Anti-Bias Training Doesn’t Stop People From Making Biased Decisions

Andrea S Kramer, Alton B Harris

May 24, 2023

Share Download

Structural changes are the only real hope we have of confronting structural discrimination so we can achieve fair, equal, and equitable career outcomes for women and men.

Approximately 98 percent of American companies have some sort of gender DEI programs.

Because of this, DEI consulting has become big business. Very big business.

Globally, more than $7.5 billion was spent on DEI programming in 2020, and it is anticipated to increase to $17.2 billion by 2027. The United States market for DEI alone is estimated to be $3.7 billion, and it is expected to compound at an annual growth rate of more than 13 percent through that seven-year period.

Every year business, professional and nonprofit organization executives work hard to strengthen workplace diversity, with 86 percent of North American financial executives expanding DEI training budgets, according to a survey by OneStream Software.

In addition to this enormous budget spend, the popular press, social media, and a large number of books and articles promote the benefits of DEI and condemn the harms caused by gender-biased workplaces.

Yet, despite the time, resources, and attention, very few organizations have made meaningful progress toward increasing the proportion of women in their senior leadership ranks, much less realizing the benefits in doing so.

With so much time, attention, and money being paid to this issue, something must be wrong with the common approaches to solve this problem.

We believe there are three fundamental mistakes with the current approaches to gender diversity training:

  • Most diversity training is focused on “de-biasing” individuals. This simply doesn’t work: people cannot be trained out of their biases with de-biasing training.

  • A great deal of effort is devoted to “fixing” women. Women are urged to change their leadership style; increase their confidence; and become more vocal. Such efforts are misguided, demeaning to women, and ineffective in reducing workplace gender inequality.

  • Most diversity training is focused on the symptoms of gender inequality but not the fundamental causes of gender inequality (systematically discriminatory workplace systems, processes, and practices).

Let’s consider each of these approaches in turn.


Through workshops, videos, role-playing exercises, prescribed reading, and moderated discussions, most DEI programs focus on training individuals. This training (which includes anti-bias and anti-harassment training) is intended to help people become knowledgeable about the nature and extent of bias; to become aware of their own (generally unconscious) biases; and to aid them in ensuring that their biases do not influence their behaviors.

The assumption underlying such training is that increasing knowledge and awareness of bias will reduce the presence of bias in the workplace and reduce the discriminatory barriers women face in pursuing their careers.

Much of the impetus behind individualized de-bias training stems directly from Title VII of the Civil Rights Act of 1964, which forbids employment discrimination because of sex.

There are benefits to increasing knowledge and information about how biases result in different workplace experiences for women and men. Making people aware of their own biases, however, has proven to be largely ineffective in changing their attitudes and behavior. Let’s examine why this is the case.


Research Regarding Anti-Bias Training

A 2009 comprehensive review by Harvard and Yale academics of 985 studies of programs designed to reduce biased behavior by increasing awareness of one’s own bias found little or no evidence that they were effective.

These programs were wide-ranging in their approaches, including:

  • workplace diversity initiatives

  • anti-bias education

  • multicultural education

  • dialogue groups

  • cooperation training

  • moral and values education

  • intergroup contact

  • education in reconciliation techniques

  • media interventions

  • intercultural and sensitivity training

  • cognitive training

  • miscellaneous other techniques and interventions

The researchers did not find evidence that any of these approaches reduced the biased behavior of the people who had participated in the training. In fact, the researchers concluded that the studies “[do] not reveal whether, when, or why [anti-bias] interventions reduce prejudice in the world.”

More recently, a 40-year analysis of 260 studies on discreet anti-bias training programs similarly concluded that while many programs increased participants’ cognitive knowledge about bias and its effects, none was particularly effective in changing attitudes or behaviors.

Another study that looked at 17 promising anti-bias training programs found that only eight of them reduced unconscious (implicit) biases. Even when the programs did reduce unconscious bias, the reduction vanished after 24 hours. Worse still, none of these programs resulted in any measurable changes in the participants’ explicit biases or actual behaviors.

Finally, a 2018 study of unconscious bias training (UBT) by the Equality and Human Rights Commission of the United Kingdom could find no evidence to indicate that UBT is effective for behavior change.

