Does Diversity Training Work? Clockwise’s CEO on the Challenges of Fighting Unconscious Bias was originally published on Vault.
Earlier this month, President Trump issued a memo to federal agencies ordering them to cease diversity training for employees, which he labeled “divisive”. The order comes at the end of a summer fraught with nationwide protests against racial injustice and renewed calls to reform the institutions that reinforce systemic racism. The President’s messaging is problematic at best, but such a move does have the potential to ignite a more productive conversation: does diversity training work?
Studies have shown that corporate diversity training is often a mixed bag. It may call attention to issues of unconscious bias in the workplace, but the jury is still out on whether or not it results in real change. While categorically dismissing all diversity initiatives as “divisive” is counterproductive, companies can begin to rethink how they foster a workplace environment of inclusion and respect. To better understand where diversity training falls short and how companies can implement effective, sustainable inclusion efforts, we spoke with Matt Martin – CEO of the calendar optimization platform Clockwise.
Vault: What is diversity training? It sounds self-explanatory, but not everybody has undergone diversity training at work. When and why is it used, how is it applied, and how is its efficacy or success is measured?
Matt Martin: Diversity training, also called bias training, comes out of research from a team of social psychologists at the University of Washington and Yale. These researchers developed a test to measure how unintentionally biased someone is. The Implicit Association Test (IAT) finds that 90-95% of people have unconscious biases that impact their decision-making. The goal is to make people aware of their unconscious biases so they can begin the work of correcting them.
Nearly every Fortune 500 company offers a variety of diversity training, yet systemic discrimination is still very much a part of corporate America. One study showed that when hiring managers saw identical resumes, they were 74% more likely to hire candidates with white-sounding names. In tech, just 3.1% of workers are Black, and Silicon Valley is just 3% Black. This is despite research showing that teams with more diversity outperform more homogeneous teams.
Most organizations that implement diversity training either don't measure its effectiveness or don't report the results publicly. So it's difficult to know whether or not it’s working for them. However, researchers have measured whether and to what extent people behave and believe differently after diversity training using questionnaires.
Vault: And what do those measurements show? Does diversity training change minds or behavior?
MM: Hundreds of studies show that bias training doesn't actually decrease unconscious bias. Frank Dobbin, professor of sociology at Harvard, and Alexandra Kalev, associate professor of sociology and anthropology at Tel Aviv University, did a wide-ranging meta-analysis of the research and found little evidence that training reduces bias.
However, some researchers take the opposite view. “We believe that pessimism is premature,” Alex Lindsey, Eden King, Ashley Membere, and Ho Kwan Cheung write for HBR. “A recent meta-analysis of over 40 years of diversity training evaluations showed that diversity training can work, especially when it targets awareness and skill development and occurs over a significant period of time.”
Vault: When diversity training falls short, what do you believe are the underlying causes of its ineffectiveness? How are current models of diversity training lacking?
MM: At the end of the day, the short answer is that we still don't really understand how to change people's minds or their behavior. At best, bias training makes people aware of how they currently see things and gives them another way to see them. Unfortunately, that isn't enough. My takeaway from the research is that bias training is not effective on its own, and absolutely shouldn’t be seen as an effective way to change behavior. But, if implemented as part of a larger set of efforts, it can be a helpful starting point for a larger conversation.
Vault: You echo that conclusion in a piece you recently wrote for Fast Company where you said that “trainings are not leading to behavioral change” and that “changing attitudes does not necessarily change behavior.” What is the difference—and the relationship—between changing attitudes and changing behaviors? How do you change an attitude? How do you change behavior?
MM: Researchers Dobbin and Kalev couldn’t find evidence that diversity training reliably changes participants' minds. They examined 426 studies in which participants took the IAT test before and after anti-bias training. Their answers were pretty much the same after as they were before. Just eight studies showed white people's unconscious bias against Black people diminished after training. Yet a follow-up experiment with a fake intervention showed that people might just be learning how to game the test. And the results dissipated after a few days anyway.
So changing people's minds in any kind of lasting, real way is tough. It certainly takes more than one short training session.
But then changing their behavior is a whole different animal. Even if you can change people's beliefs, which is hard enough, that doesn't always translate easily into behavior change. There were even fewer studies showing that diversity training caused people to act with less bias.
Vault: In the same piece, you write, “There's surprisingly little correlation between most people's attitudes and behavior… people will routinely discriminate against people without feeling biased against or thinking poorly of them.” Is this what we call “unconscious bias?” Where does this come from? If bias can truly be unconscious, can we ever overcome discrimination in the workplace?
MM: Yes, this is generally what people refer to as “unconscious bias.” It’s unpleasant to think about or admit, but we all grow up inundated with false information about marginalized people that biases our perception of them. This is part of what the academic world refers to as “systemic racism.”
When most people think about racism, they imagine deliberate animosity toward non-white people on the basis of their race. But in reality, most racism is much more subtle and less conscious.
The examples are endless. Studies show people associate Black neighborhoods with poverty and crime. David French wrote an essay about how adopting Naomi, who is Ethiopian, opened his eyes to systemic racism. French writes that one of Naomi's classmates said that she couldn't come to Naomi's house for a play date because “My dad says it's dangerous to go Black people's neighborhoods.”
This father is likely ignorant of the role the federal government played and local zoning ordinances play in locking Black families out of wealthy, low-crime neighborhoods. Rather, he assumes poverty and crime are endemic to Black culture. This isn't racism as generally understood by white people, but has the impact of separating Black and white students and further entrenching Black poverty.
