Few things attract more attention in the business world than new ways of making groups work well. As any fool knows, groups are a pain. They argue, dither, drift off course, waste time and resources, and produce loads of rubbish. Worse, all those participants draw salaries, so treasure is wasted. Surely, bosses think, any technique that promises to make groups productive will be better than what they have now: a dysfunctional collection of pointless individuals, wasting their time by rushing off in aimless directions? They reject the absurd notion that one person should do the job, and that the dysfunctional team should be disbanded. Leadership: that is what is required, they proclaim.
So, the poor hapless managers are sent off to Leadership courses, and come back with interesting theories which get nowhere, because the rest of the staff have not been sent off to Followership courses. Leaders always require followers. While everyone loves Leadership courses, being recommended for a Followership course would probably cause great offence. A pity. It is fatal to any enterprise when people can neither command nor obey.
While the supposed leaders have been away at an expensive hotel ,the remaining staff have sorted out the problem to their own satisfaction, sometimes with good effect, and most often by cobbling together a patch to protect their own interests. So, it is with grim satisfaction that one learns of Group Performance Enhancement Theory No 347, namely that people working in groups on complex problems:
“show a strong general-ability or IQ factor, with significant differences between groups on this factor. Surprisingly, group-IQ, or “collective intelligence” is not strongly correlated with the average or maximum individual intelligence of group members but is correlated with the average social sensitivity of group members, the equality in distribution of conversational turn-taking, and the proportion of females in the group.”
So, if I have understood this correctly, the individual IQs do not matter too much, so long as the group is socially sensitive, takes turns in speaking, and includes women. Personally, having heard this proposal I would have thought it unlikely. Bates and Gupta, however, are truer to the spirit of empiricism, and embody Carl Sagan’s injunction (my summary) that scientists should be kind to hypotheses and tough on proofs.
What is more, since the original study by Woolley et al in 2010 was cited over 700 times, this finding is likely to be the cornerstone of a myriad of training courses, as participants attempt to be sensitive, willing to wait their turn, and womanly. Bates and Gupta have bothered to find out if collective intelligence (group IQ) actually exists.
https://drive.google.com/open?id=0B3c4TxciNeJZLWFibW9jYnlFQVU They point out:
For some time, it has been known that work-groups whose team-members have higher IQ out-perform teams of less-able members (Devine & Philips, 2001). Against this background, Woolley et al. (2010) asked whether groups themselves exhibit a general-factor of intelligence, if this might be distinct from individual IQ, and, if so, what the origins of such a collective intelligence might be.
To assess group-IQ, subjects were allocated to small groups and performed tasks including brainstorming, matrix reasoning, moral reasoning, planning a shopping trip, and collaborative text editing. They did all this in 3 studies, so there is a lot of detail in the paper about the findings from their individual studies, and further work on the combined results, usually studies 2 and 3. Woolley et al. (2010) came to the conclusions described above, namely that it is the collective IQ which develops (due to sensitivity, turn taking in conversation, and women members) which is important, and not the IQ of the members of the group. Bates and Gupta sum up the findings of their replication thus:
What allows groups to behave intelligently? One suggestion is that groups exhibit a collective intelligence accounted for by number of women in the group, turn-taking and emotional empathizing, with group-IQ being only weakly-linked to individual IQ (Woolley, Chabris, Pentland, Hashmi, & Malone, 2010). Here we report tests of this model across three studies with 312 people. Contrary to prediction, individual IQ accounted for around 80% of group-IQ differences. Hypotheses that group-IQ increases with number of women in the group and with turn-taking were not supported. Reading the mind in the eyes (RME) performance was associated with individual IQ, and, in one study, with group-IQ factor scores. However, a well-fitting structural model combining data from studies 2 and 3 indicated that RME exerted no influence on the group-IQ latent factor (instead having a modest impact on a single group test). The experiments instead showed that higher individual IQ enhances group performance such that individual IQ determined 100% of latent group-IQ. Implications for future work on group-based achievement are examined.
