by David Robson
We don’t yet know the exact cause of these gender differences. There may be a biological basis – testosterone has known effects on behaviour and higher levels make people more impulsive and dominant, for instance – but some differences in social sensitivity could be culturally learnt too.
Woolley told me that these findings have already changed opinions. ‘Some organisations have taken what we have found and just flipped it into hiring more women.’
Whether or not you would deliberately change the gender balance on the basis of these findings, hiring people of both sexes with greater social sensitivity is an obvious way to boost the collective intelligence of an organisation.
The very name – soft skills – that we attribute to social intelligence often implies that it is the weaker, secondary counterpart to other forms of intelligence, and the tests we use to explore interpersonal dynamics – such as the Myers-Briggs Type Inventory – are poor predictors of actual behaviour.13 If you are attempting to recruit a smart team, Woolley’s research strongly suggests that these social skills should be a primary concern, and in the same way that we measure cognitive ability using standardised tests, we should begin to use scientifically verified measures to assess this quality.
By showing that collective intelligence has such a weak correlation with IQ, Woolley’s tests begin to explain why some groups of intelligent people fail. But given the research on the individual intelligence trap, I was also interested in whether high performers are ever at an even higher risk of failing than teams of less average ability.
Intuitively, we might suspect that especially smart or powerful people will struggle to get along, due to the over-confidence or closed-mindedness that comes with their status, and that this might damage their overall performance. But are these intuitions justified?
Angus Hildreth at Cornell University can offer us some answers. His research was inspired by his own experiences at a global consulting firm, where he often oversaw meetings from some of the top executives. ‘They were really effective individuals who had got to where they were because they were good at what they did, but when they came together in these group contexts I was surprised at the dysfunction and the difficulties they faced,’ he told me during one of his frequent visits back home to London. ‘I was expecting this platonic ideal of leadership: you put all the best people in the room and obviously something good is going to happen. But there was an inability to make decisions. We were always running behind.’
Returning to study for a PhD in organisational behaviour at the University of California, Berkeley, he decided to probe the phenomenon further.
In one experiment, published in 2016, he gathered executives from a multinational healthcare company, assigned them to groups and asked them to imagine that they were recruiting a new chief financial officer from a selection of dummy candidates. The spread of power between the groups was not equal, however: some were composed of high-flying executives, who managed lots of people, while others were mostly made up of their subordinates. To ensure that he was not just seeing the effects of existing competition between the executives, he ensured that the executives within the groups had not worked together previously. ‘Otherwise there might be this past history, where someone had beaten someone else to a position.’
Despite their credentials and experience, the groups of high-flyers often failed to reach a consensus; 64 per cent of the high-power teams reached an impasse, compared with just 15 per cent of the low-power teams; that’s a four-fold difference in the groups’ effectiveness.14
One problem was ‘status conflict’ – the high-power members were less focused on the task itself, and more interested in asserting their authority in the group and determining who would become top dog. But the high-flying teams were also less likely to share information and integrate each other’s points of view, making it much harder to come to a successful compromise.
You could argue that you need greater confidence to get ahead in the first place; perhaps these people had always been a little more self-serving. But a further lab experiment, using students as participants, demonstrated that it takes surprisingly little effort to send people on that kind of ego trip.
The students’ first task was simple: they were separated into pairs and told to make a tower using building blocks. In each pair, one was told that they were a leader and the other a follower, ostensibly based on their answers to a questionnaire. Their success or failure in the task was unimportant; Hildreth’s goal was to prime some of the participants with a sense of power. In the next exercise, he rearranged the students into groups of three, composed either of all leaders or all followers, and set them some tests on creativity, such as inventing a new organisation and laying out its business plan.
Drunk on the tiny bit of power bestowed in the previous exercise, the former leaders tended to be less cooperative and found it harder to share information and agree on a solution, dragging the groups’ overall performance down. They were demonstrating exactly the kind of mutually sabotaging behaviours, in other words, that Woolley had found to be so destructive to a team’s collective intelligence.
Hildreth says the power struggles were plain to see as he observed the groups at work. ‘They were pretty cold interactions,’ he told me. ‘Reasonably often at least one student in the group withdrew, because the dynamics were so uncomfortable – or they just didn’t want to engage in the conversation because their ideas weren’t being heard. They thought, “I’m the one who makes the decisions, and my decisions are the best ones.” ’
Although Hildreth’s study only explored these dynamics in a single healthcare company, various field studies suggest these dynamics are ubiquitous. An analysis of Dutch telecommunications and financial institutions, for instance, examined behaviour in teams across the company’s hierarchies, finding that the higher up the company you go, the greater the level of conflict reported by the employees.
Crucially, this seemed to depend on the members’ own understanding of their positions in the pecking order. If the team, as a whole, agreed on their relative positions, they were more productive, since they avoided constant jockeying for authority.15 The worst groups were composed of high-status individuals who didn’t know their rank in the pecking order.
