Content uploaded by Hillary Anger Elfenbein
All content in this area was uploaded by Hillary Anger Elfenbein on Mar 06, 2015
Content may be subject to copyright.
Team Emotional Intelligence:
What it can mean and how it can impact performance
Hillary Anger Elfenbein
Manuscript version of:
Elfenbein, H. A. (2006). Team Emotional Intelligence: What it can mean
and how it can impact performance. In V. Druskat, F. Sala, & G. Mount (Eds.), The
link between emotional intelligence and effective performance (pp. 165-184). Mahwah,
NJ: Lawrence Erlbaum.
The author wishes to thank Nalini Ambady, Sigal Barsade, Jeff Polzer, and Anita
Williams Woolley for their contribution to this article and to the research it
We are only just beginning to understand the consequences of emotional
intelligence (EI) for workgroups in organizational settings. High among the
benefits emphasized for emotionally intelligent individuals has been greater
effectiveness in working together with colleagues. Thus, EI could be a crucial
component of high-functioning teamwork. However, little academic research has
examined the impact that EI can make for teams. The goal of this article is to
review existing and recent evidence documenting that the emotional intelligence
of teams is a substantial predictor of effective team performance.
I begin by emphasizing that there are two very different ways of thinking
about the EI of teams: first, by examining the EI of the individual members on the
team, and second, by examining how much emotional intelligence team members
display in their interactions with each other. These perspectives do not compete
with each other. Rather, both are valuable, and each provides different insights
and opportunities for both researchers and practitioners. After briefly outlining
these two perspectives, I describe the design of a recent study that provides data
relevant to each perspective. Then, I review in greater detail the evidence for
emotional intelligence as an important predictor of team effectiveness.
What do we mean by “Team Emotional Intelligence”?
What does it mean for a workgroup to be emotionally intelligent? There is
more than one way to think about the emotional intelligence of groups. This
article reviews the two main perspectives addressing this question. First, we can
consider the emotional intelligence of the individual members of the team. A
team may be more effective if its members have greater emotional intelligence,
which is an individual resource that each person can use in their work. Second, we
can consider the degree of emotional intelligence that team members appear to
use when they interact with each other. It is reasonable to expect an emotionally
intelligent team to have healthy and effective emotional dynamics, and to use
emotion productively in order to conduct their work with each other. Instead of
being an individual resource that members can use, the second method looks at
emotional intelligence as a set of norms or patterns about the way people behave
with each other.
Although these two perspectives may at first seem very similar, there can
be important differences. Many of us have worked on teams in which the whole
was more—or less—than the sum of its parts. A team with emotionally average
members might have a spark that ignites them towards exceptional sensitivity
and adeptness in how they relate to each other. Members of some teams just
“get” each other—sometimes after working together extensively or perhaps after
a shared experience, and sometimes right from the start. Conversely, some teams
fall short of their promise, when individuals who are normally quite effective on
their own appear to be “off” when they work with each other.
The major difference between the two perspectives is the focus on the
resources that a team has versus the style of interacting that a team uses. In the
first perspective, examining the EI of individual team members allows us to
understand the individual emotional resources that members have available for
teamwork—that is, the “sum of the parts”. By contrast, in the second perspective,
examining how teams actually use their emotional skills when working together
allows us to understand the dynamics of a workgroup—that is, the “whole,” or
the team emotional intelligence that may or may not be the same as the sum of its
These two perspectives compliment each other, rather than compete with
each other. Figure 1 summarizes the perspectives, with detail to be filled in over
the course of this article, and highlights how they each ask very different
questions about teams. Before presenting the research evidence that team EI
predicts greater effectiveness, I first describe the design of a recent study
conducted in order to examine both perspectives on what it means to examine
emotional intelligence in teams.
Data linking team emotional intelligence and effectiveness
For these two perspectives on group emotional intelligence, below I
discuss relevant examples of previous research. Recent work has documented
links between effective teamwork and team emotional intelligence as measured
by both perspectives. In order to compare and contrast the perspectives more
directly, I highlight data from a new multi-method longitudinal study that I
conducted along with colleague Nalini Ambady from the Department of
Psychology at Harvard. This is the first project to examine group EI using both
perspectives—examining the EI that individuals have and also the EI that team
members use with each other. This study demonstrates that groups’ emotional
intelligence is an important predictor of a range of team-level performance
measures, including ratings by senior staff members, retention, and self-reported
outcomes such as performance, liking of colleagues, and team learning.
Why examine accuracy in communicating emotion?
In the longitudinal study that I discuss below, the particular aspect of
emotional intelligence that I examined was accuracy in the communication of
emotion. That is, to what extent can team members understand their colleagues’
emotional expressions? Likewise, to what extent can team members express their
own emotions clearly? At first this skill may seem out of place in a business
setting, but in fact we use it continually in order to get our work done. For
example, a supervisor might believe that an employee has just made an excellent
presentation. In that case, does the employee correctly perceive the supervisor’s
positive reaction to the presentation—or, instead, is the employee uncertain what
the supervisor thinks, or perhaps does the employee even believe that the
supervisor did not like the presentation at all? In this example, note that the
emotional content is related directly to the work itself, where the employee needs
to understand the supervisor’s emotional reaction as a form of feedback.
