How Diversity Defeats Groupthink

NeuroLeadership Institute INDUSTRY RESEARCH

The NLI

Business Case

How Diversity Defeats Groupthink Smarter thinking isn’t about you. It’s about your team. By Khalil Smith Head of Diversity & Inclusion NeuroLeadership Institute

Learn the history of groupthink Discover how diversity unlocks better decisions Implement inclusive culture in your organization

All content provided within is protected under copyright law and is for non-commercial educational use only. Other uses, including reproduction, selling or licensing copies, or posting to personal, corporate or third-party websites, are prohibited except where provided by law and by permission of NeuroLeadership Institute, Inc.

How Diversity Defeats Groupthink By Khalil Smith Head of Diversity and Inclusion, NeuroLeadership Institute

The term “groupthink” was first coined in 1971 by the psychologist Irving Janis. Writing in the pages of Psychology Today, Janis set out to identify a psychological mechanism that could explain the disastrous decision by John F. Kennedy and his advisers to authorize the Bay of Pigs Invasion of 1961—the bungled American assault on Cuba that ultimately came to be seen as one of the most embarrassing foreign policy decisions of all time. The attack plan Kennedy signed off on—along with Robert McNamara, Robert Kennedy, Arthur Schlesinger, Jr., Allen Dulles, and about 40 other advisers from a team Janis described as “one of the greatest arrays of intellectual talent in the history of the American Government”—went something like this. Around 1400 Cuban exiles, trained by the CIA and armed with American artillery, would storm the beachhead at Bahía de cochinos and march on Havana. Inspired by American heroism, the Cuban people would rise up against Fidel Castro, the encroaching tide of communism would be turned, and Kennedy would glory in a righteous victory against a despicable enemy.

Sounds like a masterful plan, right?

What happened instead was that resources that were meant to be made available were pulled at the last moment, and a mission that was meant to hide America’s involvement was revealed as 1400 exiles encountered a superior Cuban force of 20,000 soldiers, who easily captured or killed the smaller fighting force. The whole thing was over in three days, Castro became a hero to his people, and Kennedy was humiliated on the world stage as the Bay of Pigs paved the way for Cuban and Russian partnership and a deepening of the Cold War. “There were 50 or so of

us, presumably the most experienced and smartest people we could get,” Kennedy would later recall. “But five minutes after it began to fall in, we all looked at each other and asked, ‘How could we have been so stupid?’ To answer that question—the question of how a team of self-evidently brilliant people can pool their intellectual powers and still somehow arrive at such an unquestioningly catastrophic decision—Janis proposed the existence of a previously undiagnosed, unnamed, and unknown problem afflicting groups tasked with making decisions: a phenomenon he termed groupthink. When groups work together to make a decision, Janis posited, they suffer from a process problem that, although unnoticed by the members of the group, nevertheless distorts their perception of reality and leads to reckless, outrageous decisions like the Bay of Pigs Invasion. A half-century later, it’s hard to read the headlines and not surmise that whatever happened to JFK and his advisers might still be happening— not just in foreign policy but in the corporate world as well. Groupthink has been blamed for many of the most visible collapses in the history of business, from Lehman Brothers to Enron to Worldcom, and even national travesties like the subprime mortgage bubble leading up to the global financial crisis.

Why bad decisions happen to smart people Decades earlier, during World War II, Irving Janis himself had been drafted, and he spent his tour carrying out studies of military morale. After the war, when Janis joined the psychology faculty at Yale, he remained fascinated by the bonds people form during times of unthinkable stress.

During combat, he’d seen, soldiers came together as brothers, and Janis believed it was their loyalty to their tiny combat unit that kept them going through the onslaught of enemy tanks, bombs raining down from the air, and unending German propaganda. The observation seemed to apply not just to soldiers in combat, but to villagers preparing for air raids and miners who get trapped in underground caves for days on end. What allowed them to survive, Janis concluded, was cohesion— the sticking together of the group. But if group cohesion was so unambiguously good and universally desirable, Janis wondered, then what explained policy fiascoes like the Bay of Pigs Invasion? Up until that point, it had always been assumed that group cohesion was just as important in the boardroom as it was in the trenches—that promoting a congenial, task-oriented atmosphere among colleagues could only serve to facilitate group discussion. How could that be a bad thing?

This was the conventional wisdom of the day against which Janis ultimately took his stand. Surviving combat was different from making a sound decision, he argued—where group cohesion could actually be unproductive and possibly dangerous.

