Unconscious Bias Training That Works

Modern theories of racism and sexism propose that we simultaneously hold unconscious prejudices and strong pro-equality ideals. Our unconscious bias develops mostly from biased media representations, but also from our social experiences. As we are repeatedly exposed to stereotypical associations and prejudices from an early age, these become automated in our long-term memory. At the same time, the outward expression of internalised bias is curbed by strong social norms and legal restrictions against expressed prejudice and discrimination.

The result is a subtle, hidden, and often unconscious negative bias towards particular social groups. Even individuals who truly believe they are non-prejudiced hold unconscious biases. Studies show racism and sexism manifest as prejudiced attitudes and discriminatory behaviour in individuals who explicitly endorse egalitarian values, particularly when there is not enough time to regulate responses.

A disturbing example involves a series of experiments in which participants played a video game during which an individual, who was sometimes White and sometimes Black, appeared spontaneously carrying either a gun or a different, non-threatening object. The participants in the experiment were told to ‘shoot’ when the intruder was carrying a gun, but to press another key if the intruder was carrying a different object. The results showed that the number of times the participants accidentally perceived the object to be a gun when it was not, was much higher when the intruders were Black, compared to when the intruders were White. The results were similar for White and Black participants, indicating how negative stereotypes can exist intragroup as well as intergroup.

Unconscious bias is also more likely to manifest as discriminatory behavior when it can be justified or rationalised on non-discriminatory grounds. In one example of modern racism, participants were similarly likely to help a Black (94%) or White victim (81%) in need of medical assistance when they believed they were the only person who could help. However, when they believed that others were present who could assist, they helped the Black victim much less (38%) than they helped the White victim (75%). In the latter case, claiming that they believed others would help, meant the participants could legitimately mask their discrimination.

The phenomenon of hidden bias manifesting as discriminatory responses when it can be justified on non-discriminatory grounds has also been demonstrated in the assessment of college applications. When both Black and White college applicants rated consistently above or below par on application criteria, applications were rated similarly by assessors. But when college applications were ambiguous, such that applicants excelled on some criteria and performed below average on others, prejudiced assessors rated the Black applicants less favourably than the White applicants. The assessors masked their discriminatory responses by claiming that those criteria on which the Black applicants were weak were more important than those criteria on which those applicants excelled.

In the workplace, unconscious bias in recruitment, selection, promotion, development, and everyday workplace interaction creates inequality, establishes social and emotional distance between different groups, increases legal and reputational risk, and threatens employee well-being, engagement, commitment, and productivity. Unconscious bias limits the strategic potential that that flows from a diverse workforce for problem solving and decision-making, innovation and creativity, accessing diverse customers and suppliers, and attracting and energising top global talent. 

Addressing unconscious bias at work is problematic because, by its definition, an individual will most likely be unaware of its influence. 

Overriding Unconscious Bias

Fortunately, organisations seeking to reduce unconscious bias have a wealth of social psychological theory and research from which to draw upon when designing interventions.

Human beings process social information by two routes. One route is automatic, unconscious, and largely driven by emotional factors. The other route for processing social information involves controlled and conscious thought or reflection.

Unconscious bias is an example of automatic processing. In social settings, stereotypes and associated prejudices and discriminatory responses occur fast and outside conscious awareness.

However, research shows automatic and unconscious mental responses can be overridden with conscious and deliberate effort. A person who is motivated to be unprejudiced—because of legal sanctions, social pressures, or strong personal egalitarian values—can suppress biased responses.

Increasing One’s Motivation to Respond Without Prejudice

Unconscious bias training seeks to motivate individuals to override their automatic biased tendencies by confronting individuals who hold explicit egalitarian values with evidence of their hidden biases.

Cognitive dissonance refers to the uncomfortable emotional state experienced when individuals are made aware of an inconsistency in their beliefs, attitudes, or behaviours. Research indicates that when egalitarian values are central to an individual’s self-concept, highlighting an inconsistency between the individual’s anti-prejudice values and their biased responses is effective at evoking dissonance. In turn, dissonance motivates the individual to make conscious adjustments to their attitudes (reduction in prejudice) and behaviours (less discrimination) such that they better align with their strongly held values of tolerance and equality.

One popular technique for enhancing awareness of one’s unconscious bias is the Implicit Association Test (IAT). This test measures the reaction time of individuals to a series of words or pictures presented on a computer screen. For example, the individual may be asked to type a particular key if the word presented on the screen is a ‘female name’ or a ‘weak word’ (e.g., delicate, small, flower) and a different key if the word is a ‘male name’ or a ‘strong word’ (e.g., powerful, mighty, robust). This activity is repeated numerous times and the average reaction time for a correct response is recorded. 

