What makes teams effective? People have been trying to find out for centuries, but researchers looking into that question recently discovered something new. According to psychologists at MIT, Carnegie Mellon, and Union College, groups take on a character of their own that’s distinct from the individual natures of their members. In other words, they become living things. Which raises a second question: If groups have an intelligence all their own, how can we measure it?
The idea is especially intriguing because individual general intelligence as measured by an IQ test is a concept that wouldn’t necessarily even have to exist. An IQ test isn’t just one test; it’s 10 subtests that require the test taker to perform widely different tasks. You have to demonstrate the extent of your vocabulary and do arithmetic and solve visual puzzles and even demonstrate hand-eye co-ordination, among other things.
There wouldn’t seem to be any reason all these abilities should be related. A large vocabulary and good hand-eye co-ordination don’t appear to call on the same skills, for example. You could even suppose that performance on the subtests might vary in opposite directions; someone could do well on the tests of abstract verbal reasoning and vocabulary by reading all the time, thus having less time to work on arithmetic.
Yet, rather amazingly, people’s scores on the subtests are correlated. If you do well on one, you’ll probably do well on all of them. The pioneers of modern psychology realized there was some common factor that influenced a person’s performance on all the different subtests. They could have called that factor any number of things, but they called it general intelligence.
Might the same phenomenon occur in groups? Could there be some common factor that affects a group’s performance on a wide range of disparate tasks? If so, it would be extremely valuable to know about because the measure of that common factor in individuals, IQ, turns out to be surprisingly powerful.
IQ has long been controversial, and it certainly doesn’t measure everything we want to know about a person, but its predictive power is solidly established. As the researchers observed, it “is a reliable predictor of a very wide range of important life outcomes over a long span of time, including grades in school, success in many occupations, and even life expectancy.” What if something similar existed for groups—a factor that measures the general effectiveness of a board of directors or a sales team or a project group, and that also predicts performance beyond the tested domains?
That’s what the researchers went looking for, and they found it. They measured almost 200 groups performing many widely varying tasks that groups in real life may have to do, such as brainstorming, making collective moral judgments, and negotiating over limited resources, for example. Just as with individuals, a group’s effectiveness on these separate tasks was positively correlated. The common factor exists.
What’s more, this factor, which researchers called “c” for “collective intelligence,” is a strong predictor of performance on tasks that weren’t used in measuring the factor to begin with. Some of the groups, after being tested, were asked to do a simple group task, competing against a computer at checkers. Others had to perform a much harder task, an architectural design task modeled on a complex research and development problem. In both cases, c was a strong predictor of a group’s performance. It does exactly what we would wish for.
Then the researchers asked a crucial question: Does c tell us anything we couldn’t figure out just by measuring the IQs of a group’s individual members? So they measured all those IQs and found that the average IQ of a group’s members was worth little in predicting the group’s performance. The IQ of the smartest member was worthless. That is, even the most effective group doesn’t need the smartest person. The winning Ryder Cup team doesn’t need the greatest golfer.
What the group has, the researchers said, “is a property of the group itself, not just the individuals in it.”
The researchers then asked the most important question: If individual intelligence doesn’t explain a group’s effectiveness, what does? They investigated some popular candidates: group cohesion, motivation, and satisfaction. None of them meant a thing.
The stability of the team and its size mattered only a little.
How about such fine-sounding concepts as having a clear, challenging, meaningful vision, and specifying well-defined roles and responsibilities, and giving team members appropriate rewards, recognition, and resources? They were unimportant.
What meant a lot, however, was social sensitivity. Group members each took a widely used test called “Reading The Mind In The Eyes,” which requires you to choose a word that best describes people’s thoughts or feelings based only on photos of their eye region. If you wanted to predict a group’s effectiveness, the best thing you could do was look at the members’ average score on that test. Conversational turn-taking also made a big difference; groups dominated by a few talkers were less effective than those in which members took more equal turns.
Another way to predict the most effective group was much simpler: Just count the number of women. That’s because in this research and much other research, women perform a great deal better than men in measures of social sensitivity. That finding is just one of several reasons that women seem especially strongly positioned to excel in the emerging economy.
“Collective intelligence” seems too narrow a term for what the researchers detected; it’s more than intelligence as we usually think of it. Because it’s based in large part on tasks that groups really perform, it might better be called group effectiveness. And the keys to it—to a group’s ability to come up with ideas or plan an activity or reach a moral judgment or solve a hard problem—are the subtlest elements of human interaction, sensing the proper meaning of a furrowed brow or noticing someone’s silent wish to contribute to a conversation.
Group performance remains an increasingly valuable human activity even as computers learn to read facial nuance and otherwise detect feelings. The simple reason is that we, not computers, decide what our problems are and ultimately choose the best solutions, and groups can be far better at choosing solutions than can any individual.
“Can” is the operative word. Groups aren’t always better; our own life experience (and plenty of research) shows that some groups are utterly dysfunctional and achieve nothing. But when they work, they’re superior. We’ve seen also that, as the world grows more complex, more tasks are being handled by groups.
As technology takes over more tasks, choosing our problems and how to solve them will still be among the jobs reserved for people, not machines. Understanding how to do those jobs best, which in many cases means in groups, is therefore vital to our success.
This article is adapted from Humans Are Underrated, with permission of Portfolio, an imprint of Penguin Publishing Group, a division of Penguin Random House, LLC. © Geoff Colvin, 2015