Algorithmic bias and the dangers of homogeneous teams It’s not always possible to know how decisions get made behind closed doors, but when teams composed of objectively smart people release inexplicably terrible products, groupthink is often to blame. Perhaps the most embarrassing headsmacker in recent years involves commercial facial recognition. In recent years, several companies have developed machine learning technology to identify faces in photographs. Unfortunately, studies show that these systems don’t recognize darkskinned faces as well as light-skinned ones—a blatant case of what researchers call “algorithmic bias.” That’s a problem now that facial recognition is used not just in consumer electronics, but also in law enforcement agencies like the FBI. In a study presented at a conference called Fairness, Accountability, and Transparency, MIT researcher Joy Buolamwini assessed commercial facial recognition systems from IBM, Microsoft, and a Chinese company called Face++. Her goal was to compare their accuracy in recognizing faces of different ethnicities.

When people come together in a group like a war cabinet or an executive board, Janis suggested, they all want to belong. They’re so delighted to be part of the club that on an unconscious level, they begin to prize their membership in the group more highly than they do the rigor and accuracy of their decisions. In this happy, clubby atmosphere, everyone wants to agree with one another. The unstated goal of any group discussion, he argued, is to get to quick and painless unanimity as quickly as possible. In an effort to minimize conflict and reach consensus, individuals stifle their inner doubts, silence dissenters, and try their best to reflect the leader’s opinion. It all may sound subtle and harmless, but the consequences can be devastating. The group races to consensus without properly vetting ideas, and the results can include unrealistic optimism, biased analysis, and a disregard for ethics. Groupthink was at least partially responsible not just for the Bay of Pigs Invasion, Janis believed, but other foreign policy fiascoes as well: the decision by American military commanders in 1941 to ignore warnings of a surprise attack by Japan at Pearl Harbor; the decision in 1950 by Harry Truman and his advisers to cross the 38th parallel into North Korea; and the decision by Lyndon B. Johnson and his advisers in the late 1960s to escalate American involvement in the Vietnam War.

To ensure a broad spectrum of skin tones, she tested the systems using 1270 photographs of parliamentarians from around the world. The results showed that the systems were better at classifying white faces than darker ones, and more accurate for men’s faces than for women’s. IBM’s system, the Watson Visual Recognition service, got white male faces wrong just 0.3 percent of the time, compared to 34.7 percent for black women.

After Janis’ study, the word groupthink entered the lexicon, and the concept became an everyday concept in political science, psychology, and management. In the decades since Janis first published his article, psychologists have blamed groupthink as a factor in the Watergate scandal, the Iran-Contra affair, both the Challenger and the Columbia space shuttle explosions, and the 2003 decision to invade Iraq with the expectation of finding weapons of mass destruction.

Groupthink in the business world

Today, groupthink is widely blamed not just for foreign policy fiascoes, but for bad decisions throughout the corporate world, from everyday decision making on teams and executive boards to global financial bubbles and large-scale corporate fraud. When the postmortem analysis gets done in the aftermath of corporate disasters, it often emerges that the leaders behind the decisions have fallen prey to groupthink, says Princeton economist Roland Bénabou. As Bénabou explains in a 2012 article in The Review of Economic Studies, the pattern is always the same: The pressures of group conformity result in levels of collective denial and willful ignorance so high that they distort individuals’ perception of reality. This denial is contagious, and as the group is lulled into a state of “mutually assured delusion,” decision makers become “colorblind in a sea of red flags.” Dissent halts, subordinates draw cues from the leader, and mounting warning signs are cast aside, until finally the bubble bursts and people are left to wonder how it happened.

Any group of people that makes decisions is potentially vulnerable to groupthink. So how can businesses inoculate themselves? What can

Buolamwini’s study went viral and IBM, to its credit, responded swiftly, retraining its system with a fresh dataset and improving its recognition rates tenfold in a matter of weeks. But if IBM can 10x its accuracy overnight just by using a less biased dataset, it’s hard not to wonder: When a company forgets to train its system on dark-skinned faces, could that be because of groupthink?