Together, these exhaustive research studies demonstrate that anti-bias training has little or no effect in reducing discriminatory decisions affecting women’s career advancement. It might be argued, however, that it is not that anti-bias training in and of itself is ineffective, rather that the training must be done in specific ways to be effective. Let’s examine this idea.


Compulsory versus Voluntary Programs

Putting aside for the moment the question of whether anti-bias training of any sort can reduce biased behavior, there is the issue of whether it is better to make such training compulsory or voluntary.

The argument for compulsory training is obvious: unless you make the training compulsory, you may not reach those people who need it the most.

On the other hand, people who are required to attend compulsory training resent the training and are likely to ignore its positive message or worse.

Let’s look at the pros and cons of both approaches.


Compulsory Programs

Compulsory training ensures that all supervisors are aware of the extent, nature, and discriminatory consequences of unconscious bias. Supervisors—in particular middle managers—are positioned to make biased career advancement decisions that can have direct, adverse consequences for more junior women’s progress. Yet male supervisors are likely to be the most reluctant group to voluntarily participate in anti-bias training.

We know that compulsory training provokes anger and hostility. One study, for example, found that participants reported greater animosity to diverse group members after such training than before the training.


Voluntary Programs

We have found that voluntary training is far more likely than compulsory training to be accepted, making it more effective for those people who participate.

Its drawback, however, is that the employees least likely to participate are likely to be those with the lowest levels of awareness about their own biases and the most in need of training.

Management that chooses to provide voluntary programs might encourage participation through incentives, continuing education credits, networking event, access, and senior leadership sessions.

It is worth noting that Google’s Unconscious Bias @ Work, Facebook’s Unconscious Bias Training, and Microsoft’s Self-Paced Online Unconscious Bias Training—three thoughtfully designed diversity training programs—are all voluntary.


Six of One, Half Dozen of the Other

In the end, though, it may not make much difference if anti-bias training is compulsory or voluntary, or who attends. There is no conclusive evidence that either type of program reduces biased behavior, diminishes discriminatory outcomes, or increases the promotion of women in leadership.

As Frank Dobbin, of Harvard University, and Alexandra Kalev, of Tel Aviv University, point out, whether anti-bias training is voluntary or compulsory, such “training alone [does] not change attitudes or behavior, or not by much and not for long.”

The ineffectiveness of traditional anti-bias training was confirmed by Edward H. Chang and his research colleagues in a large field experiment involving employees of a global company.

The experiment consisted of a voluntary training program that sought to determine if a stand-alone online diversity program could affect employees’ attitudes and behavior toward women at work. The format and length of the program (limited to one hour) was typical of the types of diversity training frequently offered by most major American corporations. The program was designed in accordance with what are considered to be best practices to accomplish the following:

  • raise awareness about the pervasiveness of stereotypes and bias

  • provide scientific evidence of the discriminatory effect of stereotyping on important workplace behaviors

  • destigmatize and expose participants to their own stereotyping

  • provide evidence-based strategies for overcoming stereotyping

  • allow employees to practice evidence-based strategies to combat stereotyping by responding to different workplace scenarios

More than 2,000 people took part in this program. Participants were randomly given one of three types of training: gender bias training, general bias training, and psychological safety and active listening training that did not mention stereotypes.

Neither gender-bias nor general-bias training had any measurable effect on participants’ behavior except for a small effect on the behavior of women participants from the United States, hardly the primary group diversity trainers seek to influence.

The researchers concluded, “The lack of change in the behaviors of dominant group members indicates additional remedies [beyond such diversity training] are needed to improve the overall workplace experiences of women and racial minorities.”


Diversity Training That Might Work

There has been quite a bit of pushback against the conclusions that all diversity training programs fail to change behavior. For example, researchers writing in the Harvard Business Review argue that two types of diversity training show some promise in changing behavior.


Perspective Taking

The first type of diversity training that could offer benefits involves participants engaging in perspective taking mentally walking in someone else’s shoes.

The researchers had students write a few sentences about the challenges they imagined members of a marginalized group might experience. They were then tested with respect to their “pro-diversity attitudes and behavioral intentions toward those marginalized groups.”

The researchers reported that participants said that their attitudes and “behavioral intentions” improved, after this exercise, and these positive effects persisted when the participants were retested eight months later. No information was provided, however, about the extent of the students’ real-world behavioral changes, or whether the report of their “behavioral intention” resulted in actual changes in their behavior.