After doing their 3 studies and re-analysing the results, they conclude: Smart groups are (simply) groups of smart people. By contrast, we found little to no evidence for two proposed causes of group-IQ: numbers of women in the group and turn-taking, and found evidence for a weak and specific impact of RME on one group task, but not on latent group-IQ.
Here is the relationship between the IQs of the individuals in the group, and the resultant group intelligence. If particular group IQs develop, then the group IQs will differ from the mere sum of the individual IQs. In fact, there is a close match.
Here are all their results summarised in a best fitting model.
Individual’s IQs lead to the group average IQ, which explains the performance the group achieves on all of the tasks. On one task, Missing letters, the Mind in the Eyes task makes a small additional contribution. They add:
The present findings cast important doubt on any policy-style conclusions regarding gender composition changes cast as raising cognitive-efficiency.
Their findings should not be interpreted as meaning that groups are useless. On the contrary, given that the management of clever people is so important for success, care must be taken to let the best thinkers concentrate on the hardest problems. Also, it implies that organizations should pay close attention to the intelligence of their staff members, and very probably to pay more attention to the opinions of their brighter workers.
And there the silly story would end, but there is a sting in the tail. Not only was the original paper cited 700 times, but it was cited without the benefit of a replication. All researchers may be tempted to do that, particularly when a study buttresses a position they like. However, since so many psychological studies fail to replicate, there is now general agreement that replications should be given as much attention as the original claims. So, how did reviewers respond to Bates and Guptas’s replication? With considerable reservations, it appears.
While every paper has to run the gauntlet of reviewer criticism, this one seems to have experienced unusual opposition. In their discussion section the authors reply to the objections raised by unnamed reviewers. A reviewer complained about lack of statistical power, but the main analysis of studies 2 and 3 had a power of 95%. This is a technical discussion, but I think the reviewer got it wrong.
A reviewer judged that the replication was not a replication. Bates and Gupta used the same IQ tests, the same test of empathy, and those tests of successful team work which had best shown the effects which Woolley et al. claimed in the original research. Looks like a replication to me.
Turn-taking was measured by a simpler technique in the replication, but turn-taking was not shown to be independently predictive of group IQ, rendering the point moot.
An anonymous reviewer suggested that (paraphrasing) there clearly must be an unidentified moderator which accounts for why individual IQ and collective intelligence correlated so strongly. Readers should evaluate this claim for themselves. It is far from clear to us that an unidentified moderator “must” exist.
Bates and Gupta were polite, but they could have responded “You show us why you think there has to be a moderator. Evidence, please”.
A friend of the authors, speaking to me in a dark car park on condition of reviewer-type anonymity, said:
The back story is that this paper went through 4 revisions, in which one reviewer every time demanded 10, 20, or even 38 new changes, none of which involved a single new analysis. They demanded that Bates and Gupta remove study 2, remove variables, include a statement that they had not done a replication, and conclude that this area is vigorous and needs more research. They claimed the work was sloppy, error-filled, and so under-powered no one should publish it. They suggested that no peer-reviewed journal would ever publish such awful work. So, if you think science is an efficient hunt for the truth…Think again.
The impression I get is that the reviewers were being unreasonable, and even obstructive.
You might like to look at this link: http://www.unz.com/jthompson/does-peer-review-give-too-much-power-to
All this further silliness aside, in what I consider to be one of the most important findings about team work, the authors identify a crucial result:
It is interesting also that groups did not perform better than individuals – a genuine group-IQ might be expected to enable problem solving to scale linearly (or better) with number of subjects. In group-IQ tasks, coordination costs appear to prevent group problem-solving from rising even to the level of a single individual’s ability. This implicates not only unsolved coordination problems, which are well-known barriers to scale (Simon, 1997) but also reiterates the finding that the individual problem-solver remains the critical reservoir of creativity and novel problem solution (Shockley, 1957).
So, if you want a problem solved, don’t form a team. Find the brightest person and let them work on it. Placing them in a team will, on average, reduce their productivity. My advice would be: never form a team if there is one person who can sort out the problem.
As regards team work and collective intelligence, another idea bites the dust, at least until a new hypothesis comes along, claiming you can boost team productivity by a training in…(insert something warm and friendly).
No teams were assembled to write this post.