Defend Your Research
HBR: But gender does play a role?
Malone: It’s a preliminary nding—and
not a conventional one. The standard
argument is that diversity is good and you
should have both men and women in a
group. But so far, the data show, the more
women, the better.
Woolley: We have early evidence that per-
formance may atten out at the extreme
end—that there should be a little gender
diversity rather than all women.
You realize you’re saying that groups of
women are smarter than groups of men.
Woolley: Yes. And you can tell I’m hesitat-
ing a little. It’s not that I don’t trust the data.
I do. It’s just that part of that nding can be
The ﬁnding: There’s little correlation between a group’s collective
intelligence and the IQs of its individual members. But if a group
includes more women, its collective intelligence rises.
The research: Professors Woolley and Malone, along with Chris-
topher Chabris, Sandy Pentland, and Nada Hashmi, gave subjects
aged 18 to 60 standard intelligence tests and assigned them
randomly to teams. Each team was asked to complete several
tasks—including brainstorming, decision making, and visual
puzzles—and to solve one complex problem. Teams were given
intelligence scores based on their performance. Though the
teams that had members with higher IQs didn’t earn much higher
scores, those that had more women did.
The challenge: Are brainy people overrated? Are women the true
key to success? Professors Woolley and Malone, defend your
is the Patrick J.
of Management at
the MIT Sloan School
and the founding
director of the MIT
Center for Collective
explained by dierences in social sensitiv-
ity, which we found is also important to
group performance. Many studies have
shown that women tend to score higher on
tests of social sensitivity than men do. So
what is really important is to have people
who are high in social sensitivity, whether
they are men or women.
So you didn’t see a negative correlation
with individual IQs—just a very weak
positive correlation. In theory the 10
smartest people could still make a great
Woolley: In theory, yes, the 10 smartest
people could make the smartest group,
but it wouldn’t be just because they were
the most intelligent individuals. What do
you hear about great groups? Not that the
members are all really smart but that they
listen to each other. They share criticism
constructively. They have open minds.
They’re not autocratic. And in our study
we saw pretty clearly that groups that had
smart people dominating the conversation
were not very intelligent groups.
Can teams be too group oriented?
Everyone is so socially sensitive that
there’s no leader?
Woolley: Anecdotally, we know that
groups can become too internally focused.
Our ongoing research suggests that teams
need a moderate level of cognitive diver-
sity for eectiveness. Extremely homoge-
neous or extremely diverse groups aren’t
In some ways, your ﬁndings seem
blindingly obvious: that teams are
more than just a collection of the best
cmu.edu) is an
theory at Carnegie
Woolley: We’ve replicated the nd-
ings twice now. Many of the factors you
might think would be predictive of group
performance were not. Things like group
satisfaction, group cohesion, group
motivation—none were correlated with
collective intelligence. And, of course,
individual intelligence wasn’t highly cor-
Malone: Before we did the research, we
were afraid that collective intelligence
would be just the average of all the
individual IQs in a group. So we were
surprised but intrigued to nd that group
intelligence had relatively little to do with
What Makes a
2 Harvard Business Review June 2011
HBR puts some surprising ﬁndings to the test
Malone: Sure. This is well-known in sports.
Our study shows it with intellectual tasks.
We realized that intelligence tests are a way
to predict individuals’ performance on a
range of tasks, but no one had thought of
using the same approach to predict group
Woolley: There was a step change in
psychology once the eld had an empirical
method of measuring individual intelli-
gence through IQ tests. We’re hopeful that
this work can create a similar seismic shift
in how we study groups.
Can we design teams to perform better?
Malone: We hope to look at that in the fu-
ture. Though you can change an individu-
al’s intelligence only so much, we think it’s
completely possible to markedly change
a group’s intelligence. You could increase
it by changing members or incentives for
collaboration, for instance.
Woolley: There is some evidence to sug-
gest that collective intelligence exists at the
organizational level, too. Some companies
that do well at scanning the environment
and setting targets also excel at manag-
ing internal operations and mentoring
employees—and have better nancial per-
formance. Consistent performance across
disparate areas of functioning suggests
an organizational collective intelligence,
which could be used to predict company
So this phenomenon could extend beyond
the small groups you studied?
Malone: Families, companies, and cities
all have collective intelligence. But as
face-to-face groups get bigger, they’re less
able to take advantage of their members.
That suggests size could diminish group
intelligence. But we suspect that technol-
ogy may allow a group to get smarter as it
goes from 10 people to 50 to 500 or even
5,000. Google’s harvesting of knowledge,
Wikipedia’s high-quality product with
almost no centralized control—these are
just the beginning. What we’re starting to
ask is, How can you increase the collective
intelligence of companies, or countries,
or the whole world?
HBR Reprint FD
The Female Factor
The chart plots the collective
intelligence scores of the 192
teams in the study against the
percentage of women those
teams contained. The red bars
indicate the range of scores in
the group of teams at each level,
and the blue circles, the average.
Teams with more women tended
to fall above the average; teams
with more men tended to fall
The Revival of Smart
16425_HBR_1third_vert.indd 3 12/3/10 1:16 PM
FOR ARTICLE REPRINTS CALL OR , OR VISIT HBR.ORG
Many factors you
might think would
be predictive of
were not. Group
intelligence had little
to do with individual