The longitudinal study focused on emotional communication skill for three
reasons. First, the ability to use emotion as a channel of communication is a core
component of emotional intelligence (Mayer, Caruso, & Salovey, 1999; Mayer,
DiPaolo, & Salovey, 1990), and so far some of the best scientific evidence for the
importance of EI in the workplace has come from the positive relationship
between job performance and emotion recognition accuracy (Elfenbein, Marsh, &
Ambady, 2002). For example, in one study, business executives and Foreign
Service officers who were better at identifying the emotional content expressed in
voice samples and video clips also achieved greater performance ratings and
were promoted to higher-level positions (Rosenthal, Hall, DiMatteo, Rogers, &
Archer, 1970). This is because we need to be able to judge our colleagues’
reactions, intentions, preferences, and likely future behaviors, in order to work
productively with them.
The second reason to focus on the effective use of emotion as a
communication tool is that, among the various components of EI, it is one of the
most inherently social aspects. Communication—unlike emotion regulation, for
example—simply cannot occur alone. Thus, it is particularly relevant to teams.
The third reason for focusing on the communication of emotion is that it
has the most valid, reliable, and sophisticated set of measurements available
within the field of emotional intelligence. When a new area fascinates researchers
and managers, it can still take many years to reach the level of scientific standards
associated with psychological research. However, the communication of emotion
has been a topic of scientific study for several decades. During that time,
researchers have validated methods for measuring how accurate communication
is—using judgments of photographs, audio recordings of the voice, and video
recordings of body movement. These types of measures are more valid and
reliable than self-report and written test measures (Ciarrochi, Chan & Caputi,
2000; Davies, Stankov and Roberts, 1998; Roberts, Zeidner, & Matthews, 2001).
Self-report measures are often limited because, even when people try to describe
themselves honestly, they can vary greatly in how much self-awareness they have
about their own emotional skills. Pencil-and-paper performance questions
(Mayer, Caruso, & Salovey, 1999), that have a “correct” answer, are also limited
because it can be challenging to capture accurately in words the richness of
emotional intelligence. By contrast, 360-degree performance appraisals can be
extremely valuable when the appraisers have had extensive contact and
experience with the person they are rating. However, these methods are also
vulnerable to rating bias and subjectivity, and are less applicable for initial
screening and hiring. In spite of these challenges, I am optimistic that further
developments within the field will enable the measurement of other components
of EI to catch up with the several decades “head start” for psychologists studying
the communication of emotion.
A study of group emotional intelligence in teams.
In order to examine the effects of team-level emotional intelligence, my
colleague Nalini Ambady and I recently conducted a large-scale longitudinal
study of workgroups (for more detail, see Elfenbein & Ambady, 2002a, 2002b,
2002c). Participants were members of a non-profit public service organization
based in a medium-sized city in the northeastern United States. The organization
is part of the national service program Americorps, which serves as the domestic
version of the United States Peace Corps, providing community service in
underprivileged neighborhoods. Team members were young adults between 17
and 23 years of age serving as full-time employees for one academic year.
Members worked in teams to perform a variety of public service jobs such as
serving as assistant teachers, after-school and day-camp counselors, disaster relief
workers, assistants to local community charities, and in many other public service
roles working mostly with “at-risk” societal groups. The organization paid them
modest compensation and benefits in addition to university scholarships if they
completed the challenging year-long program.
This organization was an ideal environment in which to study emotional
intelligence in teams. First, the organizational design made it easy to study teams
over time, beginning when they were first formed. The groups conducted all of
their work in teams, with 16 teams total and 5-6 active members each. Teammates
were unacquainted before the program began. Senior staff members determined
team composition using a random assignment process that maximized the
demographic diversity of team members. Second, members conducted difficult
work that made emotional skills an important ingredient for their individual and
team effectiveness. The organization is demographically diverse, including a
wide range of ethnic and educational backgrounds.
Participants completed a range of measures associated with the different
perspectives on emotional intelligence in teams. I describe the specific measures
as well as the results below.
Two models of “Team Emotional Intelligence”
The two methods to conceptualize emotional intelligence in groups each
provide a valuable—yet distinct—perspective. In this section, I review the
underpinnings and evidence for thinking about team EI both in terms of the EI
that individual members have, and also in terms of the degree of emotional
intelligence that team members appear to use when they interact with each other.
Emotional intelligence of individual group members
Because we know that emotional intelligence has important consequences
for individuals in the workplace, we suspect—but do not necessarily know—that
the emotional intelligence of individual members should also have consequences
for teams. Indeed, researchers often find it valuable to think about workgroups in
terms of the individuals who are in the group. Emotional tendencies can be
considered as individual traits, and these traits of individuals combine and create
the emotional composition of a group (Kelly & Barsade, 2001). The emotional
composition of a team involves not only the average value for each team member,
but also includes the maximum value, the minimum value, and the diversity in
values across teammates, each of which I discuss below.
Group-level Average of Emotional Intelligence
The most common method of thinking about a psychological phenomenon
at the team level is to take an average value, which “aggregates” individual-level
scores into a single score for the group. The underlying assumption is that
emotional intelligence can be viewed a resource that team members draw upon,
and that members of the team can pool their abilities to share and compensate for
one another. Thus, a higher average level of EI among the individuals in a team
provides a benefit to the team’s performance.
Before going into the research evidence showing that teams with higher
average EI outperform teams with lower average EI, it is worth taking a brief
detour to address an academic debate about whether it is meaningful to use an
average value across individuals in order to describe a team as a whole. Scholars
have debated extensively about whether it is necessary first to demonstrate that
there is a high degree of similarity among team members before calculating an
average value (e.g., Chan, 1998; Klein, Dansereau, & Hall, 1994; Rousseau, 1985).