These negative, wrong assumptions about Black people lead to real-world harm. Studies show teachers give Black students harsher punishments than white students for the same offenses. Same for judges and Black defendants. Until recently, medical school textbooks claimed Black patients have a higher pain tolerance than white patients, despite there being no evidence this is true. As a result of systemic racism, doctors routinely undertreat Black pain.
French describes a white woman demanding that Naomi prove she was allowed to swim despite wearing the bracelet to indicate she was allowed. And how a police officer asked Naomi who she was with and what she was shopping for at a department store. This never happened to their white daughter, and studies show Black shoppers are trailed and harassed by store security at much higher rates than white shoppers.
If we want to overcome discrimination in the workplace, I believe the first step is acknowledging systemic racism, sexism, and other forms of unconscious bias. Then we need to seek effective remedies.
Vault: Does diversity training ever do more harm than good?
MM: The Guardian Science Correspondent Hannah Devlin pointed out that diversity training can exacerbate bias. Devlin points to an experiment where aboriginal Canadians reported feeling less valued by white people who took a race-based version of the IAT than white people who took a different version.
Dobbin and Kalev found that anti-bias training can actually make some people even more racist and feel more animosity toward other groups, particularly people who are aware of and comfortable with their racial stereotypes. According to Dobbin and Kalev, participants who either refuse to see their own bias or think it's fine to be biased end up more racist after training. For example, SLCPD Officer Jeffrey Denning said he still viewed racism as a “buzz topic,” overblown relative to its importance after bias training. Denning said that disparate use of force on Black people is more likely due to their “higher degree of crime,” and doubted that his biases might influence his thinking and behavior.
Vault: Diversity training is often reactionary—the response to a conflict or controversy. Aside from one-off expensive, disruptive, and often tone-deaf trainings, what do companies need to do to foster a culture of respect and inclusion? What kinds of programs and policies can companies implement on a wider scale to create lasting change within their workplace culture?
MM: There are five things companies should do to reduce discrimination. First, you need to start measuring discrimination. Regularly survey employees to determine whether they're seeing improvements in diversity, equality, and inclusion. Ask employees to anonymously share where they're seeing bias, counterproductive work behaviors like off-color jokes or comments, and alienating incidents or microaggressions. That way you know whether the problem is getting better or worse over time. It can also tell you whether your interventions are working. You also need to measure and set goals around the diversity of your company's recruitment, promotions, and leadership.
Second, implement diversity training that is ongoing, includes a skills component, and is just one part of a larger effort around diversity, equality, and inclusion. Find internal advocates who will push for the reforms that enable diverse hiring such as inclusive job announcements and search criteria. Make sure you also implement systems of accountability with clear support from company leadership.
Third, make sure your diversity training is voluntary. One reason diversity training backfires is that some workers feel personally attacked and afraid they'll be discriminated against. Plus, mandatory training causes some people to resist because they dislike feeling controlled. It's also a good idea to make it clear that the purpose is to make the company better, not to avoid getting sued. You might also want to point to research showing that diversity makes companies more profitable and productive.
Fourth, it's important to evaluate workers on inclusion. Training workers on a topic they know they won't be evaluated on doesn’t stick. To get workers invested in the material, let them know you'll evaluate them on their diversity, equality, and inclusion efforts. You could evaluate employees based on how well they moderate discussions to make sure everyone has a chance to contribute, solicit opinions from the appropriate people, give people proper credit for their work, assume that their colleagues have reasonably advanced knowledge, and productively navigate disagreements.
Finally, rather than putting all the responsibility on HR, engage leadership. Anti-discrimination efforts that put the onus on decision-makers to find ways to solve the problem themselves worked best in Dobbin and Kalev's research. Companies markedly increased their managerial diversity after they tasked corporate managers to find women and minority recruits. Leaders from various departments should form task forces to gather and examine hiring, retention, pay, and promotion data. They should also identify areas for improvement, research potential solutions, and sell their initiatives to their departments. For example, leaders should ensure job postings don't include language that alienates diverse candidates. And they should measure how whiteboard interviews or alcohol-centric social events, for example, impact diversity.
Vault: Can you tell us how you, as a CEO, cultivate a culture of inclusion in your own organization? How does Clockwise champion these values, and what success have you seen with the methods you use?
MM: We've always felt strongly that a diverse workforce ultimately creates a stronger company that builds a better product. After all, Clockwise is a general-purpose application with a user base as diverse as the general population of knowledge workers. We need diverse opinions, perspectives, and experiences to build a product that is going to work well for everyone.
We've always stressed sourcing under-represented groups in our hiring efforts so that we're always evaluating a diverse set of candidates. That said, our results, until recently, were falling short. To that end, we've doubled down not only on sourcing a diverse set of candidates, but measuring our efforts and holding ourselves accountable to the results.
We recently worked with CultureAmp to put out our first Diversity and Inclusion Survey to establish our current baseline and track our internal diversity over time. This survey will be sent out regularly so that we can track our progress over time and evaluate the effectiveness of our efforts. Notably, the survey does not simply seek to establish what percentage of our employees come from underrepresented groups, it also measures equity and inclusion by probing on access to resources, career opportunities, decision-making practices, and overall belonging.
We're also experimenting with a variety of ways to ensure that everyone at the company builds empathy with their coworkers. According to Dobbin and Kalev, people who regularly interact with diverse groups engage in less stereotyping. As one example, at Clockwise, we hold weekly Zoom lunches that have random breakout rooms where we talk to coworkers we may not have had any face time with in a while.
Matt Martin is CEO and Co-founder of Clockwise. Before Clockwise he helped build RelateIQ, a company acquired by Salesforce for $390 million. He has previously been a front-end software engineer, a civil litigator, and a founder of a variety of software startups. He holds a B.A. from Dartmouth and a J.D. from UPenn in Philadelphia.