A few weeks ago I came across the following quote:
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.
Anita Woolley & Thomas W. Malone
It’s the key finding from a piece of research conducted by Anita Woolley (an assistant professor of organisational behaviour and theory at Carnegie Mellon University), and Thomas Malone (the Patrick J. McGovern Professor of Management at the MIT Sloan School of Management and the founding director of the MIT Centre for Collective Intelligence). The purpose of the study was to try to determine what makes teams effective. They published their findings in June 2011, on HBR.
Before I dig into the study itself, I think collective intelligence as a concept is worth explaining. Here’s a quote from Woolley:
The journey we are following now very much parallels the discovery process that was involved with individual intelligence.
Individual intelligence was originally based on the observation that certain individuals who perform well in some domains also perform well in other domains. And so intelligence was actually an inference, based on this observation of consistency of performance across domains.
And that’s very much how collective intelligence has emerged. It’s the observation that teams that do well in one domain also do well in other domains.
Now, even though we’ve understood individual intelligence for quite a long time, we still lack a full mentalistic account of all of what goes on inside one’s brain when you’re functioning intelligently. And likewise, with collective intelligence, we’re still trying to understand all of what comprises a collectively intelligent group.
Anita Woolley
Fascinating, right? But how do you go about understanding and measuring this? Professors Woolley and Malone, along with Christopher Chabris, Sandy Pentland, and Nada Hashmi, conducted their study as follows:
They gave subjects aged 18-60 standard intelligence tests, and assigned them randomly to teams. Each team was then asked to complete several tasks, including brainstorming, decision making, visual puzzles, and to solve one complex problem. Teams were then given intelligence scores based on their performance.
Essentially, based on these team (or collective) intelligence scores, Woolley and Malone could easily see which teams were most effective; however, figuring out why they were more effective proved much harder:
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 correlated, either.
Anita Woolley
Malone noted that before they did the research, they thought that perhaps collective intelligence would be just the average of all the individual IQs in a group. As such, they were surprised and intrigued to find that group intelligence had relatively little to do with individual intelligence.
So it seems if you want to create a really effective team (i.e. one with high collective intelligence), getting a bunch of very smart people (i.e. those with high IQs) into a room together, won’t necessarily mean you’ll see successful outcomes.
This, in and of itself, I didn’t find too surprising. Over time, I’ve observed lots of examples of very smart people getting together, but failing to accomplish much of anything. Marvellous debates, lots of intelligencing, but little to no valuable output.
But what I was surprised about was that factors like group satisfaction, group cohesion, and group motivation also didn’t correlate.
In fairness, I’m not entirely clear on how those factors were measured, and I may be misinterpreting this element of the study, but it sounds to me as though groups which weren’t working together particularly effectively, thought that they were.
By this, I mean, if you were to ask me to rate things like satisfaction, cohesion, and motivation of a group; I think that I would be unable to divorce those factors from how well or otherwise I felt we’d performed.
Of course those things are different – a group might be highly motivated, but ultimately achieve very little; or highly cohesive but ultimately achieve very little; but I can’t imagine, even in instances like that, that I’d rate my satisfaction with the group highly.
I say that, but of course it’s entirely possible that in fact, I would.
Maybe if I was having a lovely time working within a motivated, cohesive group, I’d feel satisfied, and just fail to notice that what we were doing wasn’t very good. Possibly that is what was going on there.
But I digress. Back to the findings.
Woolley was interviewed by HBR’s Berinato on the findings of this study, an excerpt from the interview is below:
SCOTT BERINATO:
Now, of course, the headline-grabbing finding here I want to briefly explore, was that you found a mostly linear relationship between the number of women in a group and its group intelligence. Can you explore what you think is behind this finding?
ANITA WOOLLEY:Sure, and I should mention, as we’ve mentioned in other places, this is not something we expected. And in fact, the first time we observed it, we didn’t pay much attention to it until we actually replicated it in a few different studies. So it wasn’t something we expected going in.
However, in exploring it a bit further, we are coming to understand a few things that underlie this.
First of all, in our own study, what we discovered is that it’s really explained by the social skills of the people involved. So on average, women tend to score higher on tests of social skills and social intelligence, and that was true in our sample.
So that seems partially explain at least what was going on in our groups.
It wasn’t just that having more women was better, it was that having more socially skilled or socially intelligent individuals in the group was better.