In the case of emotional intelligence, I argue that this requirement does not apply.
Demonstrating similarity can be a worthwhile safeguard when examining
psychological phenomena such as attitudes or group culture, because it is difficult
to say that group attitudes or cultures exist if colleagues cannot agree upon them.
However, emotional intelligence is different in the sense that it can be viewed as a
kind of individual resource. This analogy makes it clear that it is meaningful to
compare teams with high versus low average values, whether or not individual
team members are similar to each other in EI.
Past research has documented performance consequences for group-level
averages across emotional personality traits. Jennifer George (1990) studied the
emotional tendencies of individuals in teams, and defined “affective tone” as
consistent emotional reactions by members of a workgroup. She found that
groups with more positive affective tone tended to have lower absenteeism, and
groups with less negative tone tended to have greater helping behavior among
members. Likewise, Bouchard (1969) found that group problem-solving
performance was higher in groups that had more sociable members. More
recently, Neuman and Wright (1999) found that teams whose members had
positive, “agreeable” personalities were better able to work cooperatively
towards team objectives. Their social skills allowed the team to communicate
openly and to resolve conflicts and disruptions.
Although this past work had documented the effects of group-averages
with personality traits rather than emotional intelligence, it strongly suggested
that this would be a valuable method of examining team EI. Along with
colleagues Jeff Polzer and Anita Williams Woolley at the Harvard Business
School, I recently studied teams of MBA students participating in business plan
competitions (Elfenbein, Polzer, & Woolley, 2002). The work of these teams is
more than just a course project—approximately one-third of the teams involved
in this contest are developing their plan as the roadmap for a new business
venture. At the beginning of the contest, participants completed a survey that
included a long-standing test of emotion recognition called the Diagnostic
Analysis of Nonverbal Accuracy (DANVA; Nowicki & Duke, 2001). They viewed
a series of photographs of facial expressions, and they indicated the emotion that
they thought was best represented in the photo. The photos included expressions
of happiness, sadness, fear, and anger. This test has been used for nearly a decade
in many dozens of studies by researchers across areas of psychology. It strongly
predicts important elements of life functioning such as academic success and
social adjustment among children and adolescents, and in more recent studies it
has also predicted workplace success among adults. For example, Nalini
Ambady and I found that individuals with higher total scores in recognizing the
emotional expressions on the DANVA test also had higher performance
appraisals from both supervisors and peers (Elfenbein & Ambady, 2002c).
At the end of the business plan competition—but before the official contest
results had been announced—participants completed another survey that
included questions about the way their team had functioned during their work
together. We found that teams whose members had higher average scores on the
DANVA reported that they felt greater psychological safety with each other, had
lower levels of conflict, made decisions more collaboratively together and
experienced greater team learning over the course of their project. These results
argue for the effectiveness of teams with individuals high in emotional
Turning our attention to the large-scale longitudinal study of emotional
intelligence in teams that I conducted along with Nalini Ambady, described
above, we documented a similar pattern of findings. This replication is helpful
because the two studies examined workgroups in very different contexts. The
members of business plan contest teams were older, they had more previous
work experience, and they were students at a prestigious business school. Most
importantly, they chose their own teammates. By contrast, members of the public
service group were younger and less experienced, and they were full-time
employees working with their teammates on a daily basis. They were randomly
assigned to their teams. An additional difference is that we were able to give the
public service participants a longer version of the DANVA test, which included
vocal tones in addition to photographs of facial expressions.
Yet, in both cases, teams with greater average emotional intelligence also
experienced better team functioning. At our public service group, these teams
with high average DANVA scores reported that they had accomplished more in
their work together, and they also had greater retention of their members
throughout the challenging year-long program. Thus, teams with higher average
levels of individual emotional intelligence appear consistently to outperform
teams with lower average levels.
Group-level Minimum and Maximum Emotional Intelligence
Although a group average is valuable as a single measure to summarize
the overall emotional intelligence of a group, the average value is not the only
worthwhile number. Depending on the type of group task, other values may be
more appropriate to describe important features of the group. Closely related to
the average value are several other mathematical functions, such as the maximum
and the minimum value in a team (Barsade & Gibson, 1998).
In 1972, Steiner outlined a typology of group tasks that is helpful to
consider for this purpose. The group average is most useful for examining
“additive” tasks, in which each group member’s contribution is added together
into a common pool of output. Emotional intelligence serves as a resource for
teams, and for many types of tasks it may not matter how that resource is
distributed across individuals—as long as it is available for use.
By contrast, a group’s maximum level of EI among individuals is useful for
exploring what Steiner called “conjunctive” tasks, in which a group output
represents the performance of its strongest member. For some types of work,
having one teammate with exceptionally high emotional intelligence may be
sufficient to assist the entire team. For example, in a negotiation setting with
multiple representatives from each party, one person who is particularly adept at
sensing the interests and tone of the other party can share this information with
teammates, so that the entire group can act appropriately. In other settings, it is
possible for a “good-cop-bad-cop” routine to develop in which the teammate
acting as a “good cop” can undo any tension caused in the process of productive
work by the “bad cop.” In other cases, one colleague with very high EI can serve
as a lightning rod to detect and dissipate tensions that can arise during a team’s
Few researchers have examined the impact on team effectiveness of the
highest level of skill among team members. A notable exception, Williams and
Sternberg (1988) conducted a study using teams of students working on difficult
marketing assignments that required analysis and creativity. They tested a type
of social intelligence—an unwillingness to participate in socially unpleasant
tasks—and found that the maximum level was highly predictive of team
effectiveness. In the study of business plan competitions that I conducted with
Jeff Polzer and Anita Williams Woolley, teams that had a very large maximum
level did not necessarily appear to benefit from that exceptional skill of one or
more individuals. These teams did report that they relied less on rules and
procedures in order to govern their work interactions, and they were less
overwhelmed by the day-to-day work in their teams. However, they reported
that they had somewhat less satisfying relationships among colleagues. This
suggests that individuals who were very highly skilled found it challenging to
use their exceptional skill for the benefit of the whole team. Perhaps a single
individual who stands out from teammates in EI has greater difficulty integrating
socially with them. For the public service group that I examined with colleague
Nalini Ambady, by contrast, no pattern appeared to emerge for the highest scores
for each team.
