New Study: Motivation Doesn’t Improve Group Success

Many theories for the evolution of religion draw on the apparent survival advantage of human social cohesion and organization—groupwork, if you will. Whether you buy any of these theories or not, it’s hard to disagree with the assertion that religion is a profoundly social enterprise.

Is there something inherently special about groups and how they think—something more than the sum of their individual members?

As an educator and observer of learners, it makes intuitive sense to me that human groups have what scientists call ‘collective intelligence’. As a scientist, it makes sense to me that humans’ biological evolution mirrors our social evolution—and vice versa.

Psychologists can measure a single intelligence factor for an individual person, which they can fairly quickly determine and then use to predict how well that individual will do on a diversity of tasks. The most striking characteristic of this intelligence factor (first realized in 1904) is that it predicts performance even if the tasks involve entirely unrelated approaches or concepts and regardless of how the tasks are assigned or administered. This is a bit counterintuitive. Why would being good or bad at solving Sunday’s crossword predict how well you’d do at negotiating over limited resources? But this kind of individual intelligence factor testing has been replicated many times over the last decades. The intelligence factor is a predictor of ‘important life outcomes’, career success, and ‘even life expectancy’.

Now, even more surprising is a new study in the journal Science demonstrating that just such a factor exists for collective intelligence in the performance of human groups. Woolley et al. show the same intriguing correlation between a group’s ability to perform one type of task and entirely unrelated ones—a group good at one is more likely good at the other.

This is provocative stuff. Not only were groups (from 2-5 people) which were measurably good at solving puzzles also measurably good at pretty profound tasks like making collective moral judgments, but the collective intelligence factor was a ‘much better predictor of group performance… than the average or maximum individual intelligence’ in that group. There’s something new, some emergent property, about being in a group—if the group works.

That’s a big if. We’ve all been in groups—whether religion-based or otherwise—that drive us nuts. What makes for a group that works, a group with high collective intelligence? Woolley and friends find that some of the kinds of things you’d expect to improve collective intelligence don’t (including the motivation and cohesion of the group), while others you might expect do, including the average social sensitivity of group members, better shared engagement of all members (groups with conversation-dominators didn’t do as well), and groups with more women (not surprising, since women tend to score higher in social sensitivity).

We worry a lot about individuals in America, but society—our families, jobs, labs, companies, congregations—function (or not) as groups. This new research gives us at least two hors d’oeuvres of insight into human behavior, before enjoying what is a potential feast of ways to improve the way we do things: (1) collective intelligence is an emergent and apparently powerful property of interacting humans, and (2) it’s much easier to change groups to adjust for the factors that predict higher collective intelligence than it is to change single individuals to improve their intelligence. Think about it with some friends.