This article is excerpted from a Chronicle special report, “The Future of Diversity Training,” available in the Chronicle Store.
Diversity-training programs are now practically a rite of passage for college faculty and staff members, yet the evidence that they are effective is underwhelming. Even after years of diversity training on college campuses, academe has made little progress in diversifying faculties, to take just one example.
While there is a significant body of research on diversity training dating back decades, many studies rely on surveys that ask how participants felt about the training or assess what they’ve learned, while relatively few try to determine whether the training changed how people behave.
Few of the studies use randomized controlled trials, considered the gold standard to measure the effectiveness of treatments and interventions.
The studies that do exist have found mixed results. Some show that participants learn about people from other backgrounds and that training can have an effect on beliefs and behaviors (although the latter fades over time). Others show that diversity-training programs can trigger negative feelings in participants and even harm the very groups they’re intended to help.
Despite the evidence, a survey of 670 colleges in 2016 found that two-thirds had some kind of diversity training for faculty. “Ironically, universities believe the consultants rather than looking at the research themselves,” says Frank Dobbin, chair of the sociology department at Harvard University, who conducted the survey with Alexandra Kalev, chair of the department of sociology and anthropology at Tel Aviv University. Their research found that some kinds of diversity training, such as harassment training that focuses on avoiding litigation, is largely ineffective and can even backfire. “You would think if anybody would look at the research it would be college presidents, but apparently not,” says Dobbin.
Some researchers argue that part of the problem with determining whether diversity training works is that a lot of the existing studies were poorly designed.
Patricia G. Devine, a professor of psychology at the University of Wisconsin at Madison, says that there is more enthusiasm and investment in diversity training than proof of its efficacy. “All too often, there is not sufficient analysis of the problem to be solved, but organizations think diversity training should be provided,” Devine wrote in an email. “The disappointment that follows the delivery of diversity training often results from unrealistic expectations about what diversity training would do and/or failure to systematically evaluate the content of the diversity training to be delivered.”
Devine was the lead author of a multidisciplinary meta-analysis published in 2022 that found researchers often assess the success of diversity-training programs by using measures that are far removed from the actual goals of such programs. Studies may ask whether participants rated programs favorably, rather than assess the promotion rates of employees belonging to historically marginalized groups, for example.
“Most diversity-training programs have not been rigorously tested for their effectiveness, and often rely on intuition and wishful thinking (i.e., the practitioners offer programs that seem to make sense and they hope they will work),” Devine wrote.
Similarly, a review of 418 prejudice reduction experiments reported from 2007 to 2019 found that most evaluations of diversity training were less than revealing. They “look like customer-satisfaction surveys (‘How much did you appreciate this?’) or elementary-school worksheets (‘Tell me what you learned today about stereotyping’),” wrote the lead author, Elizabeth Levy Paluck, a professor of psychology and public affairs at Princeton University, in The Washington Post. The study concluded that much of the research was “ill-suited to provide actionable, evidence-based recommendations for reducing prejudice.”
Paluck also led a 2009 meta-analysis that reviewed 985 papers studying interventions aimed at reducing prejudice. The authors examined studies of the most common approaches to reducing prejudice and concluded that antibias, multicultural, and moral education “have not been examined with a great deal of rigor” and that many of the training programs were “theoretically ungrounded.” The study points to cooperative learning, which includes lessons in which students teach and learn from one another, as an approach that is promising but still requires more evidence.
Paluck theorizes that one hurdle to gathering more evidence about what works and what doesn’t is that organizations may not want to participate in high-quality research. Executives may be reluctant to share their data and risk prompting lawsuits or learning that their diversity training is ineffective, for example.
Other researchers have found modest or mixed results from diversity-training programs. Katerina Bezrukova, chair and associate professor of organization and human resources at the University at Buffalo’s School of Management, led a team of researchers who conducted a meta-analysis, published in 2016, of 260 studies examining the impact of diversity training over time. Over all, the study found a small to moderate, but positive, impact on cognitive outcomes such as what participants learned about discrimination and smaller impacts on attitudes and behaviors. The results also suggest that the effects on cognitive knowledge are often maintained while changes to behavior and attitudes decline over time.
The study found that whether diversity training was mandatory or voluntary made no difference in the results. It also found that diversity training was more effective when paired with other kinds of diversity programs, such as efforts to improve recruitment, over a period of time.
Bezrukova said the study reinforced the idea that while people can learn very quickly, they are “fundamentally resistant to any behavioral changes.” Based on her research, she advises pairing diversity-training courses with other diversity efforts over a sustained period of time for the greatest impact.
A large study led by Edward H. Chang, now an assistant professor of business administration at Harvard Business School, found mixed results. The 2019 study, of about 3,000 employees at a global company, tested the impact of a voluntary one hour online training primarily intended to encourage more inclusive attitudes and behaviors toward women. Among groups who started off relatively less supportive of women — many of whom were employees outside the United States — the training had no significant impact on behaviors but did have a positive impact on attitudes, meaning employees acknowledged that discrimination against women exists, for example. The training also prompted female employees in the United States to connect with more experienced colleagues as mentors, perhaps, the authors hypothesize, by signaling to women the need to be proactive in their careers to advance.