The most striking example of these powerplays – and the clearest evidence that too much talent can be counter-productive – comes from a study of ‘star’ equity analysts in Wall Street banks. Each year, Institutional Investor ranks the top analysts in each sector, offering them a kind of rock star status among their colleagues that can translate to millions of dollars of increased earnings; they are also regularly picked as media pundits. Needless to say, these people often flock together at the same prestigious firms, but that doesn’t always bring the rewards the company might have hoped.
Studying five years of data across the industry, Boris Groysberg of the Harvard Business School found that teams with more star players do indeed perform better, but only up to a certain point, after which the benefits of additional star talent tailed off. And with more than 45 per cent of the department filled with Institutional Investor’s picks, the research department actually becomes less effective.
The groups appeared to be particularly fragile when the stars’ areas of expertise happened to coincide, putting them in more direct competition with each other, and it was less of a factor when they fell into different sectors, and were therefore in less direct competition with each other. Then, the company could afford to recruit a few more stars – up to around 70 per cent of the workforce – before their rutting egos destroyed the team’s performance.16
Hildreth’s theory is based on the group interactions among the powerful. But besides disrupting communication and cooperation, status conflict can also interfere with the brain’s information processing ability. At least for the duration of the meeting, the individual members can themselves be a little bit more stupid as a result of their interactions.
The study, which took place
at Virginia Tech, gathered small groups of people and gave them each some abstract problems, while broadcasting their progress – relative to the other team members – on their computer interface. The feedback turned out to paralyse some of the candidates, lowering their scores compared to their performance on a previous test. Despite having started out with roughly equal IQs, the participants eventually separated into two distinct strata, with some people appearing to be particularly sensitive to the competition.17
The diminished brainpower was also evident in fMRI scans taken at the time of the test: it appeared to be associated with increased brain activity in the amygdala – an almond-shaped bundle of neurons, deep in the brain, associated with emotional processing – and reduced activity in prefrontal cortices behind the forehead, which are associated with problem solving.
The team concluded that we can’t separate our cognitive abilities from the social environment: all the time, our capacity to apply our brainpower will be influenced by our perceptions of those around us.18 Given these findings, it’s easy to see how the presence of a brilliant – but arrogant – team member may hurt both the collective and individual intelligence of his or her more sensitive colleagues, a double whammy that will reduce their productivity across the board.
As one of the researchers, Read Montague, puts it: ‘You may joke about how committee meetings make you feel brain dead, but our findings suggest that they may make you act brain dead as well.’
The sports field may seem a far cry from the boardroom, but we see exactly the same dynamics in many sports.
Consider the fate of the Miami Heat basketball team in the early 2010s. After signing up LeBron James, Chris Bosh and Dwayne Wade – the ‘Big Three’ – the team was overflowing with natural talent – but they ended the 2010?11 season ranked twenty-ninth out of thirty teams. It was only after Bosh and Wade fell out of the game with injuries that they would eventually win an NBA championship the following year. As the sportswriter Bill Simmons put it: ‘Less talent became more.’19
To find out if this is a common phenomenon, the social psychologist Adam Galinsky first examined the performance of football (soccer) teams in the 2010 World Cup in South Africa and the 2014 World Cup in Brazil. To determine the country’s ‘top talent’, they calculated how many of its squad were currently on the payroll of one of the top thirty highest-earning clubs listed in the Deloitte Football Money League (which includes Real Madrid, FC Barcelona and Manchester United). They then compared this value to the country’s ranking in the qualifying rounds.
Just as Groysberg had observed with the Wall Street analysts, Galinsky’s team found a ‘curvilinear’ relationship; a team benefited from having a few stars, but the balance seemed to tip at about 60 per cent, after which the team’s performance suffered.
The Dutch football team offered a perfect case in point. After disappointing results in the Euro 2012 championships, the coach, Louis van Gaal, reassembled the team – reducing the percentage of ‘top talent’ from 73 per cent to 43 per cent. It was an extraordinary move, but it seems that he had judged the dynamics correctly: as Galinsky and his co-authors point out in their paper, the Netherlands did not lose a single game in the qualifying rounds of the 2014 World Cup.
To check the ‘too-much-talent’ effect in a new context, Galinsky then applied the same thinking to basketball rankings, looking at the ten NBA seasons from 2002 to 2012. To identify the star players, they used a measure of ‘estimated wins added’ – which uses the statistics of the game to calculate whether a team member was often a deciding factor in a match’s outcome. The ‘top talent’, Galinsky’s team decided, lay in the top third of these rankings – an admittedly arbitrary cut-off point, but one that is often used in many organisations to decide exceptional performance. Crucially, many of the players within their ranking coincided with the players selected for NBA’s own All-Star tournament, suggesting it was a valid measure of top talent.
