The Open Secret of What Works—and What Doesn’t—for Diversity, Equity, and Inclusion

The 1960s heralded all kinds of innovations—lava lamps, bubble wrap, the birth control pill, and the BASIC programming language, just to name a few.

Today, most of us use more modern versions of the inventions the 1960s produced.

With one notable exception: corporate antibias and diversity training. “There are dozens, if not hundreds, of studies [of these programs], most of them showing that diversity training has no effects,” says Alexandra Kalev, an associate professor of sociology and anthropology at Tel Aviv University, and coauthor with Harvard sociology professor Frank Dobbin of the new book Getting to Diversity.

And yet, despite the stack of evidence, “it is hard to find a Fortune 500 company without a diversity and harassment training these days,” they write.

When antibias trainings first emerged, they were rooted in mainstream psychological beliefs at the time: that attitudinal change precludes behavior change. In other words, you need to change what you think before you can change what you do.

Many corporate trainings were a by-product of President John F. Kennedy’s 1961 legislation to stop racial discrimination, Dobbin and Kalev explain. Eager to prove that they were on the right side of history, organizations swiftly developed recruitment and training programs for Black employees and “race relations training” for others on staff, despite there being no federal mandate to do so. When the Civil Rights Act of 1964 made sex discrimination illegal, organizations expanded these trainings to cover sex discrimination.

In their new book, Dobbin and Kalev explore not only why these trainings have endured despite the evidence, but also what works to make progress on diversity, equity, and inclusion inside organizations. Their book is informed by two big questions: “What goes on within companies to prevent women and people of color from flourishing in the way that white men with the very same qualifications flourish?” And, “How can firms tear those barriers down and work toward true inclusion?” To answer them, Dobbin and Kalev use data from over 800 companies captured between 1971 and 2015, as well as interviews with managers from another 100 companies conducted after 2007. The result is a book that is tailored for anyone who is craving actionable, evidence-based advice about how to create effective programs.

The week they published their book, I spoke with Dobbin and Kalev about why we continue to rely on unconscious bias training, the most impactful interventions for diversity, equity, and inclusion (DEI) progress, a DEI “stealth program,” and why organizations tend to lay off more women and people of color during recessions. Our conversation, edited for clarity and length, is below.

What have we learned from the past few decades of studying the effectiveness of diversity trainings?

Kalev: For decades, there has been so much research, and most of it shows that diversity training has no effect. Usually, the effects examined are short term changes—understanding of certain concepts, or plans for future behavior or, attitudes. Some studies find positive short-term effects, and some of them find negative effects.

What we’ve learned is that it’s very hard for people to be forced to change their attitudes. They act with reactance, needing to take control over their decisions and regain autonomy. We’ve learned that trying to suppress stereotypes makes those stereotypes more accessible. We’ve learned that the message of multiculturalism makes white men feel excluded. And when it comes to attitudes, we gain them through the life course, so no single afternoon session will undo that.

The barriers to diversity are diverse. Different companies, departments, groups, and industries have different barriers. If you bring on an over-the-counter diversity training, what are the chances you’ll be spot on with what’s going in your organization?

Nothing I told you is a secret. That research is not the innovation of our book. It’s out there and ignored.

Why do so many organizations opt for trainings that aren’t supported by evidence?

Dobbin: Part of it is that it’s easy to do. It’s an easy way to show that you’re trying to do the right thing, and it’s conspicuous because every person in the organization is exposed to it. Part of the reason it remains popular is that it doesn’t change what executives and managers have to do from day to day. Business can go on as usual.

Part of the reason trainings remains popular is that it doesn’t change what executives and managers have to do from day to day. Business can go on as usual.

We also tend to think, erroneously, that our behavior is driven entirely by our values and our ways of thinking about the world. So, if you can change people’s values and ways of thinking about the world, you can change their behavior. But often it goes the other way. If you change their behavior, you can often change how they think about the world.

The main indicator you look at to understand whether a program is working, or not, is how it impacts the numbers of women and people of color in management. Why is this so important? 

Kalev: Our data looks at several decades, and we look at the change in the number of managers from each demographic group over time. Getting to management means that you succeeded in getting promotions, that you get opportunities, you feel included, your talent is visible, you’re able to show your skill. We think that looking at the numbers actually includes all other dimensions of diversity and inclusion, because the number of people of color, or women or whatever underrepresented group in management is the bottom-line outcome of all other dimensions.

You find the two most impactful programs are formal, democratized mentoring systems and having a diversity manager or task force. Let’s talk about mentoring systems first. Why do they work?

