Human Abilities: Emotional Intelligence

ANRV331-PS59-19
ARI
5 November 2007
19:32
Human Abilities: Emotional
Intelligence
John D. Mayer,1 Richard D. Roberts,2
and Sigal G. Barsade3
1Department of Psychology, University of New Hampshire, Durham,
New Hampshire 03824; email: jack.mayer@unh.edu
2Center for New Constructs, R&D Educational Testing Service, Princeton,
New Jersey 08541; email: RRoberts@ets.org
3Wharton School, University of Pennsylvania, Philadelphia, Pennsylvania 19104;
email: sigal.barsade@wharton.upenn.edu
Annu. Rev. Psychol. 2008. 59:507–36
Key Words
The Annual Review of Psychology is online at
emotional intelligence, cognitive abilities, emotional knowledge,
http://psych.annualreviews.org
emotional perception, psychological assessment
This article’s doi:
10.1146/annurev.psych.59.103006.093646
Abstract
Copyright c 2008 by Annual Reviews.
Emotional intelligence (EI) involves the ability to carry out accu-
All rights reserved
rate reasoning about emotions and the ability to use emotions and
0066-4308/08/0203-0507$20.00
emotional knowledge to enhance thought. We discuss the origins of
the EI concept, de?ne EI, and describe the scope of the ?eld today.
We review three approaches taken to date from both a theoretical
and methodological perspective. We ?nd that Speci?c-Ability and
Integrative-Model approaches, adequately conceptualize and mea-
sure EI. Pivotal in this review are those studies that address the re-
lation between EI measures and meaningful criteria including social
outcomes, performance, and psychological and physical well-being.
The Discussion section is followed by a list of summary points and
recommended issues for future research.
507

---

ANRV331-PS59-19
ARI
5 November 2007
19:32
Contents
EMERGING RESEARCH IN
Adequate Test Design . . . . . . . . . . . . 516
EMOTIONAL
Validity Evidence from Factor
INTELLIGENCE . . . . . . . . . . . . . . . 508
Structure. . . . . . . . . . . . . . . . . . . . . . 517
THE SCOPE OF EMOTIONAL
Test Relations to Key
INTELLIGENCE . . . . . . . . . . . . . . . 509
Benchmarks . . . . . . . . . . . . . . . . . . . 517
What Is Emotional Intelligence? . . 509
Measurement Issues Regarding
Can Emotional Intelligence Be
Mixed-Model Scales . . . . . . . . . . . 519
Conceptualized Validly? . . . . . . . 509
WHAT DOES EMOTIONAL
The General Scope and Boundaries
INTELLIGENCE PREDICT
of Emotional Intelligence . . . . . . 510
(OR NOT) IN LIFE
APPROACHES TO EMOTIONAL
OUTCOMES? . . . . . . . . . . . . . . . . . . 520
INTELLIGENCE IN THE
Social Relations in Childhood
SCIENTIFIC LITERATURE . . . . 511
and Adolescence . . . . . . . . . . . . . . . 521
Theoretical Approaches to
Social Relations in Adulthood . . . . . 521
Emotional Intelligence . . . . . . . . 511
Scholastic Outcomes from Grade
Speci?c-Ability Approaches
School to College . . . . . . . . . . . . . 523
to Emotional Intelligence . . . . . . 511
Emotional Intelligence at Work . . . 523
Integrative-Model Approaches
Psychological and Physical
to Emotional Intelligence . . . . . . 513
Well-Being . . . . . . . . . . . . . . . . . . . 524
Mixed-Model Approaches
Overall Trends and Intriguing
to Emotional Intelligence . . . . . . 514
Findings . . . . . . . . . . . . . . . . . . . . . . 525
Relating Emotional Intelligence to
DISCUSSION . . . . . . . . . . . . . . . . . . . . . . 526
Other Psychological Variables . . 514
The Scope and Measure
MEASURES OF EMOTIONAL
of Emotional Intelligence . . . . . . 526
INTELLIGENCE . . . . . . . . . . . . . . . 514
Outcomes of Emotional
An Evaluation of Emotional
Intelligence . . . . . . . . . . . . . . . . . . . 526
Intelligence Measures. . . . . . . . . . 514
Concluding Comments . . . . . . . . . . . 526
EMERGING RESEARCH IN
and intuition, and that such feelings were truer
EMOTIONAL INTELLIGENCE
than reason (Reddy 2001). The recently intro-
duced concept of emotional intelligence (EI)
Emotion: an
In Ancient Greece, the development of logical
integrated feeling
thought—syllogisms, arguments, inquiry—
offers a new way of looking at the debate—
state involving
was the burgeoning information technology
that people can reason about emotions and
physiological
of the day. The Stoics of Ancient Greece be-
use emotions to assist reasoning.
changes,
If EI were to exist, some argued, it could
motor-preparedness,
lieved that logic was superior to feelings be-
strengthen our current understanding of both
cognitions about
cause people could agree as to rational ar-
action, and inner
guments but often disagreed as to feelings.
emotions and intelligence (e.g., Sternberg
experiences that
Although Stoic philosophy was in?uential, the
2001). It might enrich our sense of the func-
emerges from an
idea that rationality was superior to emotion-
tionality of human emotion and the breadth of
appraisal of the self
human intelligence. EI also directs attention
or situation
ality was not accepted by all. For example, the
sentimentalists of eighteenth-century Europe
to the role of emotion at home, in schools,
EI: emotional
and at the workplace and how the effects
intelligence
espoused a “follow your heart” credo, arguing
that truth might be a property of one’s feelings
of emotion may ripple through groups and
508
Mayer · Roberts · Barsade

