THE CONTRARY ASSUMPTION
We have been wrong. We must change our lives, so that it will be possible to live by the contrary assumption that what is good for the world will be good for us.
—WENDELL BERRY, The Art of the Commonplace: The Agrarian Essays
The venue varies. I might be standing in North Dakota State University’s Memorial Union; the Great Hall of the US Department of Justice; Roper High School’s gymnasium in Birmingham, Michigan; or Bloomberg’s gleaming auditorium on Lexington Avenue in New York City. The composition of the audience differs even more. I might be speaking to college deans, high school students, NASA engineers, Justice Department lawyers, Wall Street titans, or Silicon Valley disrupters.1
The event will have been advertised as a talk on diversity. I am cast in the role of expert based on The Difference, a book I wrote a decade ago. The introductions run the gamut. A college administrator might read verbatim from my bio, a Fortune 500 CEO might all but ignore a prepared script and sing the praises of the University of Michigan, or a high school junior might nervously tick off bullet points from a notecard held in her trembling hand.
The next thirty seconds always play out the same. Polite applause (who is this guy again?). I walk on a stage. I shake a hand or accept a hug. I face the audience. All is quiet. I pause, smile, and begin, knowing that the conversation about to take place will not be what anyone expects.
I communicate five points before anyone can take a breath. One: I will focus on the pragmatic, bottom-line benefits of diversity in a knowledge economy. Two: I will not present an ideological argument, that is, base the case for diversity and inclusion on equity and social justice. Three: by diversity, I will mean cognitive diversity, differences in how we interpret, reason, and solve. My primary focus will be on differences in how we think. Identity diversity will contribute to cognitive diversity, but will not be the only cause. Four: as a mathematical social scientist, I will use models and formulae to structure arguments. At this point, if talking with students, I might ask security to lock the doors to prevent people from running for the exits. Five, I expect the interaction to be fun.
All five points—pragmatic benefits; not about social justice; cognitive diversity; mathematics; and a fun time—run counter to expectations for a diversity and inclusion presentation. They lay the groundwork for equally unexpected conclusions. I do not conclude that diversity always improves performance. Instead, I demonstrate the value of cognitive diversity on high-dimensional, complex tasks like engaging in scientific research, developing marketing plans, formulating technical trading strategies, and building a robust supply chain. I make no claim that diversity always helps. In fact, on simpler tasks like packing boxes or serving coffee, it likely has no effect.
The central takeaway is that diversity can produce bonuses. By bonuses, I mean the literal dictionary definition: something extra. I mean one plus a different one making three. These bonuses do not arise by magic. They come about when people with diverse cognitive repertoires work inclusively on complex tasks.
Though most often people frame diversity and inclusion efforts as the right thing to do, I am putting forth a contrary assumption that creating inclusive, diverse teams and workforces is the sensible and innovative thing to do—that on complex tasks, we need diversity.
That diversity cannot be arbitrary. The space of possible diversities is enormous. We cannot convene a random collection of diverse people and expect diversity bonuses. We need theoretical understandings of whether and how diversity can produce benefit on particular tasks. We need to make reasoned judgments about what type of diversity might be germane to the task at hand.
Education, life experiences, and identity—for example, race, gender, age, physical capabilities, and sexual orientation—can all contribute relevant cognitive diversity. How much each of these matters depends on the task. Diverse educational training may be more valuable on a team designing a new rocket-propulsion technology, diverse life experiences may matter most for a team rewriting a health care plan, and identity diversity may be crucial for a company launching a new product. To navigate the space of possible combinations, we need a science of diversity. And we need to gather and marshal evidence to guide our actions. As W. Edwards Deming once said, “In God we trust. All others must bring data.”
I have spent the last decade talking with people from businesses, government, and education about how to put together successful diverse teams. What I offer in this book are theory and frameworks for how to realize diversity bonuses. The theory provides a road map and structures our thinking.
The first step will be classifying the type of task: are you solving a problem, making a prediction, seeking creative ideas, or trying to discern the truth? The second step will be identifying task-relevant cognitive diversity. Not all diversity will be beneficial. You cannot expect bonuses from a random group of people.
