Thomas Malone on Building Smarter Teams
CCI’s most provocative finding so far is that, by and large, the higher the proportion of women on a team, the more likely it is to exhibit collective intelligence (and thus achieve its goals).
Thu May 26 03:41:45 2016 - permalink -
The critical factor appears to be social perception. Women are, on average, more perceptive than men about their colleagues. Social perceptiveness is a kind of social intelligence; it’s the ability to discern what someone is thinking, either by looking at their facial expression or through some other means of human observation. When it comes to the effectiveness of groups, we are what we see in each other. And if this kind of acumen can be learned, Malone’s research suggests that the performance of teams (and companies) can be dramatically improved.
S+B: Are you more likely to have mediocre people become a smart group, or are you more likely to have smart people become a mediocre group?
MALONE: Statistically, either could happen. Of course, we all know from our own experience that you can have very ineffective groups made up of very smart people. Now we have a precise, scientific demonstration of that.
But we did find three additional factors that were significantly correlated with the group’s collective intelligence. The first was the average social perceptiveness of the group members, the second had to do with the equality of contribution, and the third was the ratio of men to women in the group.
The first was the most obvious: the intelligence of the individual team members. We had expected that the group intelligence would correlate with the average or max-imum intelligence of individual group members. But we were surprised to find that the correlation was not very strong
What we really wanted to know was whether there is an equivalent of the g factor for groups. As far as we could tell, no one had ever asked this question before.
So to answer the question ourselves, we brought about 700 people into our laboratories, in groups that ranged from two to five people each. We gave each group a set of tasks to perform together, ranging from brainstorming uses for a brick, to solving IQ test problems as a group, to planning a shopping trip with a number of constraints, to typing long text passages into Google Docs. Each group spent about three hours working together on these tasks.
The second factor was the equality of contribution: the degree to which the group members participated evenly. When one or two people dominated the conversation, the group on average was less intelligent. Here again was a precise confirmation of what many people have perceived in their own team meetings.
In our results, this third factor was largely explained, at least statistically, by the first result. It was known, before our work, that women on average score higher than men on the test of social perceptiveness. So one interpretation of our results is that what you really need for a group to be intelligent is to have lots of people in the group who are high on this measure of social perceptiveness, regardless of whether the people are men or women.
Notice that this is not a standard diversity result. A standard diversity result would have been that the best-performing groups would have about the same number of men and women.
But in our results so far, the groups with half men and half women had some of the lowest scores. And it appears as if the highest scores go to groups composed mostly of women, with just a few men.
when put into a regression at the same time, the only one that is statistically significant is the first one, social perceptiveness. So, and this is somewhat speculative, one might conclude that the most important factor in collective intelligence is having groups where people are good at perceiving one another’s emotions accurately, or, more generally, where they have high social intelligence.
S+B: Is this a learnable or cultivatable skill?
MALONE: Excellent question, and we don’t know for sure. This quality in people appears to have some genetic component. It may also be influenced by hormones; it’s been negatively correlated with high levels of testosterone. Those are reasons to believe it’s not very changeable.
But there are other reasons to believe it might be possible to affect it.
Having good theory of mind skills is not necessarily the same as having empathy. Of course, they’re related. You couldn’t be empathetic without some theory of mind skill, because you wouldn’t even be aware of other people’s feelings. But you could accurately perceive what other people were feeling and thinking, while not caring about them. If you didn’t have any actual sympathy for them, you could use that accurate perception to manipulate or take advantage of them.
For instance, there are now ways of designing nonhierarchical organizations, like crowd-based organizations, that have the potential to be even more intelligent than the best-designed hierarchies.
With groups, however, it seems quite possible that we could change their collective intelligence.
For instance, as a group gets larger, the way you organize the group can have a major effect on its collective intelligence.
One interesting possibility is that with the right kinds of digital electronic collaboration tools, we could greatly increase the size up to which a group can continue to increase its intelligence by adding members. Right now, the optimal size is probably somewhere between five and 10, but with the right collaboration tools, you could imagine having a group that kept getting more intelligent, up to 50, 100, or even 500 or 5,000 people. That’s one of the most intriguing long-term research questions we’re starting to work on. Now that we have a way of measuring the intelligence of a group, we can use that to find ways to allow the group to scale to a much larger size without being overcome by the “process losses” that inhibit the performance of large groups today.