---

ANRV331-PS59-19
ARI
5 November 2007
19:32
society (Barsade 2002, Barsade et al. 2003,
1966) and to promoting personal and social
Ciarrochi et al. 2006, Elias et al. 1997, Izard
improvement more generally (Beasley 1987,
2002, Matthews et al. 2007).
Payne 1986).
Intelligence: a
In this review, we describe research on EI
During the 1980s, psychologists expressed
mental ability (or set
covering a roughly 18-year span from 1990
a renewed openness to the idea of multiple
of mental abilities)
to early 2007. During that time, work on the
intelligences (Gardner 1983, Sternberg 1985).
that permit the
topic expanded from a few articles and book
recognition,
Simultaneously, research on emotion and on
chapters to an active research area. Over the
learning, memory
how emotions and cognition interacted were
for, and capacity to
same period, research continued in emotion,
on the ascendancy (for historical background,
reason about a
intelligence, and their interaction, as re?ected
see Matthews et al. 2002, Mayer 2000, Mayer
particular form of
in Annual Review of Psychology coverage (a par-
et al. 2000a, Oatley 2004). It was amid such
information, such as
tial list includes Cacioppo & Gardner 1999,
verbal information
lively inquiry that scienti?c articles on EI ?rst
Eisenberg 2000, Lubinski 2000, Oatley &
began to appear (Mayer et al. 1990, Salovey
Nomological
Jenkins 1992, Phelps 2006, Rosenbaum et al.
network: the
& Mayer 1990).
2001, Sternberg & Kaufman 1998, Voss &
interconnected terms
Interest in studying EI grew dramatically
and ideas that
Wiley 1995). EI is related to both emotion
throughout the late 1990s, propelled by a pop-
scientists use to
and intelligence, but it also is distinct from
ularization of the topic (Goleman 1995). With
understand their
them.
the term’s newly found cachet, and with the
?eld of study.
Our aim has been to collect what rep-
Scientists’ ideas are
excitement surrounding the identi?cation of
resents, to us, some of the best and most
characterized as
a potential new intelligence, many used the
connected with one
promising research in the EI ?eld. A review
term—but often in markedly different ways
another in logical
of such research can help de?ne EI, indi-
(Bar-On 1997, Elias et al. 1997, Goleman
fashion, and as tied
cate its relation to other concepts, and illus-
1995, Mayer & Salovey 1993, Picard 1997).
to real-world
trate its in?uence on practical outcomes. In
phenomena, in an
So, what does the term “emotional intelli-
the opening of our review, we provide a con-
integrated,
gence” really mean?
meaningful way
text for the present-day ?eld, examine uses
of the term “emotional intelligence,” and de-
scribe the scope of research in the area. Our
Can Emotional Intelligence Be
challenge in covering the ?eld is considerable
Conceptualized Validly?
because the term “emotional intelligence” is
By 2007, the wide diversity of those interested
used in many different ways. One of our goals
in EI was matched by the wide diversity in the
is to identify the core elements of EI and its
conceptions of EI they employed. Some re-
study.
searchers de?ned EI as an ability to reason
about emotion; others equated the concept
with a list of traits such as achievement moti-
THE SCOPE OF EMOTIONAL
vation, ?exibility, happiness, and self-regard.
INTELLIGENCE
Still others found the addition of such traits,
which seemed to be ad hoc, to be troubling,
What Is Emotional Intelligence?
and wondered whether a theoretically sound
The term “emotional intelligence” has been
conceptualization of EI could be identi?ed
employed on an occasional basis at least since
(Locke 2005).
the mid-twentieth century. Literary accounts
of Jane Austen’s Pride and Prejudice refer to
The conceptual network of psychologi-
various characters possessing this quality (Van
cal concepts. A scienti?c concept such as
Ghent 1953, p. 106–107). Scienti?c refer-
EI arises in the context of associated scien-
ences date to the 1960s. For example, emo-
ti?c terms and their meanings. Cronbach &
tional intelligence had been mentioned in re-
Meehl (1955) referred to this context as a
lation to psychotherapy treatments (Leuner
nomological network—a system of meanings
www.annualreviews.org • Emotional Intelligence
509