The minimum level of EI among individuals in a group is most useful for
exploring what Steiner called “disjunctive” tasks, in which a group performance
is only as strong as its weakest link. In the case of emotional intelligence, this may
be true for certain types of teams, for example those that represent their
organization to outside stakeholders, such as a sales team with a goal of 100%
customer satisfaction. In these teams, individual behavior that is emotionally
inappropriate and lacking can reflect poorly on the entire group. Barsade and
Gibson (1998) also note that the lowest value can be important if individuals may
be able to infect their colleagues with their negativity.
Some evidence suggests the benefits for teams that have a high minimum
level of emotional intelligence. In Williams and Sternberg’s (1988) teams working
on marketing problems, the minimum value of their social intelligence measure
did not predict team performance. By contrast, in our business plan study
(Elfenbein, Polzer, & Woolley, 2002), it appeared that there was a large benefit for
teams that had a high minimum score. These are teams in which no one is left
behind, in which each member has a relatively strong level of emotion
recognition. In these teams, the minimum standard was high, and the benefit to
team performance for having a high minimum standard was even greater than
the benefit of having a high average level. Similarly, in the longitudinal study of
our public service group, groups with a higher minimum level of emotion
recognition skill reported a somewhat greater sense that they accomplished their
goals, although this effect was relatively small.
In summary, the research findings demonstrate that a high average level of
individual emotional intelligence of team members predicts stronger team
performance. Teams also appeared to benefit from having a high minimum
standard of EI across individuals. However, teams did not necessarily appear to
benefit from the exceptionally high skill of any one individual. The research
results regarding minimum and maximum skill levels appear to be promising,
but showed some inconsistencies across studies that suggest a degree of caution
in interpreting the findings. However, the results for high average levels of
emotional intelligence were consistent and robust.
Group-level Diversity in Emotional Intelligence
An additional way to examine EI at the team level is to consider the
amount of diversity, or variability, across individual scores in a group. The
underlying assumption is that emotional intelligence can also be viewed as a trait,
and that members of the team who are similar may fit together more smoothly
and may be better able to coordinate their activities.
This perspective draws on research examining diversity in terms of
personality traits, workplace goals and values, demographic characteristics and
functional background and training—which shows that diversity provides
helpful perspectives, but unfortunately can be accompanied by greater challenges
as well. Whether team diversity helps or hinders team performance depends on
the type of diversity as well as the context and environment of the workgroup.
Although diversity along dimensions such as personality and technical skills can
be beneficial, diversity along demographic characteristics such as ethnicity and
gender is often associated with poorer group functioning and performance
(Williams & O’Reilly, 1998).
In general, one would expect that greater similarity in emotional
intelligence among team members could provide a benefit to team performance.
Psychologists frequently find that people show favoritism towards others they
believe are similar to themselves (Byrne, 1971). Barsade and Gibson (1998)
applied this finding specifically to similarity along emotional characteristics.
Thus, they argue that individuals may work better with colleagues who share
their own emotional styles. Barsade and colleagues (2000) recently documented
evidence that emotional diversity does present a challenge for the effective
functioning of top management teams. In their study of Fortune 500 companies,
top management teams benefited both from higher levels of positive affect as well
as from greater similarity in their emotional tendencies. Emotional similarity was
associated with better financial performance of the company as well as more
effective group processes. Further, these two effects interacted with each other, so
that the very worst performers in their study were those teams with both low
average positive affect and high affective diversity.
Although emotional diversity might generally pose a challenge to effective
team functioning, there may also be some contexts and environments in which
emotional diversity could be valuable. Emotional diversity could help teams to
succeed if it provides differences in perspective that are helpful for the team’s
work, and if the diversity is accompanied by a supportive organizational climate
that respects the differences among individuals. Particularly for personality and
other social traits, teams can benefit from a mix of styles. Sometimes, having a
group that is homogeneous can be “too much of a good thing.” As early as the
1950s and 60s, researchers found that participants preferred working with
colleagues with complementary—rather than similar—personality traits
(Haythorn, 1968; Hoffman & Maier, 1961; Rychlak, 1965). More recently,
researchers have found that this is particularly true for extraversion, so that
individuals benefit from having colleagues who differ from themselves (Barry &
Stewart, 1997; Kristof-Brown, Barrick, & Steven, 2001). Although some similarity
can be helpful, researchers found that it was overload to have colleagues who
were all exceptionally outgoing and gregarious.
These findings, taken together, argue for the importance of examining the
impact of diversity in emotional intelligence among individuals in a team.