The third step involves having a culture that enables successful interactions. Team members must get along, trust one another, share a common mission, and be committed for the long haul.2 Most organizations I have visited recognize that hiring people who look different or have different training is not sufficient to produce good outcomes. Hence, management consultants and human resource professionals speak not just of diversity, but of diversity and inclusion.
Just as the theory suggests that not any type of diversity will do, it also provides insights into how inclusion varies by task. Different tasks—creative, predictive, and so on—require different levels of engagement. On a creative task, diverse ideas can be compared. On a predictive task, diverse predictions can be averaged. Other tasks, like discerning the truth, will require deep engagement and interrogation of models, and hence, greater levels of trust.
On any task, achieving bonuses requires the right people and the right practices. Organizations that add diversity indiscriminately to reach targets and benchmarks will be disappointed. Team performance will suffer and people will come to see diversity efforts as counter to the organization’s core mission.
The sophisticated diversity practitioner recognizes that diversity bonuses require thoughtful hiring practices, and the creation of a culture that enables meaningful, organic interactions between people with different life experiences, educational backgrounds, and identities. This book provides a framework for achieving those bonuses.
When presenting to an audience, I try to gather information beforehand on how to partition the audience across three groups. The first consists of true believers, the diversity advocates. They want their organization and society writ large to be more inclusive for normative reasons—to redress past wrongs or because it is the right thing to do. The second group consists of the deniers. They believe that diversity and inclusion initiatives hinder performance. Some see these initiatives as affirmative action in disguise. The third group enters the room thinking that they would rather be almost anywhere else. They suffer from what the Economist in 2016 called diversity fatigue.3
I need all three groups to change how they think. The believers must pull back from the unrealistic, magical thinking that holds that identity-diverse groups perform best on all problems. Wanting something to be true or not true does not make it so. The deniers, many of whom attribute their own success to individual abilities, must open their minds to possible benefits. They must see the shortcomings of hiring people who look just like themselves or come from the same five schools. Those suffering from diversity fatigue must see diversity and inclusion as something other than a waste of time thrust upon them by a compliance officer.
I can only accomplish so much in sixty or ninety minutes. I ask that people set aside their political and normative positions and entertain the contrary assumption that diversity and inclusion can produce bonuses. I ask that they listen with open, skeptical minds. I ask that they challenge the logic and evidence. And challenges do come—whether delivered by a precocious seventh grader wondering how gender could influence the way a person approaches a scientific problem, by a skeptical mid-level executive buying into the logic but questioning the magnitude of diversity bonuses, or by the chairman of the Joint Chiefs of Staff asking how an organization whose members must follow orders can also be innovative.
Those answers will follow, as will a summary of what I learned on a decade-long tour that has included stops in Houghton, Michigan; Monroe, Louisiana; Frankfort, Kentucky; Redmond, Washington; and Princeton, New Jersey. I learned that rewriting a company or university’s diversity and inclusion statement matters far less than changing what people believe and how they behave toward one another. The right culture cannot be imposed from above by a bureaucracy or through elaborate diversity and inclusion strategic plans.
As Wes Pratt, the Chief Diversity Officer at Missouri State, remarked to me, we must be the process. The behaviors that produce diversity bonuses must emerge from the bottom up, organically. Each person who belongs to a diverse team brings a history and set of beliefs shaped by their identity. Those histories and beliefs must be validated and appreciated.
Achieving an organic, bottom-up inclusion requires that people believe in the value of interacting across differences and thus seek diversity bonuses. They must be all in. This more organic form of inclusion will be more likely in teams and organizations with a shared mission or goal. This more bottom-up inclusion will be bolstered by a shared understanding of diversity bonuses because people will see different ideas as worthy of deeper investigation.
In the best cases, people begin to appreciate the resonance between our identities and how we contribute. They see how identity differences add more than pragmatic contributions (more accurate predictions, better policies, more innovative solutions, and so on); they also add beauty, grandeur, and meaning to what is produced.