But as we look at this further and look at some other lines of research, we are coming to understand that having more women in a group actually seems to change group process in some fairly beneficial ways. So groups usually experience higher-quality collaboration as a function of at least having gender parity, and we also have some evidence that having more women in the group can change the behaviour of the men. So these suggest a few other places that we’re hoping to look into more in our ongoing studies, to try to understand these findings.
In the same interview, I also found this gem:
SCOTT BERINATO:
Of course the danger in such an, I would say, attention-grabbing finding is that it can drown out a lot of the other parts of the research or more nuanced parts of the findings. What else is interesting to you about some of the findings that are turning up in this group intelligence research that maybe won’t get as much play?
ANITA WOOLLEY:
Well, I think it’s been interesting to us, actually, that people weren’t more surprised by this notion that a group might be consistent in how it performs across different domains.
So we certainly know of sports teams that excel at sports, or orchestras that excel at producing music, but this notion that you can have a particular combination of people that will perform well in a relatively consistent manner across domains and potentially over time is really intriguing, and it has some important implications, certainly, for organisations.
So as I mentioned, being able to predict which combinations of people you can really trust in going into highly dynamic situations where a lot of money or even lives are at stake seems to have really important implications.
Also, over the past few decades, we’ve gotten better at identifying individuals at organisations that are highly valuable, that are huge contributors whom we should really try to retain.
But perhaps, if we can actually quantify the value of what a group can produce, through understanding what the collective intelligence of this particular group is, we can make better decisions about which groups to keep intact when we’re going through yet another reorganisation, so that we don’t destroy value without even realising it.
It seems to me that Woolley was proved right.
The “if a group includes more women its collective intelligence rises” headline played really well.
I think it started a lot of positive conversations, and I’m really happy about that.
Many companies now, are on the way towards accepting that more diverse teams lead to better outcomes. Many companies are actively trying to build more diverse teams as a result. This is good thing. There’s still more to be done, and we’re not nearly there yet, but I’m hopeful things will continue to move in the right direction.
But what’s troubling me is the other direction in which she was right. Woolley was surprised that more people weren’t interested in this:
…being able to predict which combinations of people you can really trust in going into highly dynamic situations where a lot of money or even lives are at stake seems to have really important implications.
Anita Woolley
We keep saying that we need to figure out how to build effective teams, but, we’re not, I don’t think, making tremendous progress on that front.
A snippet of Woolley’s quote bears repeating here:
…over the past few decades, we’ve gotten better at identifying individuals at organisations that are highly valuable, that are huge contributors whom we should really try to retain.
Anita Woolley
She said this in 2011. The problem is, I don’t think we’ve moved on. We’re still in the same place.
Today, when we evaluate performance, we still typically do so at an individual level – we do individual performance reviews, individual pay rises.
I feel like we’re still (almost exclusively) spending time identifying highly valuable individuals, not valuable teams.
Not intentionally perhaps. But we’re doing it nevertheless.
Now sometimes of course, what’s valuable about those individuals is indeed what they contribute to the team that they’re working in.
But sometimes it really isn’t.
Sometimes the importance of team contributions are downplayed, or even actively dismissed as unimportant. This can create dysfunctional or even toxic workplaces.
Possibly you think I’m exaggerating.
I’m just going to leave this tweet here. Perhaps you’ve already seen it, if not, by all means click through and read the thread, it’s an absolute horror show on just about every level.
I’m really glad I don’t work for that dude, but that’s not really the point.
In the course of putting together this post, I’ve come to suspect that some of those frameworks we’re using to evaluate performance, aren’t just contributing to the 10x problem, they are the problem.
Moreover, the 10x problem might be why we’re struggling to build better teams.
When we create performance evaluation frameworks what we’re actually doing is defining what our companies value in their employees and what (in terms of behaviour, performance, outcomes, etc) they want to reward.
Those frameworks created the 10x individual mythology (the notion that we should put a higher value on individual as opposed to team contributions), and if we’re not careful they will continue to perpetuate it.
It occurs to me that if we acknowledge that a team is indeed more valuable than the sum of its contributing members, shouldn’t we all be working harder at trying to figure out how to better measure and evaluate team performance, or, at the very least trying to better understand how individuals contribute to team dynamics?
And if we did, not only might we be able to actually build more effective teams, it’s possible we might also finally see the demise of the toxic 10x individual mythology which has been making so many, so miserable, for so long.