Emotion can serve as a language. When expressing important messages, people
use nonverbal methods of communicating just as much—or more—than verbal
methods. Thus, the way that we use emotions in the workplace can function like
a language that we speak simultaneously with our spoken language. Using this
metaphor, diversity in the levels of emotional intelligence among teammates can
serve as a language barrier. If some members are skilled with—and,
consequently, accustomed to—using their emotions as a channel for
communicating and coordinating with others, then it may be challenging for
them to work with others who prefer a different method. In this case, diversity
can imply that some colleagues speak one language, and other colleagues speak
This suggests that diversity in emotional intelligence is likely to serve as a
hindrance to team effectiveness. Indeed, recent evidence shows that this is the
case. In our survey examining Harvard Business School students writing
business plans, described above, we also examined the level of diversity in
emotional intelligence. High levels of emotional diversity in the team predicted
poorer team functioning. Teams with more variability in emotion recognition
levels reported that they felt less psychological safety and had more conflict with
their teammates, did not collaborate on decisions as well together and
experienced less team learning. This suggests that teams with diverse levels of
emotional intelligence can find it more challenging to work together.
Our longitudinal study of public service teams, described earlier, also
found that affective diversity presented challenges for group effectiveness.
Teams with less similarity in levels of emotional intelligence reported that they
had accomplished less in their work together, had lower retention through the
end of the year-long program, and were rated less highly by senior staff members
at the organization.
Interestingly, these trends were stronger for the section of the DANVA test
of emotion recognition that included photographs of facial expressions—more so
than the section using audiotapes of vocal tone. Researchers studying the
communication of emotion often distinguish among the various “channels” of the
body through which we express ourselves—facial expressions, vocal tones, and
body movements. Among these, the face is considered the most “controllable.”
That is, we can more easily control our own facial expressions, and we generally
pay more attention to facial expressions than to other types of emotional
expressions (DePaulo, 1992; Elfenbein, Marsh, & Ambady, 2002; Ekman &
Friesen, 1969; Rosenthal et al., 1979). By contrast, the voice is considered the most
“leaky.” That is, it is relatively more difficult to control our vocal tone, and often
our true feelings can leak out through our voice. This is why some newer lie
detection machines use stress analysis to examine small tremors in the voice. The
differences across “channels” of communication suggest that facial expressions
are the expressions of emotion that are the most likely to be noticed,
acknowledged, and discussed in a workgroup setting. Therefore, differences
among teammates in accuracy with facial expressions would be particularly
detrimental. One person may act on, and attempt to discuss, a signal that another
colleague did not even notice. If some members are more sensitive than others, it
can be as if they are speaking a different language. Correspondingly, in the case
of our public service group, teams that were very diverse in understanding facial
expressions had lower liking among colleagues, whereas teams diverse in
understanding vocal tones did not have the same difficulty.
These results argue for the complexity of emotional intelligence, and for
the need to assess EI using a range of methods that assess multiple components.
Although teams appear to experience greater functioning and effectiveness when
their members are highly emotionally intelligent, teams also appear to work better
when members have similar levels of emotional intelligence. The detrimental
effects of “affective diversity” are particularly strong for the components of
emotional intelligence that are the most public and discussable among team
“Team EI”: Using emotional skills when working in a team
The evidence reviewed above focuses on the emotional intelligence of a
team by examining the emotional intelligence of individual group
members—their average value, their minimum value, their maximum value, or
the diversity in their values. However, this is not the only way to focus on the
emotional intelligence of a team, and it is not always necessary to measure the
scores of individual team members. The second main perspective for examining
EI at the team level is to examine the emotional savvy exhibited when the team
members interact with each other. I refer to this second perspective as “team EI.”
The underlying assumption of the second perspective is that emotional
intelligence can be viewed as a process, and that this process can differ across
interaction partners. That is, one person may display more emotionally
intelligent behavior when interacting with colleague A than with colleague B. A
person may display more emotionally intelligent behavior in situation A than in
situation B. We each have a unique emotional style, and the style we use fits
better with certain people and with certain contexts than it does with
others—even after accounting for the individual’s general level of emotional
intelligence. Thus, it can be worthwhile to examine the team-specific emotional
intelligence—that is, the emotional quality of interactions in the team
context. Researchers have often used this perspective by administering surveys
that tap into the use of effective interpersonal processes among teammates.
Researchers can also engage in participant observation, and can conduct
controlled exercises with intact teams. The core distinction between this
perspective and the perspective used in the work with results described above is
the focus on how much emotional intelligence is displayed and actually used in the
interactions among teammates--rather than the fixed individual attributes of
teammates—as a predictor of team performance.
This approach to examining team EI is a natural extension to the definition
of “intelligence.” Psychologist Robert Sternberg (1984) defines intelligence as
“adaptation to, selection of, and shaping of real-world environments relevant to
one’s life” (p. 285). This suggests that the intelligence of a group should be the
ability of that group to collaborate and work interdependently. This is the
“functional intelligence of a group of people working as a unit” (Williams &
Sternberg, 1988, p. 356). By examining the group as a whole, rather than the
individuals who are in it, we can gain an important perspective on what it means
to be emotionally intelligent.