companies and organizations do to safeguard their decision-making processes? Today, one solution has become increasingly obvious: diversity. Whether it’s bringing new people into the room or inducing people to consider things in new ways, more views help. In the last 20 years, a wide body of research has accumulated showing that diverse teams consistently outperform homogeneous teams, especially on tasks that are creative, nonlinear, or complex. In 2015, for instance, McKinsey & Co. conducted a survey of 366 public companies. They found that those with more ethnic and racial diversity among managers produced superior financial returns. In fact, the top 25 percent most diverse companies generated 53 percent more return on equity than the least diverse companies. And a study in 2009 found that racial and gender diversity correlate with greater sales revenue, more customers, and higher profit. Research shows that diverse teams also make superior decisions. It’s not just that demographic diversity means cognitive diversity, though that’s part of it—hiring employees with different backgrounds does mean an influx of new ideas and fresh perspectives, which means the group is more inclined to consider alternative options. But there’s also a benefit to decision making that’s inherent to demographic diversity. As it turns out, the mere presence of someone perceived as an outsider changes the behavior of the group, improving the thinking of those in the majority even when divergent viewpoints aren’t even expressed.

In a study published in The Journal Of Personality and Social Psychology, researchers examined how the ethnicity of jury members affected the way they evaluated evidence in a mock trial about an African-American man charged with sexual assault. The study found that compared to homogeneous juries, ethnically diverse juries considered a broader range of information and reasoned more accurately than same-race teams. It wasn’t just that African-American jurors contributed a different perspective to the discussion. It was that when African-American jurors were present, white jurors actually processed information more carefully—deliberating longer, referencing more case facts, making fewer factual errors, and discussing a wider range of personal perspectives. In fact, when black jurors were in the room, white jurors were twice as likely to notice “missing” pieces of evidence: why was there no evidence from fingerprints? Why hadn’t an important child witness testified? The presence of diversity seemed to trigger a motivation to avoid prejudice, prompting white jurors to consider more information, remember more accurately, and ultimately arrive at a sounder conclusion.

In another series of experiments, researchers set up faux financial markets in Texas and Singapore and asked participants to buy and sell stocks in either a same-race group or an ethnically diverse group. The studies found that same-race groups, with their implicit sense of

Three ways to defeat groupthink The best way to prevent these kinds of groupthink fiascos is to assemble a team with meaningful diversity. But diversity on its own isn’t enough—you also need to promote an environment of inclusion. After all, diverse perspectives don’t help anyone if people don’t feel empowered to voice them. To that end, here are three inclusive strategies to make sure all perspectives are heard.

Encourage dissenting opinions

The problem with group decisions is that even when team members have doubts, they often softpedal their objections or even silence themselves entirely for fear of being seen as an annoying buzzkill or a non-team player. The solution is to change the group’s social norms, creating a social context in which opposing viewpoints are welcomed and rewarded. To do that, leaders need to demonstrate that they truly want to hear opinions different from their own and emphasize that the goal of discussion is not group harmony, but sound decisions.

shared identity, trusted each other more. But that trust led them to put too much faith in each other’s judgment, fueling a cycle of blind conformity in which people copied each other’s mistakes and ultimately creating a price bubble. Ethnically diverse groups were more skeptical of each other and encountered more friction and conflict. But since they scrutinized each other’s decisions more closely, they wound up being far more accurate in how they priced stocks. As the researchers put it in a New York Times op-ed, diversity increased critical thinking, contributed to error detection, and produced “cognitive friction that enhanced deliberation.” The business benefits of diversity are present with regard to gender. In a 2013 study published in Management Science, researchers partnered with an entrepreneurship program to assess how gender diversity would affect business performance. As part of their degree, graduate students were assigned to go out in teams of 10 to 12 and start a business. To assess the impact of gender diversity on the teams’ success, the researchers varied the gender composition of the teams. Some teams had just one woman, others were all women, and the rest were somewhere in between. The study found that gender-diverse teams had more sales, higher profits, and higher earnings per share than male-dominated teams. But gender-mixed teams also outperformed female-dominated teams. It wasn’t that men or women were better at business. Rather, performance peaked when the proportion of women hovered around 55 percent. Analysis of the data found that diverse teams performed better not because men and women had skills and knowledge that complemented each other, but because gender-balanced teams monitored each other more closely and because the amount of learning in the group was more evenly distributed among team members.

Gather

opinions anonymously

The Delphi Method, originally developed for military and diplomatic analysis of cold war scenarios, is a decision making process in which team members submit their thoughts anonymously. This confers two critical benefits. First, since members’ submissions are private and anonymous, it removes social pressure to conform. Second, submissions are read out by a facilitator, ensuring that equal weight is given to all voices—not just the loudest or most vocal.