---

ANRV331-PS59-19
ARI
5 November 2007
19:32
with which most scientists are familiar and
Our view and de?nition of emotion. As
that have been established because of their
an emotion emerges, it entails coordinated
utility. For the term “emotional intelligence”
changes in physiology, motor readiness, be-
Mental ability: a
person’s capacity to
to be valid, it must ?t with such a network of
havior, cognition, and subjective experience
perform a
concepts (or provide a rationale for why it does
(Izard 1993; Parrott 2002, p. 342; Simon
psychological task,
not). We begin by examining some concepts
1982). For example, as a person becomes
such as solving a
that are closely related to EI and then con-
happy, she may experience lower blood pres-
problem, so as to
sider how EI might ?t within this nomological
sure and greater motor readiness to ap-
meet a speci?ed
criterion such as
network.
proach others; she also may smile, think
correctness, novelty,
happy thoughts, and feel good inside. These
or speed
emotional reactions emerge in response to
Our view and de?nition of human mental
perceived or actual alterations in the person’s
abilities and intelligence. Intelligence is a
environment. Our working de?nition of emo-
type of mental ability that concerns the han-
tion appears in the margin.
dling of—and reasoning about—information
Our de?nitions of both intelligence and
of various sorts (Carroll 1993, Spearman
emotion are consistent with longstanding—
1927, Sternberg & Detterman 1986). The in-
we would say, consensual—approaches in
formation involved can be very speci?c (re-
their respective disciplines, but there are al-
lations among auditory frequencies) or very
ternative views of both concepts (Averill &
general (strategic planning). Often, these abil-
Nunley 1993, Kleinginna & Kleinginna 1981,
ities are described as falling along a hier-
Sternberg 1985, Sternberg & Detterman
archy from simple perceptual processes and
1986). For example, some views of intel-
information processing to higher and more
ligences divide the concept into a crystal-
general forms of problem solving (Carroll
lized, learned portion, including especially
1993).
verbal aspects, and into a ?uid portion that
We view intelligence as a general de-
involves on-the-spot reasoning and empha-
scriptive term referring to a hierarchy of
sizes perceptual-organizational and spatial
mental abilities. At the lowest level of this
skill (e.g., Carroll 1993, Vernon 1971). Alter-
hierarchy are basic, discrete, mental abili-
native views of emotion exist as well (Averill
ties. These include, for example, the abil-
1992, Averill & Nunley 1993). Acknowledg-
ity to recognize words and their meanings
ing such complexities, we continue to examine
in the verbal realm, or, as another instance,
how intelligence and emotion might connect
to see how puzzle pieces ?t together in
with EI in a conceptual network.
the perceptual realm, or to understand how
objects are rotated in space. At a middle
level of the hierarchy are broader, cohesive
The General Scope and Boundaries
groups of abilities. These abilities include
of Emotional Intelligence
verbal-comprehension intelligence—a group
Emotional intelligence is a term parallel to
of abilities focused on understanding and
such others as verbal-comprehension intel-
reasoning about verbal information, and, as
ligence, perceptual-organizational intelli-
a second example, perceptual-organizational
gence, or broad-visualization intelligence
intelligence—a group of abilities focused on
(Carroll 1993). In each such term, the
recognizing, comparing, and understanding
descriptor—verbal-comprehension, percep-
perceptual patterns. At the highest level of
tual-organizational,
broad-visualization—
the hierarchy, general intelligence, or g, in-
modi?es the noun: intelligence. For example,
volves abstract reasoning across all such do-
verbal comprehension concerns an individ-
mains. Our working de?nition of intelligence
ual’s understanding and reasoning with verbal
appears in the margin.
information.
510
Mayer · Roberts · Barsade