In my travels, I have been amazed by the depth and breadth of knowledge, skills, and passions people bring to their professions—be it building rockets, designing farm equipment, maintaining quality control in computer cable production, or managing financial portfolios. I have been awed by the kindness and generosity of people and of their dedication to creating more inclusive workplaces, schools, and societies. And yet, I often hear these same people lament the failures of their diversity groups.
It has been my goal to help people build more productive diverse teams. In each interaction, I could see the large gulf between precise blackboard truths and the messy reality of the world of people. Nevertheless, my experiences reinforced rather than attenuated my belief in the necessity of formal logic and models. I witnessed the power of mathematical rigor to cut through ideologically clouded thinking to reveal truths and guide action. Metaphors and stories can spur emotions and rally the team, but they cannot produce deep understandings or reveal the conditions necessary for bonuses to exist. Those conditions guide proper action, and enable us to take claims to data. That data can then be used to test and improve the theory.
As I look out through the glass walls of Mighty Good Coffee in Ann Arbor, I think of the words of T. S. Eliot: I have ended my exploring by arriving back where I started. The logic holds: cognitive and identity differences can produce bonuses. Achieving them in the real world takes practice. We need to learn the behaviors that make the theoretical bonuses real. I am not talking about just being nice. When making predictions, we have to include less accurate, diverse predictions—because we will do better. When hiring people, we have to see value in difference.
The logic I present, revealing the pragmatic benefits of diversity, intersects with and complements arguments for diversity and inclusion based on social justice. Many people, including colleagues and close friends, have said that though they accept the logic and agree that substantial evidence supports bonuses, they believe that pragmatic logics carry less weight than normative arguments. They question how I can place improvements in economic forecasts, marketing plans, and product offerings on equal standing with considerations of social justice and equity.
I do not dispute that point. One cannot equate a 1 percent increase in a portfolio’s return or a 4 percent reduction in shipping costs to the value of creating a world free of discrimination. We should care more about creating a world in which each person has an opportunity to succeed regardless of his or her identity more than we care about the IRS creating a tax form that is 6 percent more readable.
That said, I do not see the social justice frame and the diversity bonus frame as in competition with one another. We need not choose between applying our differences to improve our lot and creating an equitable and just world.4 On the contrary, I view embracing diversity bonuses as crucial to advancing social justice.
Moreover, if we ignore the pragmatic logic and emphasize only equity and social justice, we all but rule out achieving diversity bonuses. Bureaucratic rules that impose arbitrary diversity without any understanding of how to produce diversity bonuses will result in teams that do not make scientific breakthroughs, develop better health care plans, or write captivating screenplays. An occasional group may get lucky, but most diverse groups will not perform well.5 As a result, people will see diversity as costly and see a tradeoff in their pursuit of the normative ideal. Those who want a more inclusive society will find themselves climbing uphill barefoot on streets of broken glass. If, on the other hand, we learn the logic of diversity bonuses and learn how to form teams and interact with one another so as to produce diversity bonuses, we align the normative ideal with our self-interest. We skateboard downhill on smooth pavement.
The same logic can be applied to diversity efforts based on compliance with the law and changing demographics—the idea that America will soon be a nation with no majority group, so we must be inclusive. Though each promotes inclusive workforces, each positions inclusion, at least in the short term, as a sacrifice. The diversity-bonus logic does not. It shows that diversity can improve performance.
A KNOWLEDGE ECONOMY PHENOMENON
Historically, diversity bonuses have not been a central reason for promoting diverse interactions. That should be expected. The theory will show that diversity bonuses occur most often within teams of cognitive workers engaged in nonroutine tasks. As we transition to a knowledge economy and as more people work in teams on complex tasks, diversity bonuses become more relevant. We can find diversity bonuses in scientific research, on investment teams, in groups of neonatal surgeons and script doctors, and among groups of programmers and policy makers
Diversity bonuses were less prominent in the past when most workers were engaged in routine tasks. The logic is straightforward for why routine physical tasks cannot produce large diversity bonuses. Think of a group of Domino’s employees folding pizza boxes. The number they fold as a team equals the sum of the boxes folded by the individuals. No bonus exists. Similar logic holds for the number of packages delivered by the fleet of UPS drivers. The total equals the sum. No diversity bonuses arise. However, diversity does exist among the teams of engineers and mathematicians who devise the complex algorithms that route those trucks.