Assign a devil’s advocate

Diversity’s paradoxical power

When one considers the sweeping devastation wrought by groupthink over the last 100 years, and the clear benefits diversity has for decision making, it may seem surprising that corporate America didn’t rush to embrace diversity decades ago. There are countless reasons why businesses have failed in their attempts to achieve true diversity—overt discrimination, unconscious bias, the tendency to hire for culture fit, the temptation to value “alignment” over dissent, and of course the fact that the main symptom of groupthink is that you’re not aware you’re doing it. But there’s another subtler, frequently overlooked reason why diversity is not yet a universal standard throughout the business world.

In 2009, researchers ran a study in which they asked members of fraternities and sororities at Northwestern University to assemble in groups and play a murder mystery game. Each group began with three members, all drawn from the same fraternity or sorority. After five minutes, they were joined by a fourth person—either another member of the same Greek house or an interloper from a rival house. Each team then read police transcripts drawn from interviews of a detective

Another way to combat groupthink is to assign a devil’s advocate, a team members whose job is to question assumptions, demand evidence, crossexamine, and poke holes in arguments. Since the devil’s advocate is actually rewarded for rejecting the group’s dominant position, it’s a good way to combat social pressure to conform. A variation on the devil’s advocate is the Red Team—a subgroup whose mission is to criticize the

investigating a fictional murder. After weighing the evidence and discussing the clues, each team voted on which suspect they thought was the culprit. To win the game, the team had to correctly guess the identity of the killer—in other words, a problem-solving task with one right answer. Unsurprisingly, the study found that adding an outsider to the group increased its odds of arriving at the correct solution—in line with the earlier research on how out-group members prompt careful thinking. What was more surprising was that the team members assumed the opposite to be true. After submitting their verdict to the researchers, participants answered survey questions about their impressions of the group discussion. The study found that the more diverse groups—the teams that included the outsider—judged their interactions and discussions to be less effective than did homogeneous groups. Mixed teams also felt less confident than did the homogeneous teams that their judgments were correct. They were wrong: Their discussions were more likely to lead to the right answer, even though they didn’t feel as good about their interactions. That’s the insight Janis anticipated in his original analysis: that as dangerous as it is, consensus actually feels pleasurable. Human beings are wired to prefer information that comes to us easily, and we judge it to be more true. When you collaborate with people who are similar to you, the group forms a strong sense of shared social identity. With similarity comes comfort, camaraderie, and the illusion of progress. As NLI research scientist Valerie Purdie Greenaway puts it, “fluency is fun.”

The problem is that the pleasure of group cohesion puts pressure on individuals to just get along—to not rock the boat. People enjoy cohesion so much that they’re afraid to say anything to disturb it. When you’re in a group of similar people and spirits are high, you’re not motivated to turn to the person next to you and tell them their reasoning is flawed. Motivated to preserve consensus, you don’t criticize the ideas of others or challenge the opinions of the majority. Instead you often go with the flow. Perspectives get overlooked, your ability to think critically is crippled, and mistakes get made—sometimes catastrophic ones. Rather than being guided by what is right, you’re guided by what feels right—preserving group cohesion. That’s why diversity is so important in business. When you add an outsider to a homogeneous group, it upsets the balance, causing friction and tension. Paradoxically, this feels counterproductive. People don’t understand each other as immediately. There’s more conflict, the work feels harder. But that discomfort is the precise reason why the group doesn’t lapse into mindless conformity. Instead, the group assesses information more carefully. The result is superior performance on complex tasks, and smarter, more balanced decisions. In the end, diversity often feels worse—and that’s precisely why it defeats groupthink.

primary team’s plan. Red Teams can sometimes work better than lone devil’s advocates, since having multiple people on the contrarian team provides social proof of the validity of opposing viewpoints. Red Teams are used extensively in the military before combat missions, as well as in law firms, which often run internal mock trials in which one team of attorneys is tasked with helping the primary team prepare for trial by presenting opposing counsel’s case as devastatingly as possible.