Specific, Measurable, and Challenging Goal-Setting

The second type of diversity training that the researchers claim holds promise is called “SMC goal-setting.” It involves asking participants to set specific, measurable, and challenging (but attainable) personal goals to increase workplace diversity. (The Google training course Unconscious Bias @ Work does this.)

The experiment again involved undergraduate students, and the researchers reported that goal-setting with respect to promoting diversity led to more positive diversity attitudes nine months after the training.

They gave no evidence, however, as to how much the participants’ diversity attitudes increased or whether the goal-setting had positively influenced their actual behavior.

Although perspective taking and SMC goal-setting seem to produce some positive effects, these experiments require further validation to see if the techniques can increase positive workplace outcomes for women or reduce workplace gender inequality.

While organizations should certainly consider including these types of techniques in their overall diversity efforts, the fact remains that diversity training programs need to be combined with other initiatives to increase workplace gender diversity equity, and inclusion.


Defensiveness Avoidance

It has also been claimed that diversity training specifically designed to avoid defensiveness on the part of white men can work to decrease biased behavior.

In one study, the training material was presented in the context of actual day-to-day workplace activities, and it provided participants with action-oriented bias mitigation strategies.

Researchers found that eight months after participating in this type of training, employees at a global technology company reported engaging in these strategies to mitigate the bias that had been highlighted in the training.

However, the researchers were appropriately understated in their conclusion, “Unconscious bias training can be a useful component of [other] diversity and inclusion efforts.”


The Inescapable Conclusion

Based on the available research and our personal experiences, stand-alone diversity training, whether compulsory or voluntary, does little to change attitudes, beliefs, or behaviors, even over an extended period of time. It helps educate employees about gender inequality, women’s workplace advancement obstacles, and an organization’s broader diversity goals. Indeed, there is no evidence that diversity training on its own will reduce workplace gender inequality or strengthen women’s representation in senior leadership.

This is not surprising. Providing people with the information they need to make good (unbiased) decisions doesn’t do much to stop them from making bad (biased) decisions. As public health expert Victor Strecher, of the University of Michigan, puts it, “We’ve known for over 50 years that providing information alone to people does not change their behavior.”

Many people still smoke, despite knowing the risks of tobacco use. Children exposed to drug education programs aren’t any less likely to try illicit drugs. Providing people with information about workplace sexual harassment doesn’t reduce its prevalence. Increasing people’s awareness of the role of unconscious bias in fueling workplace gender inequality and stereotypes and biases does not reduce their biased behavior.

In the simplest of terms, people can’t be trained out of their biases.

That said, diversity training programs designed to increase participants’ knowledge about bias and its harmful effects on women’s careers can be a highly beneficial part of a comprehensive and multifaceted program. It is useful for people to share a common language so they can talk about bias and the need for diversity and inclusion.



Many leaders remain unaware of major challenges in their current organizations. “They underestimate the obstacles confronting an employee of a diverse group, perceiving a workplace with far less bias than actually exists,” says DEI consultant Matt Krentz.

This lack of awareness is all too apparent from a survey of faculty, staff, and alumni at Stanford University’s Byers Center for Biodesign. The survey found that 80 percent of men believed their workplace “empowers women to reach their full potential,” while only 36 percent of women agreed. In addition, 84 percent of men believed their organization’s promotional criteria for women and men are the same, but only 35 percent of women agreed.

Moreover, men overwhelmingly attribute gender disparities within the leadership ranks to individual choices made by women and men. For example, it is a commonplace myth that women prefer to balance the needs of homelife and career. Women, by contrast, believe that they are held back by stereotypes and exclusion from important communication networks and influence because of systemic gender discrimination.

Because of such blind spots, senior leaders can fail to see, much less comprehend, the gender-based obstacles to women’s career advancement.

Because of the widely held belief that workplaces operate as meritocracies, senior leaders can believe that if women are not advancing in their careers as far and as fast as men, it is because women are doing something wrong.

The assumptions are that women are not confident enough; their behaviors are not forceful enough; and their career commitments are not strong enough.

In other words, these leaders can believe that women can succeed in their careers to the same extent as men if they have the will and display the commitment, attitudes, and behaviors for such success.