---

ANRV331-PS59-19
ARI
5 November 2007
19:32
Many forms of intelligence concern learn-
ability to reason verbally, although it may be
ing and reasoning about a particular type of
in?uenced by such reasoning; equating char-
material and then are enhanced further by
acteristics such as assertiveness with the ability
the learning they have fostered. For example,
diverts attention from the intelligence itself.
verbal-comprehension intelligence describes
Returning to EI, its primary focus has to do
the capacity to learn and reason about words
with reasoning about emotions and the use of
and their meanings. The more words one
emotions to enhance thought.
understands, however, the more the verbal
knowledge one already has gained promotes
the intelligence. Thus, verbal intelligence is
APPROACHES TO EMOTIONAL
the ability to reason about words and the
INTELLIGENCE IN THE
use of acquired verbal knowledge to promote
SCIENTIFIC LITERATURE
such reasoning. Perceptual-organizational in-
Theoretical Approaches to
telligence concerns the ability to reason
Emotional Intelligence
about visual patterns and the use of ac-
quired knowledge about patterns to enhance
EI represents abilities that join intelligence
the intelligence. Following such precedents,
and emotion to enhance thought. Some of
an initial working description of EI is as
the abilities that make up EI can be found in
follows:
the top of Figure 1 (see color insert), in the
box labeled “emotional intelligence.” The box
Emotional intelligence concerns the ability
contains speci?c skills, such as the ability to ac-
to carry out accurate reasoning about emo-
curately identify emotion, and indicates that
tions and the ability to use emotions and
these individual skills may also be viewed as
emotional knowledge to enhance thought.
forming an integrated, global EI. Theoretical
To study EI means to focus on the ability
approaches to EI, in fact, can be divided ac-
itself. Some have made the case that charac-
cording to whether they focus on speci?c abil-
teristics such as assertiveness and self-regard
ities or on more global integrations of those
should be considered part of EI because both
capacities.
involve emotion and intelligence to some de-
The speci?c-ability approaches concern
gree. Virtually all mental activities, however,
individual mental capacities important to EI.
from color perception to self-insight, poten-
The integrative-model approaches regard EI
tially involve emotion and intelligence, simply
as a cohesive, global ability. There exists a
because emotion and intelligence are active
third approach to EI as well, called a mixed-
throughout most of one’s mental processes;
model approach to the ?eld (Matthews et al.
that is, mental functions are highly intercon-
2004, Mayer et al. 2000b, McCrae 2000,
nected (Hilgard 1980, LeDoux 2000). EI is
Neubauer & Freudenthaler 2005). This ap-
distinct from other mental processes in in-
proach mixes in a variety of non-EI qualities,
volving a primary focus on a speci?c area of
and, consequently, appears to fall partway or
problem solving.
largely outside the boundaries of the concept
As an analogy, consider again verbal-
(Figure 1, bottom). These three approaches to
comprehension intelligence. The primary fo-
EI are described in detail below.
cus on the meaning conveyed by language is
crucial. Someone could argue, for example,
that assertiveness (or self-regard, etc.) is a part
Speci?c-Ability Approaches
of verbal intelligence because asserting one-
to Emotional Intelligence
self often requires words. The argument fails,
Emotional perception and identi?cation.
however, in regard to the criterion of the pri-
Speci?c-ability approaches to EI focus on a
mary focus. Assertiveness is not part of the
particular skill or skills that can be considered
www.annualreviews.org • Emotional Intelligence
511