Following this, the rules are changed such that the test taker is asked to press one key if the word is a ‘female name’ or a ‘strong word’, and a different key if the word is a ‘male name’ or a ‘weak word’. Because gender stereotyping associates female names with weak words, and male names with strong words, reaction times on the first test are relatively faster compared to the reaction times under the conditions of the second test involving a mismatch of stereotypical categories. Differential reaction times are evidence of implicit (unconscious) gender bias, and the greater the difference in reaction times between the two tests, the greater are those implicit stereotypical associations. 

Anonymous unconscious bias tests administered by Harvard University are publicly available at https://implicit.harvard.edu/implicit/takeatest.html. Over a million people have taken those tests, and results confirm that participants across a range of locations, ages, genders, races, and ethnicities hold unconscious stereotypes and prejudices regarding disability, sexual orientation, race, skin tone, age, weight, gender, ethnicity, and religion.

Practitioners should be aware, however, that there have been varied results from the use of this tool in real-world settings. Problems may arise because the theory behind the IAT is difficult to understand and participants may misinterpret the results...leading to confusion, shock, anger, and defensiveness.

When the IAT is used as an intervention tool, it is important that the facilitator is knowledgeable in the mechanisms of the IAT and adequately explains to participants that bias is inevitable as a result of social conditioning and cognitive processes—the results do not evidence or make accusations of prejudice. Rather, the facilitator must stress that exercise is undertaken to highlight the existence of hidden bias and that, contrary to our conscious intentions, we all hold hidden biases that manifest in subtle and unconscious ways.

A TEDx presentation by UCLA law professor Jerry Kang is useful for helping individuals understand the theory behind the IAT.

Bias Awareness Training Activities

In addition to the IAT test, there are a number of other activities grounded in social psychological theory that can be incorporated into unconscious bias training without triggering resistance or backlash, and that are simple to administer and understand.

One example is the Tag Game, adapted from Fowler (2006). In this exercise, participants stick badges, in a variety of shapes, colours, and sizes, somewhere between their waist and neck.

Participants are then instructed to form groups without talking. There are no instructions given as to what criteria to use to form the groups. Once formed, the participants are instructed to break up and form into new groups. This is repeated at least four times.

Participants will normally form groups based on shapes, colours, or sizes. Rarely do the participants look beyond the badges, and even less rarely do they intentionally form diverse groups in which many shapes, colours, and sizes are represented.

This powerful yet non-confrontational activity leads well into a discussion about social categorisation processes, the automaticity of “us” vs. “them” categorisations, and ingroup bias—our tendency to cluster together in groups with people similar to ourseves. 

It is also an excellent exercise for introducing the concept of diversity and the potential benefits of diverse workgroups. Group discussions following the exercise explore diversity experiences (or lack thereof) in the workplace, and prompt participants to suggest ways improve the recognition, support, and value of diverse perspectives and experiences.

Another useful awareness activity for unconscious bias training from the social psychological literature is the Father/Son activity, adapted from Pendry, Driscoll, & Field (2007). In this activity, participants are instructed to solve the following problem:

“A father and son were involved in a car accident in which the father was killed and the son was seriously injured. The father was pronounced dead at the scene of the accident and his body was taken to a local morgue. The son was taken by ambulance to a nearby hospital and was immediately wheeled into an emergency operating room. A surgeon was called. Upon arrival and seeing the patient, the attending surgeon exclaimed “Oh my God, it’s my son!’ Can you explain this?”

Around 40% of participants who are faced with this challenge do not think of the most plausible answer—being the surgeon is the boy’s mother. Rather, readers often invent elaborate stories such as the boy was adopted and the surgeon was his natural father, or the father in the car was a priest.

As such, the exercise illustrates the powerful pull of automatic stereotyped associations; for some individuals, the association between surgeon and men is so strong that it interferes with problem solving and making accurate judgments.

This exercise leads well into an ensuing discussion on the automaticity of stereotypes and the distinction between explicit and implicit bias. From here, the discussion can move to exploring ways of controlling or overcoming automatic bias. In addition, because some of the participants will solve the problem with the most plausible reason, the exercise highlights individual differences in stereotyping and opens a discussion into why stereotypes differ across individuals. 

Building Skills for Managing Unconscious Bias

As a stand-alone initiative, awareness programs are rarely effective tools for reducing bias. Negative stereotypes and prejudices are maintained and reinforced by powerful cognitive and motivational biases that act to filter out information that contradicts or challenges our preexisting beliefs or attitudes. 

In fact, interventions may be more effective if approached from the other direction—behavior often has a greater influence on our attitudes than our attitudes have on behaviours.  For this reason, it is critical that awareness training is supplemented with skills-based training designed specifically to change behavior in the workplace. This has not often been the case, however, with a large proportion of interventions focused on attitude change, assuming that behavioural change will follow.

In contrast, best practice unconscious bias training includes training in skills for overriding our automatic tendency for bias, for example, perspective-taking, making culturally appropriate attributions, and counter-stereotypical imaging. 

