Bias Training: Why Starbucks is Right

It is now a familiar story: On April 12, two young African American men were arrested in a Philadelphia Starbucks, apparently for waiting while Black.  Starbucks, in relatively little time, issued an apology and announced that they will close US stores on May 29 for racial bias training of approximately 175,000 employees.

Diversity training has existed in American workplaces for decades; unconscious or implicit bias training is becoming the new norm.  These trainings rest on the assumption that we all have deeply embedded attitudes or associations of which we are not entirely aware and that can lead us to act in biased ways.  In the last few years, the notion of implicit bias has moved beyond purely academic circles and invoked to explain a wide range of events, from the shooting of unarmed black men to Hillary Clinton losing the presidential election.  Implicit bias has also been implicated in the workplace, a contributor to the dearth of women in leadership positions, the underrepresentation of women in technology fields and the slow progress of African Americans in the corporate landscape. Given all of this, it is not entirely surprising that an increasing number of companies are offering implicit bias training.

But an air of skepticism hovers like fog over implicit bias training.  Attention grabbing headlines refuting such trainings abound (“How Diversity Training Infuriates Men and Fails Women” [Time, 2017] and “Starbucks Will Give Employees Unconscious Bias Training. That May Not Help.” [Daily Beast, 2018].  And Starbucks’ very public decision to conduct bias training in response to this incident has led to renewed questioning of the utility of implicit bias training.  Is bias training useful for anything beyond helping corporations maintain a good public image? Can trainings like this actually reduce bias and discriminatory behavior?

To me, the obvious answer is yes.  Before I get into why, let me reveal some of my own biases.  I am a social psychologist by training, who came of age in social psychology in the early 2000s. During this time, the idea of unconscious influences on behavior broadly, and implicit attitudes more specifically, captivated the field. I fully acknowledge that my perspective as a social psychologist influences my opinion on this topic.  I nonetheless believe the criticism does not warrant dismissing implicit bias training.

One major criticism of implicit bias training is that it does not work. This evidence often takes the form of pointing to persistent inequalities that exist in the workforce despite implicit bias training.  For example, at companies like Google and Facebook, early adopters in the tech world of implicit bias training, technical workers are still predominantly male.  Some also point to recent academic research that indicates that changes in implicit bias does not necessarily lead to changes in behavior.  A recent meta-analysis (i.e. a quantitative summary) of approximately 500 studies demonstrated that while interventions can change implicit attitudes, the effects were often weak and short-term (Forscher et al 2016).  The authors note that many of the procedures designed to change implicit attitudes were brief and were experienced during a single session.  Changes in implicit attitudes from such sessions did not lead to corresponding changes in behavior.

To this criticism I respond: is it really that surprising that a one-time training doesn’t eradicate centuries of history and learning that contribute to implicit bias?  Implicit bias are actions or judgments that result from automatic feelings or thoughts. These automatic feelings and thoughts, i.e. attitudes, develop over time based on both direct (e.g., interactions with people from a particular group) and indirect experiences (e.g., seeing members of a particular group on television, observing other’s interaction with members of a particular group).  Although an implicit bias course may start to challenge stereotypes and break down these implicit attitudes, it can’t eradicate these long held attitudes that our culture still reinforces in many ways.

Patricia Devine, a social psychologist doing work in this area uses a habit framework to examine the relationship between implicit attitudes and behavior (Devine & Monteith, 1993; Plant & Devine, 2009).  Our culture has led us to develop a habit, where because of repeated direct and indirect experiences with men as leaders and women as subordinate, we for example, automatically respond to the male in the group as the leader. Getting to the point where that is not our automatic reactions is not easy.

Here is analogy that puts this habit framework into concrete terms that many can likely relate to. You have a sweet tooth, which you indulge with dessert after dinner on a nightly basis.  After a recent visit to your doctor in which she told you that your blood sugar is a bit high and you are pre-diabetic, you realize you need to eat more healthfully.  You know you have an issue and you want to change.  Nonetheless, that one conversation with your doctor is unlikely to be sufficient to change your eating habits.  Initially, there will be nights when you find yourself hankering for ice cream.  On some nights, you will be successful in resisting giving in to your urges.  But on other nights, likely when you’re especially tired, or maybe a little sad, you will find yourself on the couch, with a bowl of ice cream.

Similarly, as we apply this to thinking implicit bias training, the fact that there hasn’t been a complete and immediate change in behavior doesn’t mean the information isn’t worthwhile.  It means that we acknowledge that is just the beginning and that we keep working at it. Devine argues that in order to break the bias habit, people must: 1) know about their implicit attitudes 2) know when those attitudes are likely to lead to biased responses; and 3) be motivated to change.

Changing people’s implicit negative attitudes is one long-term goal of implicit bias training.  But, the notion that change in behavior is only possible if these implicit attitudes change is mistaken.  We may still have that hankering for ice cream, but the fact that we no longer eat it every night is progress.  Thus, a more immediate and realistic goal is to make us more aware that we have these automatic associations and attitudes, be more conscious of how they may affect our behavior, and becoming more committed to behaving in unbiased ways. And there is evidence indicating that implicit bias training can help people do that.  Devine and colleagues (Carnes et. al 2015) conducting implicit bias trainings in a number of academic departments.  The training improved implicit attitudes related to women and leadership.  Perhaps more importantly, in the 43 departments that went through the training, the course led to a greater commitment to promoting gender equity and led to an improvement in department climate.  Thus, training can lead to changes in explicit attitudes and environmental factors that can facilitate different outcomes.

Not all courses are created equal. The impact of implicit bias training will be influenced by the specifics of the course and the organizations. If you are interested in conducting implicit bias training that leads to a more diverse and inclusive workplace, where there is less prejudice and discrimination, here are some recommendations for successful implicit bias training.

• Training should be voluntary. As with any habitual behavior where change is difficult, the person needs to be motivated to change.  A basic principle from social psychology is that we are most likely go beyond our implicit attitudes when we have mental energy (e.g. not tired, hungry, stressed) and when we are motivated.  Employees who have no interest in examining their own attitudes or behavior or unlikely to benefit from such courses.
• Training should provide people with concrete skills to behave differently. Implicit attitudes affect behavior in part because it’s automatic; it’s easy.  To undo that, people will need alternatives and help executing them. For example, providing employees with interviewing skills that helps them uncover important information from candidates rather than relying on their gut, which is of course highly subject to bias.
• Training should be part of a broader range of initiatives. Creating an environment that is open to acknowledging and challenging bias will support and facilitate individual changes in behavior. One of the most effective ways for companies to do this is by putting systems in place that reduces reliance on biased thinking.  For example, in the case of Starbucks, creating a bathroom policy that does not require an employee to judge whether a person is a likely customer or a trespasser, is one way to reduce the impact of implicit racial bias.   As another example, creating clear, objective, and explicit criteria prior to meeting and evaluating candidates is another.

Implicit bias training isn’t a panacea.  It won’t get rid of all disparities in outcomes.  It won’t immediately eliminate biased behavior. But being more mindful of our automatic thoughts and feelings, and increasing our commitment to being better is certainly a good start.