---

ANRV331-PS59-19
ARI
5 November 2007
19:32
fundamental to EI. In this section, we out-
of thinking. For example, positive emotions
line some of these abilities, beginning with
promote greater creativity in some contexts
accuracy in emotional perception. The study
(Amabile et al. 2005, Averill & Nunley 1992,
DANVA:
Diagnostic Analysis
of perceptual accuracy grew out of an exten-
Isen 2001, Lyubomirsky et al. 2005).
of Nonverbal
sive body of research in nonverbal perception.
Part of emotional facilitation is to know
Accuracy Scales
Nonverbal perception includes deciphering
how to include emotions in, and exclude
JACBART:
social information, such as power and inti-
emotions from, thought. On the Emotional
Japanese and
macy relationships, along with the accurate
Stroop test (Richards et al. 1992), people ?rst
Caucasian Brief
recognition of emotional expression. From
see neutral words printed in varying colors
Affect Recognition
the nonverbal research, specialized models of
and must say the colors without being dis-
Test
emotional accuracy emerged. For example,
tracted by the words. In a second condi-
LEAS: Levels of
one model aimed to study a person’s accuracy
tion, negative/anxiety emotion words are em-
Emotional
Awareness Scale
at perceiving emotion in child and adult faces,
ployed; in a third condition, positive emotion
voices, and postures (Nowicki & Duke 1994).
words might be employed. It is common for
A number of reviews and key papers provide
people to be distracted and read the emotion
excellent descriptions of research in nonver-
word rather than say the color. Those with
bal sensitivity more generally (e.g., Buck 1984,
higher EI might exhibit less interference from
Hall & Bernieri 2001, Rosenthal et al. 1979).
the emotion words (e.g., Masia et al. 1999,
Two frequently used measures of percep-
Richards et al. 1992).
tual accuracy in emotion are the Diagnos-
tic Analysis of Nonverbal Accuracy Scales
Reasoning about emotions: emotional ap-
(DANVA and DANVA-2; Nowicki & Duke
praisal, labeling, and language. Another
1994) and the Japanese and Caucasian Brief
set of speci?c-ability models concerns emo-
Affect Recognition Test ( JACBART; Mat-
tional reasoning and understanding. For ex-
sumoto et al. 2000), though there are others
ample, emotion-appraisal researchers have
(e.g., Elfenbein et al. 2006). Generally speak-
developed decision rules for matching a given
ing, these scales present pictures of faces and
emotion to the class of situation that has
of postures, gestures, or recordings of voice
elicited it. If a person experiences fear, for
tones; the participant’s task is to correctly
example, it is likely that he is facing a situ-
identify the emotion expressed. For example,
ation that is threatening, raises thoughts of
the DANVA-2 employs stimuli that express
bad things happening, and elicits a need to
one of the four emotions of happiness, sad-
escape (Roseman 1984, p. 210; Scherer et al.
ness, anger, and fear.
2001). Related to such appraisals also are the
accurate labeling and categorization of feel-
Use of emotional information in think-
ings (Clore et al. 1987, Innes-Ker & Nieden-
ing. Some speci?c-ability models address the
thal 2002). Theorists have argued that accu-
ways in which emotions facilitate thinking.
rate appraisal may be a hallmark of emotion-
For example, emotions may prioritize think-
ally intelligent responding (MacCann et al.
ing (Mandler 1975) or allow people to be bet-
2004, p. 41; Parrott 2002, pp. 354–355). If a
ter decision makers (Lyubomirsky et al. 2005).
person’s appraisal process is awry, then he or
A person who responds emotionally to im-
she may misunderstand an event or its conse-
portant issues will attend to the more crucial
quences and react inappropriately.
aspects of his or her life. By contrast, if the
As another example, emotional under-
person is constantly frustrated, say, by her sub-
standing may involve being able to describe
ordinate’s minor clerical errors, then broader
one’s own and others’ feelings. For instance,
concerns that are more important may not
the Levels of Emotional Awareness Scale
be addressed (Parrott 2002). In addition, cer-
(LEAS; Lane et al. 1990) presents 20 emo-
tain speci?c emotions can foster given types
tionally evocative situations involving the test
512
Mayer · Roberts · Barsade