The authors concluded that the one-time diversity trainings that are typical in U.S. workplaces “are unlikely to be stand-alone solutions for promoting equality in the workplace,” especially since such training appears to have little impact on the actions of the employees whose behaviors corporations would most like to improve.
Discussions of implicit bias, also known as unconscious bias, are a mainstay of diversity-training programs in both corporate America and higher education. The theory behind such training is that many people are unaware of their biases and prejudices and that the solution is to simply educate them.
A 2023 study led by Calvin K. Lai, now an associate professor at the Diversity Science Lab at Washington University in St. Louis, examined a daylong implicit-bias diversity-training program for police officers. The study found that the program helped to increase knowledge of, concerns about, and intentions to address bias, but the results did not last.
Likewise, a meta-analysis of 492 studies published in 2019 and led by Patrick S. Forscher, now a research lead at a nonprofit that uses behavioral science to help reduce poverty, found that while interventions could slightly reduce implicit bias, they generally had little or no impact on behavior. “Our findings suggest that changes in implicit measures are possible, but those changes do not necessarily translate into changes in explicit measures or behavior,” the authors concluded.
Some research has found that typical diversity training is not only ineffective but can even backfire, harming the very groups it is intended to help. Dobbin, of Harvard, and Kalev, of Tel Aviv University, wrote Getting to Diversity: What Works and What Doesn’t (Harvard University Press, 2022), which details their research on diversity training and examines alternative approaches to diversifying workplaces.
The researchers studied federal workforce data from more than 800 companies that employed eight million workers over 30 years. They found that antibias training that places blame on individual employees not only fails to increase the diversity in management, which they view as a key metric, but can even cause harm: It can activate biases and cause employees to feel resentful or angry, with a negative impact on employees in marginalized groups.
Among the findings from their research:
- According to their model, the effect of adopting training that emphasized compliance with the law and avoiding lawsuits, holding all else equal, was to decrease the share of Black women who were managers by more than 13 percent.
- Companies that instituted diversity-training programs for all employees, with all else being equal, saw the proportion of managers who were Latina women decrease by 6 percent, which they speculate may be because there are so few in management to begin with that they end up being a target of backlash.
- After corporations began providing harassment-training programs covering sexual, racial, and other kinds of harassment, the proportion of managers who were white women, Black women, and Asian American men decreased.
Other researchers have also found that diversity training can have unintended consequences, perhaps by changing perceptions of the company offering the training. A 2015 study led by Laura M. Brady found that when a company offered diversity training, women perceived the firm as being fairer toward women and less likely to have practiced sexism. As a result, the women at companies that offered diversity training were also less likely to be supportive of litigation charging the company with sexism.
The authors of the study concluded that “the mere presence of diversity structures causes women to believe that women are treated more fairly in the workplace. This belief in fairness, in turn, leads women to minimize discrimination against women and to be less supportive of women’s mobilization against discrimination.”
Kalev and Dobbin argue that if companies want to fix inequality, they should focus on bias and discrimination in the way they operate instead of trying to “fix” those problems in individual employees. In lieu of antibias diversity-training programs, for example, they advocate for strategies such as cultural-inclusion training, which aims to improve how people communicate and collaborate, and emphasize that managers can be part of the solution to workplace inequality. Their research found that when companies put in place cultural-inclusion training for their managers, management became more diverse.
They also urge companies — and colleges — to take steps to improve workplaces for everyone in ways that women and people of color, especially, may benefit.
What might that look like? Organizations can open up recruitment to widen the pool of potential employees, build mentoring programs to help ensure that women and people of color can get the help and social connections they need to advance in their careers, and expand access to programs to help employees balance work-life responsibilities. Colleges could, for example, extend the tenure clock to allow people to spend more time with their children when they are young; a similar program at the global financial-services firm Deloitte helped to boost the number of women at the partner rank.
If there’s little evidence that diversity training improves behaviors or diversity, why do so many companies and colleges continue to offer it?
To Bezrukova, the modest evidence that diversity training can help means it’s worth doing: “We are a global society, and we need to coexist with people who are different.”
As the Brady study suggests, the training may tamp down potential lawsuits alleging discrimination. At the very least, it signals a college’s commitment to improving diversity, says Musa al-Gharbi, an assistant professor in the School of Communication and Journalism at Stony Brook University and a research fellow at Heterodox Academy. “If there is a conflict, if there is a hostile work environment, if someone does resign and they want to sue their employer as a result of conflicts that occurred in the workplace, employers can point to the training that they’re doing and say, ‘Look, you know, we tried our best.’” There is some evidence that having diversity programs, such as training, in place can help protect companies legally.
Dobbin says that college lawyers like to encourage diversity and harassment training. “Sometimes legal counsels will tell them if they don’t have diversity and sexual-harassment training, it looks bad because their peers have it,” he says. And despite the available evidence, people like to believe that a quick fix will be able to change culture, he says.
Finally, he believes colleges have tended to focus too much on trying to fix bias at the individual level rather than addressing systemic issues like failing to recruit from the right places, failing to mentor all faculty, or making changes to support the work-life balance of all employees.
“As long as we try to treat the problem as an individual problem, we don’t solve the systemic problems that could make a difference,” Dobbin says.