What to Watch For When we think about increasing diversity, we usually think about bringing different people into the room. But diversity can also mean provoking the people already in the room to think differently. Here are a few common traps and what you can do to break free. ££ Bite your tongue when you’re in charge. As a leader, you naturally want to share what you know. But it’s important not to bias the discussion with the influence you wield. Next time you find yourself wanting to broadcast your opinion at the outset, remember to hold back until others have weighed in. ££ Solicit contrary perspectives. People naturally want to get along, but that can make dissent feel unwelcome. Next time you sense team members hesitating to speak their minds, remind them it’s OK to disagree. The goal isn’t harmony, but good decisions. But you need to rotate the role of contrarian, so that it becomes a habit that people can employ. ££ Amplify quiet voices. Bad decisions happen when team members keep their doubts and reservations to themselves. Next time you notice a discussion being dominated by a few vocal personalities, make a point of calling on those whose voices haven’t been heard. ££ Run the scenarios. One way to defeat conformity is to change your time horizon. Next time your team’s plans start feeling too rosy, try projecting yourself into the future, running through scenarios and thinking through what could go wrong. Shifting your perspective can inject an important dose of reality and help you see through misplaced optimism.

££ Switch it up. The more time you spend as a team, the closer you get. You learn each other’s quirks, you develop inside jokes, and you start using the same acronyms and shorthand. Camaraderie feels great, but it also has a downside: Team members begin to act like each other, talk each other, and most dangerously, think like each other. Next time that happens, make a point to shake things up. Kick people off the team, bring new people in, and shuffle roles. Change can be bittersweet, but it will help keep your team nimble and sharp.

Main References

Sidebar References

Hoogendoorn, S., Oosterbeek, H. & van Praag, M. (2013). The impact of gender diversity on the performance of business teams: Evidence from a field experiment. Management Science, 59(7), 1514–1528.

Buolamwini, J. & Gebru, T. (2018). Gender shades: Intersectional accuracy disparities in commercial gender classification. Proceedings of the 1st Conference on Fairness, Accountability and Transparency, in Proceedings of Machine Learning Research, 81:77-91.

Hart, P. (1991). Irving L. Janis’ victims of groupthink. Political Psychology, 12(2), 247. Janis, I. A. (1971). Groupthink. Psychology Today, 5(6), 43-46, 74-76. Janis, I. A. (1972). Victims of groupthink: A psychological study of foreign-policy decisions and fiascoes. Oxford, England: Houghton Mifflin. Janis, I. A. (1982). Groupthink: Psychological studies of policy decisions and fiascoes (2nd ed.). Boston: Houghton Mifflin. Levine, S. S., Apfelbaum, E. P., Bernard, M., Bartelt, V. L., Zajac, E. J., & Stark, D. (2014). Ethnic diversity deflates price bubbles. Proceedings of the National Academy of Sciences, 111(52), 18524-18529. Levine, S.S. & Stark, D. (2015, December 9). Diversity makes you brighter. The New York Times. Retrieved from https:// www.nytimes.com/2015/12/09/opinion/diversity-makes-youbrighter.html🔗 Phillips, K.W., Liljenquist, K.A. & Neale, M.A., 2008. Is the pain worth the gain? The advantages and liabilities of agreeing with socially distinct newcomers. Personality and Social Psychology Bulletin, 35(3), 336–350.

Sommers, S.R., 2006. On racial diversity and group decision making: Identifying multiple effects of racial composition on jury deliberations. Journal of Personality and Social Psychology, 90(4), 597–612.

Buolamwini, J. (2018, January 26). How does facial recognition software see skin color? NPR.org. Retrieved 21 April 2018, from https://www.npr.org/2018/01/26/580619086/joy-buolamwinihow-does-facial-recognition-software-see-skin-color🔗 Horwitz, J. (2018, February 13). If you’re a darker-skinned woman, this is how often facial-recognition software decides you’re a man. Quartz. Retrieved 21 April 2018, from https:// qz.com/1205604/if-youre-a-darker-skinned-woman-the-bestfacial-recognition-software-frequently-thinks-youre-a-man/ Lohr, S. (2018, February 9). Facial recognition is accurate, if you’re a white guy. The New York Times. Retrieved 21 April 2018, from https://www.nytimes.com/2018/02/09/technology/🔗 facial-recognition-race-artificial-intelligence.html Simonite, T. (2018, February 26). photo Algorithms ID White Men Fine—Black Women, Not So Much. WIRED. Retrieved 21 April 2018, from https://www.wired.com/story/photo-algorithms-idwhite-men-fineblack-women-not-so-much/🔗 Sunstein, C. R., & Hastie, R. (2015). Wiser: Getting beyond groupthink to make groups smarter. Boston: Harvard Business Press.