---

ANRV331-PS59-19
ARI
5 November 2007
19:32
taker and a ?ctional person. Participants write
Izard (2001) sometimes prefer to speak
both about how they and the other person
of emotional knowledge as opposed to
would feel in the situation. Responses are
emotional intelligence. Psychologists often
EKT: Emotional
scored according to whether the test taker ap-
speak about an aptitude-knowledge contin-
Knowledge Test
propriately includes emotional responses and
uum (e.g., Lichten & Wainer 2000). At one
MSCEIT: Mayer-
the degree of sophistication (complexity) of
end of this continuum, aptitude refers to the
Salovey-Caruso
those responses, including, for example, the
capacity to reason and learn; at the other end,
Emotional
individual’s capacity to differentiate between
knowledge refers to what a person actually
Intelligence Test
his or her own and others’ responses.
has learned. Both intelligence and knowledge
tests operate according to similar principles
Emotion management. Another relevant
and rely on assessing a person’s knowledge.
ability
area
concerns
emotional
self-
Generally speaking, intelligence tests empha-
management. This area grew out of clinical
size general breadth and rate of learning as
?ndings that, for example, one’s emotionality
well as the ability to reason with unfamil-
could become more positive by reframing
iar problems. Knowledge tests, by contrast,
perceptions of situations (Beck et al. 1979), as
measure attained knowledge. Both concepts
well as from the idea that when at work, in-
?t within the scope of EI studies, as de?ned
dividuals often exert considerable emotional
here.
self-control (Hochschild 1983). A sizeable
amount of research on emotional self-
The Four-Branch Model of Emotional In-
management and regulation has emerged in
telligence. The Four-Branch Model of EI
parallel with that on EI (Gross 1998, Lazarus
is another integrative approach (Mayer &
1994), including in the child development
Salovey 1997, Salovey & Mayer 1990). The
domain (Eisenberg 2000). Denham and
model views overall EI as joining abilities from
colleagues (2003), for instance, have used be-
four areas: (a) accurately perceiving emotion,
havioral observations of children in order to
(b) using emotions to facilitate thought, (c) un-
assess their frustration tolerance, asking ob-
derstanding emotion, and (d ) managing emo-
servers to rate the children’s degree of distress,
tion (Mayer & Salovey 1997, Mayer et al.
crying, and tantrums, among other indices.
2003). Each of these areas is viewed as de-
veloping from early childhood onward. For
example, in perceiving emotion, a person’s
Integrative-Model Approaches
ability to recognize basic emotions in faces
to Emotional Intelligence
is likely to precede the ability to detect the
Izard’s Emotional Knowledge Approach.
faking of emotional expressions (Mayer &
The key element in integrative models of EI
Salovey 1997, p. 10). As skills grow in one
is the joining of several speci?c abilities to ob-
area (e.g., perceiving emotions), so will skills
tain an overall sense of EI. For example, Izard’s
in other areas, such as understanding emo-
Emotional Knowledge Test (EKT; Izard et al.
tions and being able to regulate them.
2001) asks test takers to match an emotion
The Four-Branch Model has been mea-
such as sadness with a situation such as “your
sured by a series of instruments, the most re-
best friend moves away,” as well as to identify
cent of which is the Mayer-Salovey-Caruso
emotions in faces. It provides an integrative
Emotional Intelligence Test, or MSCEIT
measure of EI, focusing in particular on emo-
(Mayer et al. 2002b). This test is composed
tional perception and understanding. Izard’s
of eight individual tasks similar to those de-
test also is important because it is designed
scribed in individual areas above. Two tasks
for use with younger age groups (e.g., as early
are used to measure each branch of the
as 3–4 years old) relative to other measures
model. For example, emotional perception is
of EI.
measured by asking participants to identify
www.annualreviews.org • Emotional Intelligence
513

---

ANRV331-PS59-19
ARI
5 November 2007
19:32
emotions in faces and landscapes. Emotional
Relating Emotional Intelligence
facilitation is assessed, in one subscale, by ask-
to Other Psychological Variables
ing participants to identify which emotions
MEIS: Multifactor
Variables included in mixed models such as
Emotional
promote which kinds of thoughts and activ-
assertiveness and need for achievement surely
Intelligence Scale
ities. Emotional understanding is measured
are important to study—but are not part of EI,
Mixed Model: a
via understanding how emotions blend [e.g.,
as that concept is developed here. A clearer ap-
theoretical approach
“Which two emotions together are closest to
proach is to consider EI a discrete variable and
that equates diverse
contempt: (a) sadness and fear or (b) anger and
then study it in relation to such other charac-
psychological traits,
disgust?”]. Emotional management of oneself
abilities, styles, and
teristics. Several theorists have examined EI
and others is measured by presenting test tak-
other characteristics
in the context of positive and negative affect
to EI
ers with vignettes describing a social situation
and stress tolerance (Izard 2001; Parrott 2002,
and asking them how emotions might be man-
pp. 351–355; Zeidner et al. 2003); others have
aged in the situation (Mayer et al. 2002a). The
positioned EI, the need for achievement, and
MSCEIT replaced the earlier, lengthier, Mul-
other diverse traits in the context of person-
tifactor Emotional Intelligence Scale (MEIS;
ality (Mayer 2005, 2006). These latter models
Mayer et al. 1999).
connect EI to related variables in a way that is
consistent with the great majority of psychol-
Mixed-Model Approaches
ogists’ nomological networks.
to Emotional Intelligence
The third approach to EI is often referred
MEASURES OF EMOTIONAL
to as a Mixed Model approach because of
INTELLIGENCE
the mixed qualities that such models target.
An Evaluation of Emotional
These approaches use very broad de?nitions
Intelligence Measures
of EI that include “noncognitive capability,
competency, or skill” (Bar-On 1997) and/or
In this section, we examine more closely the
“emotionally and socially intelligent behav-
measures proposed to assess emotionally in-
ior” (Bar-On 2004, p. 122), and “disposi-
telligent skills and abilities. Our focus is on
tions from the personality domain” (Petrides
several of the scales introduced above, includ-
& Furnham 2003, pp. 278–280). Tett et al.
ing scales of emotional perception (e.g., the
(2005) drew on Salovey & Mayer’s (1990)
DANVA and JACBART) and emotional un-
original EI model, which they interpreted in a
derstanding (e.g., the LEAS), as well as mea-
broader, more mixed-model fashion than the
sures that integrate across such areas (e.g., the
authors had intended (see Mayer et al. 2000b,
MSCEIT and EKT). We categorize and sum-
p. 401).
marize these and other scales in Table 1.
More concretely, most measures in this
The key purpose of this section is to ask,
category assess one or more EI attributes,
“Do these tests measure what they claim to?”
such as accurate emotional perception, but
In particular, do they measure EI? Standards
then to varying degrees mix in other scales
of test validity have changed and developed
of happiness, stress tolerance, and self-regard
over the past century, and still are develop-
(Bar-On 1997); adaptability, (low) impulsive-
ing. We have distilled from the current Stan-
ness and social competence (Boyatzis & Sala
dards for Educational and Psychological Testing
2004, Petrides & Furnham 2001); and creative
( Joint Comm. Standards 1999) a group of de-
thinking, ?exibility, and intuition versus rea-
sirable criteria that seem particularly relevant
son (Tett et al. 2005). Relative to the concep-
to EI research at this time. These criteria are
tual development we described above, these
grouped into three broad categories: (a) ade-
mixed-in attributes lack a primary focus on
quate test design relative to theories of EI, (b)
EI, as described in this review.
the structure of EI measurement (which tells
514
Mayer · Roberts · Barsade