Dobbin: We think mentoring systems work so well because they do several things at once. They help managers to see the obstacles that people of color and white women face in moving up the ladder. They make managers aware that there are fixable structural problems that they could begin to address, or at least they could look out for. Part of it is just making one person higher up in the organization aware of you, aware of your strengths and your aspirations. People mostly get superior jobs in an organization because somebody’s noticed them—they need mentors and sponsors to move up.

What are the most important components of a mentoring program?

Kalev: You want to have a formal program. Mentoring does occur informally, but it excludes women, people of color, other underrepresented groups that don’t have those natural networks. You want to have someone in charge of the program to make sure it is actually happening. You want to match based on interest. People might think for cultural comfort you would want to match based on demographic identities, but there are not enough potential mentors, like Black female mentors or Asian American women that can serve as mentors. So there’s simply not enough senior people from underrepresented groups to mentor. And those who are underrepresented senior leaders are usually overloaded with diversity related tasks.

Probably the most important reason to match by interest is that you want the relationships to break down the segregation, those glass walls between people like white men, for example, and women or people of color. You want to create matches that are actually cross-cultural, but based on interests, because that’s what we do in organizations: we work in functions like finance or AI, based on disciplines or interests.

Finally, you want to make sure that mentoring is open to everybody, not only to the potential stars, because those who you think are stars have already succeeded. Those that you don’t recognize as stars are exactly the ones who need the chance. Not everyone can be promoted, but you want to give everybody the opportunity to show their skill.

You also talk about an intervention I wasn’t familiar with, perhaps because it’s not typically seen as a tool to enhance DEI: self-managed teams. This is a group of employees inside an organization charged with managing and executing on their work without being guided by a manager. You call them a “stealth diversity program.” Why can they have such a positive impact?

Kalev: Usually work in modern organizations is divided by color and gender lines. For example, men and women can work side-by-side, but the woman is the assistant, which reproduces stereotypes, because we don’t see women outside of supporting roles.

Self-managed teams were designed to increase productivity from workers. But an unexpected consequence of them, or maybe expected if you think about it, is that they put together people that never worked together to collaborate, to be dependent on each other to design a product that will have the best consumer interface, or the best technology. They have to solve problems together.

That’s one of the best ways to reduce stereotypes, to propel awareness. It’s a different type of awareness, right? Because now I see you as a person. And because I’m dependent on you, in this task, I also devote more cognitive resources to evaluate you and cannot treat you as a category.

During recessions, firms are more likely to lay off women and people of color, and also less likely to hire them. I’m curious if you could talk a little bit about why you think that is the case.

Kalev: There’s a structural bias component and an individual bias component. For the individual bias component, basically, when the ship is going down, you want to have the ones you trust most to manage the ship. Even when women and people of color reach management, they are still suspect. In times of stress, we rely more on our biases.

We think mentoring systems work so well because they do several things at once. They help managers to see the obstacles that people of color and white women face in moving up the ladder.

And then there is a structural bias component. Women and minorities are usually the later comers. So I would have a lower tenure in management than my white male colleague. And many of the layoffs and downsizings are “last to come first to go.”

Women and minorities are also still in marginalized jobs, such as in human resources, public relations, even in management. So women and minorities are more likely to be in units that are more likely to be axed.

One of the things that I was surprised by in the book was that having a diversity task force or diversity manager was such an impactful intervention. That’s because I’ve seen so many companies appoint someone a diversity manager, or create a task force, but then give that person or group virtually no budget or power to do anything. I’d love to hear why I’m wrong and why these interventions can be powerful.

Dobbin: We interviewed hundreds of diversity managers who would often say, “I have no power, I have no budget. I feel like I’m not doing anything.” But a number said there are a few things they can do, and one of the things that’s most effective is “I just ask a line manager why they made that decision.” Essentially what diversity managers do is they activate social accountability in people, and the knowledge that someone might ask you about your decisions.

Diversity managers usually have access to the human relation information system, and they’re looking at the numbers, and they’re looking at them by department, and they’re looking at hiring slates. One thing they frequently do is talk to managers after, say, a seasonal round of hiring or promotions. And they’ll ask, “So it seemed like half the applicants were women, but you didn’t find any women that you wanted to hire. Should we be searching somewhere else? Is there something wrong with the candidates we’re getting?”

There’s this sociological idea of the looking-glass self, which is when you put yourself in the perspective of somebody who’s observing you. You’re looking in the mirror at your own behavior and asking, How does this look to someone observing my behavior? How does it look if I interviewed 15 men and 15 women and hired seven men and one woman? It’s not the same as a grievance procedure where you’re being threatened through a formal complaint process. This is just social influence—what are the norms, and how is this going to look to someone else?