional intelligence. New York: Bantam Books.
Mayer, J. D., Caruso, D. R., & Salovey, P. (1999) Emotional intelligence meets traditional standards for an intelligence.
Intelligence, 27, 267298.
Mayer, J. D., Salovey, P., & Caruso, D. R. (2002). Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT) Users Manual.
Toronto, Canada: MHS Publishers.
Mayer, J. D., Salovey, P., Caruso, D. R., & Sitaraneos, G. (2001). Emotional intelligence as a standard intelligence.
Emotion, 1, 232242.
Mayer, J. D., Salovey, P., Caruso, D. R., & Sitaraneos, G. (2003). Measuring emotional intelligence with the MSCEIT V2.0.
Emotion, 3, 97105.
October 2003 Table
of Contents | TIP Home
| SIOP Home