---

ANRV331-PS59-19
ARI
5 November 2007
19:32
Table 1 A guide to emotional intelligence measures frequently mentioned in the review
Key test name, related tests, and
source(s)
Acronym(s)
Description of the test
Speci?c Ability measures
Diagnostic Analysis of Nonverbal
1. DANVA 2-AF
The Adult Facial version consists of 24 photographs of an equal number of happy,
Accuracy 2
2. DANVA 2-AP
sad, angry, and fearful facial expressions of high and low intensities, balanced also
The test has three versions:
3. DANVA 2-POS
by gender. For this and the related tests described below, the participants’ task is
1. Adult Facial Expressions (Nowicki
to indicate which of the four emotions is present in the stimuli. A youth form is
& Carton 1993)
also available. The Paralanguage version includes two professional actors (one
2. Adult Paralanguage (e.g., auditory)
male, the other female) who say a neutral sentence, “I am going out of the room
(Baum & Nowicki 1998)
now but I’ll be back later” in one of four emotional states. The Posture test
3. Posture Test (Pitterman &
includes 32 stimuli of two men and two women in standing and sitting postures
Nowicki 2004)
representing high- and low-intensity happiness, sadness, anger, and fear.
Japanese and Caucasian Brief Affect
JACBART
Fifty-six Japanese and Caucasian faces are presented in a video format. Each target
Recognition Test
face portrays one of seven emotions: happiness, contempt, disgust, sadness,
(Matsumoto et al. 2000)
anger, surprise, and fear. Each such facial expression is presented for 0.2 seconds
between identical initial and trailing neutral facial expressions posed by the same
individual—that is, between backward and forward masks. The test-taker’s task is
to identify correctly the emotion present.
Levels of Emotional Awareness Scale
LEAS
Twenty social scenes involving two characters, “you” and an additional individual,
(Lane et al. 1990)
elicit four types of emotion: anger, fear, happiness, and sadness. After a test taker
reads a scene, he or she is asked, “How would you feel?” and “How would the
other person feel?” Participants are required to describe their anticipated feelings
(and those of a second person) for each scene. Scoring is according to a
continuum of low emotional awareness (no emotional response) to high
emotional awareness (appropriate emotions for “you” and the character).
Integrative Model measures
Emotion Knowledge Test
EKT (or ACES,
The most recent ACES contains three subscales. Facial Expressions contains 26
(umbrella label for an evolving set
PLE, or EMT)
faces; children are asked if they are happy, sad, mad, scared, or express no feeling.
of tests, including the Assessment of
The Social Situations subscale includes 15 two- to three-sentence vignettes
Children’s Emotional Skills,
describing a social situation; the Social Behavior scale similarly contains 15 two-
Perceiving and Labeling Emotion,
to three-sentence descriptions of behavior; children respond to each scale by
and Emotion Matching Test)
estimating the emotion of the main character. An overall emotion-knowledge
(Izard et al. 2001, Mostow et al.
score is calculated.
2002, Trentacosta & Izard 2007)
Mayer-Salovey-Caruso
MSCEIT; MEIS
Eight tasks (141 items) measure various aspects of EI including emotional
Emotional Intelligence Scale
perception in (a) faces and (b) landscapes, using emotions in (c) synesthesia and in
(Mayer et al. 2002a, Mayer et al.
(d ) facilitating thought, understanding emotional (e) changes across time and ( f )
2003)
blends, and managing emotions in (g) oneself and (h) relationships. Responses are
Multibranch Emotional Intelligence
scored for correctness (e.g., against answers from an expert or consensus-based
Scale (Mayer et al. 1999)
scoring). Each task uses a different item type; different response scales are used
by different tasks. Scores for overall EI as well as Perceiving, Facilitating,
Understanding, and Managing emotions, and other composites, can be
calculated. The longer MEIS test (402 items) consists of 12 scales, also arranged
into four branches; there is considerable conceptual overlap, but no item overlap,
between the two tests.
Mixed Model measures
Emotional Quotient Inventory
EQ-i
A 133-item self-judgment inventory. Items are divided over 15 subscales such as
(Bar-On 1997)
adaptability, assertiveness, and self-regard that also can be formed into ?ve
higher-order factors: intrapersonal, interpersonal, adaptation, stress
management, and general mood.
(Continued )
www.annualreviews.org • Emotional Intelligence
515

