Malone shared his findings recently at the Techonomy13 conference in Tucson, Ariz., where tech industry elites were invited to spend three days considering the most important topics in our technologically advanced society. According to Malone, “It’s becoming increasingly important to think of businesses and organizations in terms of how intelligent, not just how productive or efficient, they are.”
There are many tests and standards for measuring individual intelligence. But, Malone says, until now there has been no way to measure the intelligence of a group. “If you give a bunch of individuals a wide range of tasks and look at how well they do, intelligence is the factor that the test measures. Nobody had ever asked if there is a similar factor for groups,” he says.
His research team at MIT set out to figure it out. Malone says he and his colleagues asked groups of between two and five people to perform tasks in the lab and applied to the entire group’s performance the same statistical techniques used to measure individual intelligence.
The surprising finding: individual intelligence is only moderately correlated with group intelligence. If it’s not smart people, then what is it that makes a group smart? “We found three significantly correlated factors,” says Malone.
One is the average social perceptiveness of the group members. The researchers measured social perceptiveness by administering a test called “reading the mind in eyes.” In that test, the study subject tries to guess what each person in a series of 36 photographs is feeling by looking only at their eyes. “When a bunch of people in the group are good at that, then group on average is more intelligent,” Malone says.
Also correlated to the level of intelligence of the group was the degree to which members participated equally in the discussion. “If one or two people in the group dominated, then on average the group was less collectively intelligent,” Malone says his research found.
Finally, the percentage of women in the group was a predictor of the group’s intelligence. “More women was correlated with more intelligence,” he says.
To be sure, this last result was largely explained statistically by the first: It was already known that, on average, women score higher than men do on the social perceptiveness test. So, Malone says, it could be that what you really need for a group to be intelligent is uniformly high scores on the social perceptiveness test. He says the study was not designed to consider that. But, he says, “as best we could tell from looking at the results, a group that is half men and half women seems to be the worst combination.” Instead, they found a linear correlation, indicating that up to and including a group all women seems to be correlated with higher group intelligence.
Interestingly, the findings hold up in electronic collaboration among a group as well as they do in verbal collaboration. In some tests, the groups came together online and could only communicate by text chat. “It turned out that the average social perceptiveness of group members was equally applied, even when they can’t see each other’s eyes at all,” Malone says. He believes this means that a high score in the ‘reading the mind in the eyes’ test must be correlated with broader range of social skills and social intelligence.
If you score poorly on the test for social perceptiveness, don’t despair. The New York Times reported recently that researchers have discovered that you can improve your score by reading literature by Chekhov or Alice Munro. Perhaps there’s hope for teams of men