When diversity bonuses exist, the best group will not, as a rule, consist of the best individual performers according to some criterion. Instead, it will be diverse. I am not saying that an organization should hire less talented people. The claim is that talent is multi-dimensional. An organization should hire people with different talents and skills, an insight that I make formal by introducing the concept of a cognitive repertoire.
Similarly, selecting an optimal group requires consideration of the cognitive diversity each person adds. To be clear, a person cannot be diverse, but a person can add diversity. The diversity a person contributes will be relative to an extant group and with respect to a given task. The same person may add relevant diversity to one group on one task and not add diversity to a different group on a different task.
From a diversity-bonus standpoint, choices about whom to hire do not involve a tradeoff between excellence and diversity. If selecting a person to add to a research team or design group, the best choice will be the person who can add the most new ideas or apply the most novel tools. That may not be the person who would perform the best on her own, or who has the highest test scores. The diversity that a person adds to the group will matter as well.
The same logic applies when selecting a cohort. Each year, the University of California–Los Angeles receives over one hundred thousand applications for admission. Leading Wall Street and consulting firms receive similar numbers of resumes from job seekers. Google receives over three million applicants for employment. To cope with these numbers, organizations develop rubrics to predict the future success of applicants. These organizations do not blindly admit and hire those individuals who score highest by those rubrics. They embrace the contrary assumption and also seek diversity. UCLA wants students from multiple communities with diverse interests. They want students from Encino, Nipoma, and Brentwood. They also want philosophers, physicists, and French majors. Google wants people who learned programming from different books and professors. They want history, psychology, and ergonomics majors in addition to engineers and mathematicians.
Choosing only those people who ranked highest by the rubric would bias admits and hires and result in too many people who grew up in similar communities, attended the same small set of schools, and have had common life experiences and interests. Thus, the cohort of the best individuals will not be the best cohort. The best cohort will be diverse. It will consist of people who possess relevant knowledge bases, analytic tools, mental models, perspectives, and information. The group will include diverse, talented individuals.
Making intelligent choices of whom to hire and admit involves contemplating the bonuses that might arise from the abundance of combinations of skills. It is not a matter of deciding between the excellent candidate and the diverse one. People who find themselves torn between the highest-ranked candidate by traditional criteria and a diverse candidate often need only think harder. The diverse candidate may add more to the group or she may not.
Identifying what types of diversity may be germane requires careful thought. Relevant cognitive diversity can stem from our life and work experiences, our educational paths, and our identities. When identity diversity correlates with or causes germane cognitive diversity—and it often will—diversity bonuses underpin arguments for identity-based inclusion in the business world, the academy, and the nonprofit sector. It will not be true that identity-diverse teams always outperform homogenous teams, only that they can and do on a range of tasks. When identity diversity produces bonuses, inclusion involves no sacrifice. We do not confront a tradeoff between excellence and diversity. Excellence demands diversity.
The relevance of diversity bonuses depends in large part on the complexity of the modern world. On simple, separable tasks, we do not need diversity. We can hand them to smart, capable people. Thus, to appreciate the logic of diversity bonuses, we must first take stock of how we produce value in the modern economy, how we produce innovations, how we create art, and how we create successful policies. The short answer: we use teams. And those teams are diverse.
Work was not always this way. I grew up in the 1960s in Yankee Springs, Michigan, about thirty miles north of Kalamazoo, a place that Carl Sandburg reduces to a “spot on the map where the trains hesitate” in “The Sins of Kalamazoo.” Sandburg’s Kalamazoo is a place where children carve their initials on the ballyard fence, with a five-and-ten-cent store and hound dogs barking on the public square, where young people run off to see the world and, upon returning, remark that it is “all like Kalamazoo.” That characterization was an accurate picture of my childhood.
Today, Kalamazoo and the rest of America differ markedly from Sandburg’s caricature. Our economy no longer runs on people who perform routine jobs: cooper, farmer, blacksmith, merchant, and baker. Experts describe the modern economy with adjectives—weightless, flat, connected, information based, and cognitive—that would have made little sense in Sandburg’s time.