Such beliefs have led to the proliferation of programs designed to “fix the women, primarily by” making them more confident and capable of self-promotion. The common underlying premise is that women don’t get ahead because they lack confidence. This lack of confidence “is presented as being entirely an individual and personal matter, unconnected to structural inequalities or cultural forces.” A misguided belief.

This “fix the women” approach to dealing with gender inequality in leadership is reflected in the advice offered in the large number of books written by women who have been successful in their careers.

A look at these books reveals that the authors present a remarkably similar set of recommendations. To achieve career success, the authors argue that women need three things: (1) confidence, (2) control, and (3) courage.



With respect to confidence, Sheryl Sandberg (formerly of Meta Platforms) displays the common theme of these authors when she writes, “Women are hindered by barriers that exist within ourselves. We hold ourselves back in ways both big and small, by lacking self-confidence, by not raising our hands.”

While Sandberg and other women authors acknowledge that some gendered barriers exist, they emphasize that, for the most part, if women just become more confident, they will be able to overcome them.

As Michelle Mone, co-founder of Ultimo Lingerie, writes, “If you’ve got those ingredients [that is, passion, determination, and a ‘can-do’ attitude], nothing will stop you.”



Another common theme is that women can achieve career success if they can just remain in control at all times in their personal lives and in their work lives. Indeed, women can “have it all,” successfully combining a high-powered career and a satisfying homelife, as long as they exercise tight control over both spheres.

These authors thus present success as a matter of personal choice and responsibility: If women make the right choices, success can be theirs; if not, failure is their own fault.



On the topic of courage, Arianna Huffington, founder and CEO of Thrive Global and founder of The Huffington Post, is typical when she writes, “[Our female] fears of sticking our necks out because of how we’ll be perceived often causes us to sabotage our careers... If you want to succeed big, there is no substitute for simply sticking your neck out.”

These advice books and company programs designed to fix women obviously don’t consider the structural discrimination that women face in their workplaces or the individual discrimination they face every day at work.

Indeed, the implicit assumption of the entire “fix the women so they can succeed” thinking seems to be that structural and individual discrimination are inevitable, that they probably cannot be changed, and they are essentially irrelevant. Rather, women can succeed if they would take more risks and demand seats at the leadership table. The implication is that gender inequality is their fault and therefore women can end it all by themselves. It is hard to think of a more wrongheaded or perverse view of the state and cause of workplace gender inequality.


Because of the focus on anti-bias training and programs to fix women, leaders have shown little interest in addressing the structural forces driving workplace gender inequality. The pervasive assumption of diversity training is that gender inequality will end if people will just become more aware of their biases, and women will just try harder.

But organizations have been pursuing this approach for at least the last 30 years, and they have achieved little in bringing about real change in their leadership.

The failure to confront the structural discrimination in their workplaces and the workplace cultures that allow individual discrimination are the primary reasons we have made little progress to end gender inequality.

Increasing individuals’ awareness of their unconscious biases and teaching women to be more confident will not reduce—much less end—gender inequality. Without a program that directly attacks both structural and individual discrimination, company management is just playing around the edges of workplace gender inequality.

Most business, professional, and nonprofit organizations currently do diversity training of one sort or another, and many have programs specifically designed to help women become more confident and self-assertive. Yet, few of these organizations have programs to end exclusionary behavior in the workplace or to change the workplace’s personnel systems, processes, and practices.

These sorts of structural changes are the only real hope we have of confronting structural discrimination so we can achieve fair, equal, and equitable career outcomes for women and men.

As University of British Columbia researchers Toni Schmader, Tara Dennehy, and Andrew Baron note, “Although raising awareness is a clear and necessary prerequisite to fostering self-regulated unbiased behavior, awareness alone does not change culture.”

Diversity training focuses on the symptoms (individual discrimination) but not the causes of workplace gender inequality (structural discrimination). We need to directly confront both structural and individual discrimination if we are to move forward, reduce negative impacts, and strengthen positive operational outcomes and rewards across the entire workplace.


Adapted from Beyond Bias: The Path to End Gender Inequality at Work (Nicholas Brealey, May 16, 2023)
Copyright © 2023 Andrea S. Kramer and Alton B. Harris


We have updated our privacy policy. Click here to read our full policy.