Previous research has validated the importance of thinking about the
emotional intelligence of groups in terms of effective functioning. Vanessa
Druskat and colleagues (e.g., Druskat & Kayes, 1999; Druskat & Wolff, 2001) have
investigated “team EI” in a variety of contexts. They found that many of the
elements of effective emotional functioning in teams came from norms that team
members developed with each other, rather than from the intelligence of the
particular individuals. That is, team emotional intelligence was often a matter of
effective interpersonal behaviors, rather than unchangeable traits. The “whole”
was more than just an average of the parts, because teams tend to take on their
own unique character. Teams acted in the most emotionally intelligent manner
when they had mutual trust among members, a sense of group identity, and a
sense of group efficacy. Note that these norms do not focus on soft areas such as
being happy and friendly, but rather they focus on the conditions for
communicating openly even under difficult circumstances. Although individuals
can contribute towards building or destroying the necessary factors, it is the
group as a whole that shapes norms. Druskat and Wolff did find that individuals
with high levels of emotional intelligence tended to be more effective at fostering
healthy norms for teamwork. However, once in place these norms took on a life
of their own, and no longer depended on the individual group members.
Another source of evidence for the importance of the examining team
EI—in terms of the interactions among teammates rather than the EI of
individuals in the team—comes from research on the linkage of moods among
colleagues. When one person in a team experiences an emotion or mood, that
person’s teammates are often influenced and can take on some of that emotion or
mood as well. Jennifer George’s (1990) study, reviewed earlier, found evidence
that colleagues tended to be consistent in describing the emotional tone of their
team, which provides evidence that emotional tone is an important part of team
culture, with implications for performance. Totterdell, Kellett, Teuchmann, &
Briener (1998) demonstrated that individuals are influenced by the emotional
tone of their teammates, and over time they tend to shift their own moods
towards those of their colleagues. In a study that used naturalistic observations
rather than surveys, Caroline Bartel and Richard Saavedra (2000) found further
evidence that members of workgroups generally converge to develop similar
moods. Team members tended to develop their similar moods through a process
known as “emotional contagion”—so that people who are nearby tend to “catch”
the moods of others. Sigal Barsade (2002) investigated more closely this
emotional contagion process. In her research, she found that the contagion of
positive emotion led to greater team effectiveness, in the form of greater
cooperation and performance, as well as lower levels of conflict. Thus, the ability
of team members to share positive mood with each other is a form of emotionally
intelligent behavior that promotes greater team effectiveness.
Thus, there appears to be strong evidence for the importance of
emotionally intelligent interactions among colleagues in predicting the success of
teams. Nalini Ambady and I examined this issue further in our own longitudinal
study of public service teams. We used an exercise that measured how accurate
colleagues were in mutually understanding each others’ emotional expressions.
This exercise had two parts. First, I conducted a one-on-one interview with each
individual joining the organization. We discussed previous occasions during
which they had felt strong emotions in a workplace or school setting, and in
which they wanted others to know how they felt. Each participant described a
separate incident each for anger, fear, amused embarrassment, happiness, and
sadness. I asked them to repeat what they had said during the incident and to
describe how they expressed themselves. If they had not said anything at the
time of the incident, I asked them what they wish they had said, or what they
might have said. Although this interview was a re-enactment, after participants
described the incidents and the words that they had used, the interviews took on
a strong emotional tone. I videotaped the interviews and edited them in order to
create brief 5-second video clips containing naturalistic samples of their
emotional expressions. These video clips used segments with words that did not
give away the emotion and did not violate the participant’s privacy.
In the second stage of this exercise, colleagues viewed these video clips
within one week of the team being formed. I created a separate cassette for each
team so that they could view the video clips from each of their new colleagues.
Colleagues made multiple-choice judgments regarding which emotion had been
expressed in each video clip. Several additional measures served to validate this
video clip exercise. At the end of this process, we had a measure of how well each
team’s members could understand their colleagues’ workplace-relevant
expressions of emotion. Note that this exercise did not merely measure the skill
of individuals on the team, because it tapped into their skill in understanding their
specific teammates—which we demonstrated was distinct from their general skill in
understanding other people’s emotions. This is because people express
themselves in a range of different styles, and it is easier to understand a style with
which we are more familiar.
We were surprised by the strength of this exercise in predicting team
effectiveness over the course of the year. In fact, the ability of team members to
understand each colleagues’ emotional expressions explained 40% of the variance
in team performance (with an adjusted R2 of 28%), which is rare for research on
psychological processes. However, these results also showed that sometimes less
is more: greater accuracy in understanding colleagues’ positive emotions
predicted better team performance, whereas greater accuracy with negative
emotions actually predicted worse team performance. Teams whose members
easily understood each other’s expressions of amusement and happiness reported
greater success in accomplishing their service goals, and greater interest in
working with each other again. By contrast, teams adept in understanding
colleagues’ anger, fear, and sadness reported lower evaluations of their team’s
work, less liking for each other, and less interest in working together again. What
I found while spending time with the groups that were very perceptive at
understanding each others’ negativity was that they were unable to translate this
sensitivity into productive use. These teams got into spirals of negative energy.
The results of this study do not necessarily argue that the mutual ability to
understand negative emotion is always unproductive. There are many situations
in which we need negative feedback among teammates in order to improve—and
in which failing to understand negative emotion would be a roadblock for
learning. In this study, the public service teams consisted of young adults largely
in their first full-time job, and they appeared not to have the skills to use negative
emotion productively. Rather than perceiving negative emotions in colleagues as
a warning sign to reevaluate the work they were doing, or to reflect on their style
of team interaction, they reacted defensively and escalated conflict. Overall, these
results emphasize the complex interaction among the various components of
emotional intelligence. In the absence of effective emotion regulation skills, it may
be better not to have strong emotional perception skills. A balance among skills is
important for emotionally intelligent behavior in teams. It is worthwhile to make
an effort to achieve this balance, in light of the promise of greater team
Implications and Future Work
It is an exciting time to be a researcher studying emotional intelligence.