---

ANRV331-PS59-19
ARI
5 November 2007
19:32
Table 1 (Continued )
Key test name, related tests, and
source(s)
Acronym(s)
Description of the test
Self-Report Emotional Intelligence
SREIT
A 33-item self-report inventory that has most often been used to assess an over all
Test (Schutte et al. 1998)
level of EI.
Multidimensional Emotional
MEIA
A 118-item self-report inventory employing 10 scales, many of which are based on
Intelligence Assessment
the original Salovey & Mayer (1990) model of EI and some of which are added.
(Tett et al. 2005, 2006)
Measures are organized according to the categories presented in the main body of the text.
DANVA scores are usually reported in articles as “Coding Errors”—we have reversed this to “Coding Skill” in the main body of the text.
us whether EI is one thing or many things),
and then compare their answers to a crite-
and (c) test relationships with key benchmarks.
rion of correctness. Such ability testing elicits
a response process in which a person demon-
Adequate Test Design
strates an ability by actively solving the prob-
lem and then recording a correct answer.
We use the term “adequate test design” to re-
Signi?cant reviews of intelligence—including
fer to evidence of appropriate test content, ev-
those covering dozens of diverse abilities—
idence that test takers employ proper response
rely exclusively on such ability testing (Carroll
processes to answer a question, and evidence
1993). In other words, response-process evi-
of acceptable test reliability.
dence for the validity of an EI measure in-
cludes that the test poses questions of a test
Content evidence of validity. Evidence for
taker and then matches the individual’s an-
a test’s validity includes the extent to which
swers to a criterion of correctness.
a test’s content addresses what should be
The Speci?c Ability and Integrative Model
measured. For example, evidence that the
scales discussed here meet such standards. For
DANVA-2 measures EI comes from the fact
example, the JACBART asks participants to
that the scale presents pictures of emotion-
look at an emotional facial expression and
ally laden faces and body postures to partic-
then match the expression to an emotion.
ipants who must then identify whether the
The correct answer is decided by reference
Response-process
content they see is mostly happy, angry, sad,
to the Facial Affect Coding system, a well-
evidence: a form of
or fearful. As another example, evidence that
regarded system for determining emotional
validity evidence that
the MSCEIT measures EI stems from its con-
facial expressions (Ekman & Friesen 1975).
concerns whether
tent, which is divided into four areas corre-
The MSCEIT has employed two scoring sys-
the questions posed
sponding to the Four-Branch model of EI:
by a test elicit the
tems. The expert-consensus scoring method
actual
the capacities to (a) perceive emotions, (b) use
involves matching a participant’s response to
to-be-measured
emotions to facilitate thought, (c) understand
the correct answers nominated by emotion ex-
mental activities
emotions, and (d ) manage emotions. For ex-
perts. The general-consensus scoring method
targeted for study
ample, item content re?ecting understand-
matches participant answers to the preferred
Speci?c ability
ing emotion provides a participant with an
responses of the standardization sample. The
approach: a
emotion de?nition and then asks him or her
rationale for the latter method is that, because
theoretical approach
to select the emotion that was de?ned (see
to EI focused on a
human beings have evolved to understand
speci?c skill area
Figure 2, see color insert).
emotional information, unselected groups of
within the domain
people can identify correct scores almost as
(e.g., effective
Response-process evidence of validity.
well as can experts. A study of test scores,
emotional
The standard practice in measuring mental
assessed across roughly 2000 individuals, in-
management)
abilities is to ask people to solve problems
dicated that these two scoring methods are
516
Mayer · Roberts · Barsade

---