Once again, the researchers calculated the proportion of star players within each club, and compared it to the team’s overall wins within each season. The pattern was almost identical to the results from the football World Cup:
In one final experiment, the team examined data from Major League Baseball, a sport that does not require so much coordination between players. Here, they found no evidence of the too-much-talent effect, which supports the idea that status is only harmful when we need to cooperate and bring out the best in each other.20 For sports, like baseball, that are less interdependent than basketball or football, it pays to buy all the top talent you can afford.
If we look back at Iceland’s unexpected victory against England in the Euro 2016 football championships, it’s clear that its success stems from many different factors. The country had spent years investing in improving its training programmes, and the team was in excellent hands under Swedish coach Lars Lagerbac and his assistant Heimir Hallgrímsson. The quality of the individual players was undoubtedly better than it ever had been. But although many worked for international football clubs, just one of them at the time (Gylfi Sigurðsson) had a contract in one of the top-thirty clubs in Deloitte’s Football Money League. They hadn’t yet achieved the kind of international status that could be so destructive.
England, in contrast, had pulled twenty-one of its twenty-three players from these super-rich teams, meaning they accounted for more than 90 per cent of the squad, far above the optimum threshold. In fact, according to my own calculations, not one of the teams that succeeded in passing through to the quarter-finals had as many star players (the closest was Germany, who had 74 per cent). England’s defeat against Iceland at the Allianz Riviera Stadium in Nice is an almost perfect fit with Galinsky’s model.
Although they may not have been aware of Galinsky’s scientific work, football pundits noted the disastrous team dynamics at the time of the tournament. ‘England, for all their individual talents, lacked so much as one’, sports writer Ian Herbert wrote in the Independent after Iceland’s win. ‘The reason why the nation struggles to feel empathy or connections with many of these players is the ego. Too famous, too important, too rich, too high and mighty to discover the pace and the fight and the new dimensions to put it on when against one of Europe’s most diminutive football nations. That is this England.’21 The eventual champions, Portugal, incidentally took only four of their players from the elite clubs in the Deloitte Money League. The team may have had Cristiano Ronaldo – arguably the biggest star in the game – but it had not exceeded Galinsky’s threshold.
The ‘Miracle on Ice’ during the 1980 Winter Olympics at Lake Placid, New York, shows exactly the same pattern. The Soviet team had an unbroken record, returning home with gold in each of the past four Games. Of twenty-nine matches, they had won twenty-seven. Eight of their players had competed in at least some of those Olympic Games and also played for high-profile teams in their home country. The US team, in contrast, were a bunch of college kids, with an average age of twenty-one – making them the youngest team of the whole tournament – and little international experience.
The coach, Herb Brooks, was under no illusions that it was a case of ‘David against Goliath’. Yet David triumphed; America beat the Soviet Union 4?3, and they entered the second medal round with Finland. They walked away with the gold medal.
And you don’t need to be an international superstar for this kind of dynamic to apply to you and your colleagues; as Hildreth found with his university experiments, elite performance depends in part on your perception of your talents relative to those around you.
Anita Williams Woolley, who first alerted me to Galinsky’s research, even saw it in her sons’ amateur soccer team. ‘They had a very good team last year and won the state cup,’ she told me. ‘And then they attracted all of these really good players from other clubs and it ruined their dynamic. They’ve now lost five games this year.’
With this new understanding of collective intelligence and the too-much-talent effect, we are now very close to b
eing able to discover some simple strategies to improve any team’s performance. Before we do so, however, we need to explore the role of the leader in more detail – and the best case-study emerges from a tragedy on Everest’s slopes.
On 9 May 1996, at an altitude of 26,000 feet, two expeditions were ready to depart from Camp IV on the South Col route in Nepal. Rob Hall, a thirty-five-year-old New Zealander, led the team from Adventure Consultants. They were joined by a group from Mountain Madness, led by forty-year-old Scott Fischer from Michigan. Each team included the leader, two additional guides, eight clients, and numerous Sherpas.
Their expertise was unquestionable. Hall had already reached the summit four times previously, successfully guiding thirty-nine clients in the process; he was known to be meticulous with his organisation. Fischer had only mastered Everest once before, but he had earned his stripes on many of the world’s most challenging peaks. And he was confident of his methods. ‘We’ve got the Big E figured out . . . we’ve built a yellow brick road to the summit,’ one of the surviving members, Jon Krakauer, recalled Fischer saying.
Although they came from different companies, Hall and Fischer had decided to work together on the final bid – but they were soon beset with delays and difficulties. One of their Sherpas had failed to install a ‘fixed line’ to guide their ascent – adding an hour to their journey – and a bottleneck of climbers began to amass further down the slope as they waited for the line to be installed. By early afternoon, it became clear that many would not be able to reach the summit and return by the time darkness fell. A few decided to turn back, but the majority – including Hall and Fischer – pressed on.