However, it can also be a challenging time as well, because the research findings
often do not stretch far enough to make recommendations that are as firm and
unambiguous as managers and practitioners will ultimately need in order to
make productive use of this research.
The initial evidence is very promising, suggesting strongly that greater
emotional intelligence benefits work groups in organizational settings. The
current article emphasizes that there are two very different ways of thinking
about what it means for a team to be emotionally intelligent: first, by examining
the EI of individual members, and second, by examining the EI displayed in
interactions among team members. The two perspectives, summarized in Figure
1, complement each other by asking different questions about teams, and thus
provide different insights and opportunities for researchers and practitioners.
The first perspective—examining the EI of individual members—offers the
chance to make predictions about team performance before a team is formed. For
this reason, it is the only practical method that can be useful for choosing team
members. By contrast, you cannot examine the team EI displayed in interactions
among team members until the team is formed. Thus, the second perspective on
team EI would be prohibitively expensive for choosing team members because
team EI is more than the combination of its parts. And, in many cases, team
membership must be driven by specific needs for the functional backgrounds and
availability of individual members, and cannot be adjusted based on emotional
However, Vanessa Druskat and Steve Wolff’s work shows us that the
second perspective on team emotional intelligence should still be crucial at the
time of team formation: rather than using EI as a selection tool to choose team
members, managers can use it as a development tool to help foster emotionally
effective norms from the first meeting onwards. Creating the conditions for
teams to communicate openly can help to build trust, a group spirit, and a can-do
attitude. Thus, emotionally intelligent behavior can develop in teams, regardless
of the test scores achieved by individuals.
Both perspectives on the emotional intelligence of teams can be useful in
crafting interventions once a team has already formed. When a problem arises
that appears related to the interpersonal dynamics among colleagues, it is
worthwhile to ask each of the questions on the right side of Figure 1. Whether a
team generally has the emotional resources that it needs, whether the team has
anyone left behind in terms of emotional competencies, whether the team has
anyone with exceptional skill who could help to build a more effective
environment, and whether the team has members who speak the same
“emotional language” all offer a chance to pinpoint possibilities for team coaching
or altering team membership. It is also important to ask, encouraged by the
second perspective, whether a team uses emotion effectively in its work. If the
answer is no, then there is rich potential for intervention by a manager or
qualified coach to develop more effective norms for emotional behaviors.
The first step in using emotional intelligence as a tool for improving team
performance is to consider carefully the nature of the team’s goals and contexts.
Some teams work in environments that are more emotionally charged, sensitive,
or sophisticated than others—for example, a negotiating team that represents a
company to outside interests would most likely benefit more from emotional
intelligence than a manufacturing team internal to the company. Likewise, a
team that is responding to a set of system failures may benefit more than a team
installing new systems. Thus, it helps to decide when it is worthwhile to
intervene in the emotional intelligence of a team. It will not always be the
case—the research reviewed above should convince you that emotional
intelligence is valuable, but complicated, in its impact on teams.
It is important for scholars to accumulate more evidence for the connection
between the emotional intelligence of teams and effective performance. For
example, what are the consequences when emotional processes such as the
understanding of emotional expressions are interrupted, which can happen
during telecommuting, electronic communication, or in virtual teams? More
research would be particularly helpful in examining critically and scientifically
the results of strategies for intervening. We are only now at the stage where we
have documented the likely impact of emotional intelligence in workplace
settings, and we need to be careful and judicious with attempts to alter a team’s
emotional landscape. However, the current research suggests cause for
optimism, as there could be great benefits for teams that can harness effectively
the power of emotional intelligence.
Barsade, S. G. (2002). The ripple effect: Emotional contagion in groups.
Administrative Sciences Quarterly. In press.
Barsade, S. G., Ward, A. J., Turner, J. D., & Sonnenfeld, J. A. (2000). To your
heart's content: A model of affective diversity in top management teams.
Adminstrative Science Quarterly. 45, 802-836.
Barsade, S.G., & Gibson, D.E. (1998). Group emotion: A view from top to
bottom. In D.H. Gruenfeld, B. Mannix, & M. Neale (Eds.), Research on managing
groups and teams: Composition (Vol. 1, pp. 81-102). Greenwich, CT: JAI Press.
Bartel, C A., & Saavedra, R. (2000). The collective construction of work
group moods. Administrative Science Quarterly, 45, 197-231.
Barry, B., & Stewart, G. (1997). Composition, process, and performance in
self-managed groups: The role of personality. Journal of Applied Psychology, 82,
Bouchard, T. J., Jr. (1969). Personality, problem-solving procedure, and
performance in small groups. Journal of Applied Psychology, 53, 1-29.
Byrne, D. (1971). The Attraction Paradigm. New York, Academic Press.
Chan, D. (1998). Functional relations among constructs in the same content
domain at different levels of analysis: A typology of composition models. Journal
of Applied Psychology, 83, 234-246.
Ciarrochi, J. V., Chan, A. Y. C., & Caputi, P. (2000). A critical evaluation of
the emotional intelligence construct. Personality and Individual Differences, 28,
Davies, M., Stankov, L., & Roberts, R. D. (1998). Emotional intelligence: In
search of an elusive construct. Journal of Personality and Social Psychology, 75,
DePaulo, B. M. (1992). Nonverbal behavior and self-presentation.