In Chicago, a more famous subject of Sandburg’s prose, people no longer butcher hogs or stack wheat. Chicagoans work in financial technology, e-commerce, plastics, medical technology, and biotechnology. The people Sandburg called “players with railroads” have been supplanted by transportation logistic engineers. The children of the hog butchers now work in biotech.
Sandburg’s Kalamazoo and Chicago have been replaced by a new world that presents an array of complex challenges and opportunities: managing elaborate supply chains, developing molecules to cure diseases, understanding the brain, preventing market crashes, and confronting climate change, to name just a few. The people who work in those jobs perform nonroutine cognitive tasks. Those tasks outstrip the capabilities of any one person, so organizations rely on teams. The best of those teams are diverse. They include people with diverse training, experiences, and identities.
The teams that excel achieve diversity bonuses, and that often requires behavioral changes. Those behavioral changes require practice guided by theory. If we understand the logic of diversity bonuses—that is, if we understand the logic and read the evidence—we are better able to identify actions that produce bonuses. Believing that they are there also helps. If we see inclusive actions as in our self-interest, we are more likely to act inclusively, engage diverse ideas, and produce bonuses.6
THE MIDDLE OF TOWN
To explain diversity bonuses, I present frameworks, construct models, evaluate empirical evidence, and explore illustrative cases. Together, these enable me to trace the boundaries of the domains in which diversity produces bonuses and those in which it does not.
I first describe the core logic for how diversity creates bonuses. I then unpack that logic by describing cognitive repertoires and linking those directly to better outcomes. Loosely defined, a person’s cognitive repertoire consists of the different ways in which that person thinks. Having established the logic, I take up the connections between cognitive and identity diversity by presenting three frameworks: icebergs, the timber-framed house, and the cloud.
That discussion lays the groundwork for interpreting empirical studies on diversity and team performance. The evidence of the benefits of cognitive diversity proves strong, bordering on overwhelming.7 Studies of identity diversity and performance also align with the theoretical models, though not as strongly. I hasten to add that in interpreting those studies, we must keep in mind that the data tell us what we currently achieve. The logic shows how much we could achieve in ideal circumstances. I conclude by embedding the diversity bonus logic within the larger business and societal cases for diversity and inclusion, offering thoughts on how we might better achieve bonuses and highlighting the need for practice and bottom-up integration.
As a preview of the type of claims that follow, one result will be that when a group of people make numerical predictions, the error of their collective prediction cannot be larger than the average of their individual errors. In other words, diverse predictive groups must be more accurate than their average member. Everyone (by that I mean the collective) is above average. Always.
Furthermore, the amount by which the group outperforms its average member depends on diversity: holding the average size of an error constant and making the predictions more diverse makes the group smarter. One analysis of forty thousand predictions by economists found that averaging two predictions reduced the expected error by 8 percent.8 That error reduction represents a diversity bonus—a significant, quantifiable improvement due to differences in how people think.
I have witnessed, read about, and heard accounts of hundreds of diversity bonuses. I believe that we can build organizations, and even a society, where we achieve them as a matter of course. Doing so requires that we understand how the bonuses occur and that we practice inclusive, productive behaviors. That is why I wrote this book—to help us generate diversity bonuses. If we can, we will improve society and expand opportunities.
My approach brackets social justice and equity-based arguments. That separation has been difficult. Over the past decade, I have met many talented people who have suffered from implicit and outright discrimination. I have met many others who have benefitted from a finger on a scale at key moments. As much as I would like to do so, I do not tell those personal stories. I omit them, not because the normative case lacks merit, but because creating diverse teams based on normative principles alone will produce few bonuses.
People will embrace the move toward a more integrated, inclusive society if they see benefits from doing so. To reach a place where each of us can contribute our unique skills, knowledge, and insights, we need to act with forethought. We need to apply logic, data, and experience in order to build effective, diverse teams capable of achieving bonuses. That type of scientific approach to diversity and inclusion veers neither left nor right. Like Main Street in Sandburg’s Kalamazoo, it runs straight through the middle of the town.