Perspective-Taking

Neurocognitive studies have demonstrated that we all possess an ability to mimic automatically the emotions, thoughts, and actions of others. From an evolutionary perspective, this ability to synchronise our intentions and behaviours with others enhances our social functioning that, in turn, increases our chances of survival.

We automatically simulate the mental and motor activity of ingroup members. This process, however, is much less responsive, to outgroup members. But with conscious effort, we can make up for this deficit in intergroup sensitivity. One way we can do this is through perspective-taking.

Perspective-taking refers to the ‘active contemplation of other’s psychological experiences’—that is, thinking and imagining the feelings and viewpoints of others.  It is now well established that perspective-taking has positive implications for intergroup relations including increased empathy, reduction of unconscious prejudiced attitudes and discriminatory behaviours, and decreased activation of negative stereotypes. 

Perspective-taking works by enhancing ‘self-other overlap’—the merging of one’s cognitive representations of their self-concept with outgroup members. In this way, social category boundaries become blurred and “they” become “me”.                                                                            

Outgroup members are now perceived to be more like the self, and are thus more likely to be afforded the positive favoritism usually reserved for ingroup members. This effect can generalise to other outgroup members because the target member is perceived as a prototype for the wider group.

Making Culturally Appropriate Attributions

While perspective-taking for empathy focuses on the emotions of the other person without reference to cultural differences, perspective-taking with intercultural awareness focuses on understanding the other person’s perspective through their cultural frame of reference.

The active consideration of alternative worldviews helps you transcend the automaticity of your cultural framework. Imagining the world from the perspectives of others helps you to interpret their intentions and behaviours more accurately, reducing the potential for misunderstandings, misattributions, and conflict. In addition, differences are less likely to be labeled as “deviant”, stereotyping and bias is reduced, communication is enhanced, and suspicion and distrust are minimised.

Prompts can help individuals to improve their cultural perspective-taking by encouraging them to reflect on how culture might be affecting their counterpart’s values and beliefs, for example,

‘Before you decide how to respond in this interaction, write down a few sentences describing your counterpart’s interests and concerns as a person living within their culture. Now consider how your counterpart’s behaviours and decisions in this situation may be guided by his or her cultural values and beliefs’.

Prompts are useful when planning for intercultural interactions. They can also be used to encourage reflective monitoring in real-time, during your exchanges with others.

Counter-Stereotypical Imagery

Multiple studies have shown that the accessibility of automatic and unconscious stereotypes is reduced by engaging in short 2-3 minutes counter-stereotypical imagery. In one example, being instructed to imagine a strong women led to less accessibility of an automatic 'weak-women' stereotyped association as measured in the IAT.

Within a training environment, counter-stereotypical imaging can be introduced as a useful unconcious bias intervention. A suggested approach is to have participants reflect briefly on a cultural identity group that they may be experiencing difficulty with (or have done so in the past) and to note this group on a piece of paper. The participants should not be given too long to think about this—the intention is to flag stereotyped groups without activating negative stereotypes. 

The participants are then instructed to identify an exemplar from this cultural identity group. The exemplar may be famous, or someone known to the presenter such as a friend or colleague. The description should focus on positive qualities and avoid comparing the exemplar to the negative stereotype—writing a short bio is a good way to avoid this trap. The intention is that the presentation of positive qualities will strengthen counter-stereotypical associations, and in turn, moderate the automatic activation of negative stereotypes.

Participants should be reminded that stereotyping and intergroup bias is inevitable and that being able to identify outgroups with which we have trouble interacting does not make us prejudiced. The purpose of this exercise is to improve intergroup relations by moderating our automatic beliefs and judgments. 

This is a useful technique that participants can take away from the training and engage when needed to help them improve their interactions with diverse groups. 

Workplace Cultures That Support Unconscious Bias Training

While training is a critical and necessary component of an organisation’s efforts to reduce unconscious bias, companies committed to reducing unconscious bias can also draw upon social psychological literature to design work settings that temper the automatic activation of stereotypes, prejudice, and discrimination.

Workplace cultures that support self-worth, offer opportunities for positive intergroup contact, promote employee well-being, and encourage and reward pro-equality norms, support unconscious bias training and encourage the transfer of knowledge and skills acquired in training to the workplace.

In a subsequent post, I examine the first of those suggestions—workplace cultures that support self-worth.

The original version of this article with research citations can be found here. 

About the Author

Felicity Menzies is an authority on cultural intelligence and diversity and inclusion in the corporate environment. Felicity’s interest in the role of culture in business began during her tenure as the Head of Private Bank, Westpac, in Singapore. There she led a culturally diverse team of bankers serving a multinational client base and learned firsthand how cultural intelligence is a necessary component of business success. 

As Principal at Culture Plus Consulting, Felicity now applies her business acumen and intercultural expertise to help global organisations respond effectively to the opportunities and challenges presented by diverse workforces, unfamiliar markets, and rapid shifts in the global competitive landscape.