Psychological Bulletin, 11, 203-243.
Druskat, V. U., & Kayes, D. C. (1999). The antecedents of team
competence: Toward a fine-grained model of self-managing team effectiveness,
Research on Managing Groups and Teams, 2, 201-231.
Druskat, V. U., & Wolff, S. B. (March, 2001). Building the Emotional
Intelligence of groups. Harvard Business Review, 81-90.
Ekman, P., & Friesen, W. V. (1969). Nonverbal leakage and clues to
deception. Psychiatry, 32, 88-106.
Elfenbein, H. A., & Ambady, N. (2002a). Accuracy in communicating emotion
in the workplace: A field research investigation. Manuscript submitted for
Elfenbein, H. A., & Ambady, N. (2002b). Affective diversity in emotion
recognition ability: A longitudinal study. Manuscript in preparation.
Elfenbein, H. A., & Ambady, N. (2002c). Predicting workplace outcomes
from the ability to eavesdrop on feelings. Journal of Applied Psychology 87, 963-971.
Elfenbein, H. A., Marsh, A., & Ambady, N. (2002). Emotional Intelligence
and the recognition of emotion from the face. In L. F. Barrett & P. Salovey (Eds.),
The Wisdom of Feelings: Processes Underlying Emotional Intelligence. In press.
Elfenbein, H. A., Polzer, J., & Woolley, A. W. (2002). Team process and team
success: A survey study of business plan competitions. Manuscript in preparation.
George, J. M. (1990). Personality, affect, and behavior in groups. Journal Of
Applied Psychology, 75, 107-116.
Haythorn, W. W. (1968). The composition of groups: A review of the
literature. Acta Psychologica, 28, 97-128.
Hoffman, L. R., & Maier, N. R. (1961). Quality and acceptance of problem
solutions by members of homogeneous and heterogeneous groups. Journal of
Abnormal and Social Psychology, 62, 401-407.
Kelly, J. R., & Barsade, S. G. (2001). Mood and emotions in small groups
and work teams. Organizational Behavior and Human Decision Processes, 86, 99-130.
Klein, K. J., Dansereau, R. G., & Hall, R. J. (1994). Levels issues in theory
development, data collection, and analysis. Academy of Management Review, 19,
Kristof-Brown, A., Barrick, M. R., & Steven, C. K. (2001). A Test of
Competing Models of Person-Team Fit on Extraversion. Paper presented at the
Annual Meeting of the Academy of Management, Washington, D.C.
Mayer, J. D., Caruso, D. R., & Salovey, P. (1999). Emotional intelligence
meets traditional standards for an intelligence. Intelligence, 27, 267-298.
Mayer, J. D., DiPaolo, M., & Salovey, P. (1990). Perceiving affective content
in ambiguous visual stimuli: A component of emotional intelligence. Journal of
Personality Assessment, 54, 772-781.
Neuman, G. A., & Wright, J. (1999). Team effectiveness: Beyond skills and
cognitive ability. Journal of Applied Psychology, 84, 376-389.
Nowicki, S., Jr., & Duke, M. P. (2001). Nonverbal receptivity: The
diagnostic analysis of nonverbal accuracy (DANVA). In. J. A. Hall, & F. J. Bernieri
(Eds.), Interpersonal sensitivity theory and measurement (pp. 183-198). Mahwah, NJ:
Roberts, R. D., Zeidner, M., & Matthews, G. (2001). Does emotional
intelligence meet traditional standards for an intelligence? Some new data and
conclusions. Emotion, 1, 196-231.
Rosenthal, R., Hall, J. A., DiMatteo, M. R., Rogers, P. L., & Archer, D.
(1979). Sensitivity to nonverbal communication: The PONS Test. Baltimore: Johns
Hopkins University Press.
Rousseau, D. M. (1985). Issues of level in organizational research:
Multi-level and cross-level perspectives. Research in Organizational Behavior, 7,
Rychlak, J. F. (1965). The similarity, compatibility, or incompatibility of
needs in interpersonal selection. Journal of Personality and Social Psychology, 2,
Steiner, I. D. (1972). Group process and productivity. New York: Academic
Sternberg, R. J. (1984). Toward a triarchic theory of human intelligence.
Behavioral and Brain Sciences, 7, 269-315.
Totterdell, P., Kellett, S., Teuchmann, K., & Briener, R. B. (1998). Evidence
of mood linkage in work groups. Journal of Personality and Social Psychology, 74,
Williams, K. Y., & O’Reilly, C. A. (1998). Demography and diversity in
organizations: A review of 40 years of research. In L. Cummings & B. Staw (Eds.),
Research in organizational behavior: Vol. 20. (pp. 77-140). Greenwich, CT: JAI Press.
Williams, W. M., & Sternberg, R. J. (1988). Group intelligence: Why some
groups are better than others. Intelligence, 12, 351-377.
Perspectives on Emotional Intelligence in Teams
I. EI of individuals in the team
Examining the individuals who make up the team
Team-level average EI
Does this team generally have the
emotional resources to be
Team-level minimum EI Does this team have anyone left
Team-level maximum EI
Does this team have a member who
could jumpstart emotional
Team-level diversity of EI
Does this team have members who
speak the same "emotional
II. “Team EI”
A team as more than the sum of its parts
Observational and self-report measures of the emotional
savvy in interactions among team members Does this team use emotion
effectively in its work?