On Implicit Bias: Part I

We’re starting off this month with a book that I read a few years ago, and recently reread: The End of Bias: A Beginning: The Science and Practice of Overcoming Implicit Bias (“The End of Bias“), by Jessica Nordell. As the first official book review of this blog, this book will lay the foundation for many of my future reviews and posts.

The End of Bias brilliantly explains the psychology and social impact of implicit bias (an attitude or stereotype towards a particular group that is not conscious or intentional, but harmful and impactful nonetheless). Drawing from numerous psychological studies and experiments, individual anecdotes, scientific research, and even her own personal reflections (on her experiences as a Jewish woman in STEM, as well as the ways in which she has found herself unintentionally harboring and perpetuating implicit biases), this book educates readers on the realities of implicit bias, while also offering practical solutions for decreasing the impact of these pervasive, pernicious habits.

Nordell starts her book by introducing the advent of implicit bias research, a relatively new field. At first, researchers were stumped by findings that showed people who consciously viewed themselves as unprejudiced behaving in prejudiced and biased ways. Further probing found that people (all people) absorb the cultural biases and prejudices of the society around them. These biases are then perpetuated by individuals through unconscious habits: “Devine deduced that people could consciously reject prejudice on one hand and behave in biased ways out of habit on the other… These were habits, she concluded, and people can engage in habits without thinking, just as someone might chew their fingernails down to the quick before realizing what they were doing.”

In many ways, this early research was groundbreaking. This body of research has found, for example, that biases can even be harbored by and towards people who we know and love: “Parents google ‘Is my son gifted?’ at two and a half times the rate they google “Is my daughter gifted?’.”

Additionally, these biases can also be harbored by people who are themselves harmed by these prejudiced habits. For example, women can behave in ways that are biased and prejudiced against other women; people of color can behave in ways that are biased and prejudiced against other people of color, including against other individuals within their own race or ethnicity: “Studies show that one need not be part of a dominant group in order to harbor harmful implicit associations.” Individual people unconsciously internalize our society’s long-held prejudices, regardless of the individual’s personal identity, experiences, or conscious thoughts and beliefs.  

While reading this book for the second time, I took the opportunity to reflect on the ways in which implicit bias can impact our work as investigators, and how knowledge of implicit bias can help us be better in our work. In these next two posts, I discuss three distinct modes in which we can use our knowledge of implicit bias in our investigations practice: (1) Understanding examples of implicit bias and microaggressions that can form the basis of a workplace complaint; (2) Eliminating implicit bias in the investigator in order to maintain neutrality and impartiality throughout the process; and (3) Identifying potential expressions of implicit bias in witnesses, and understanding how this will impact our credibility determination of the witness.

Because of the length of this review, I will discuss the first topic in the post below, and the next two topics in a separate post to be released on April 15, 2025, titled On Implicit Bias: Part II.

Examples of Implicit Bias and Microaggressions that Can Be the Basis of a Complaint

Knowledge of implicit bias, and the ability to identify common examples of implicit bias, can help us better assess complaints of harassment and discrimination. While these complaints often include allegations of more overt conduct, many complaints that I’ve investigated also include allegations involving microaggressions, or instances of more subtle, implicit bias.[1]

Assessing the “Reasonableness” of a Claim

First, knowledge of implicit bias and its prevalence in our society, along with examples of common instances of bias, can help us better determine the “reasonableness” of a complainant’s allegations, or their perception of bias. If the alleged conduct matches other common notions of stereotypes, microaggressions, and instances of implicit bias, the perception of this conduct as biased would be reasonable. For example, if a female complainant working in a male-dominated field alleges that the task of making coffee for her coworkers always presumptively falls on her, and that this is because she is a woman, her perception of bias in this instance is likely reasonable, given the common knowledge, stereotypical association between women and domestic or secretarial work. In contrast, if the alleged conduct does not match these common notions of implicit bias, the perception of this conduct as biased will appear more attenuated, less reasonable, and less credible.

Nordell’s book discusses several common stereotypes and biases that can inform our investigative work:

• Women as overemotional and hysterical: “Women have long been stereotyped as overemotional, agonized creatures with outsize responses, “hysterical” beings whose physical symptoms were manifestations of psychological problems. Research shows that adults watching a child express pain perceive the pain as less intense when the child is described as a girl.”
• Black men as dangerous or threatening: “Black men, the researchers point out, are seen as threatening and therefore evaluated negatively for behaving in a dominant, assertive way.”
• Asian women as especially passive: “One experimental study found, for instance, that dominant Asian women were more disliked as coworkers than dominant White women: they were violating pronounced stereotypes prescribing that they behave in passive and communal ways.”

Although the preceding list is by no means exhaustive, these excerpts present examples of the types of stereotypes we can keep in mind as we evaluate a complaint’s allegations. As discussed above, a complainant or witness’s perception of bias is more likely to be reasonable and credible if the alleged behavior aligns with common knowledge stereotypes and biases.

Moreover, as discussed above, someone need not be part of a dominant group to harbor these types of biases. Thus, we may have a female complainant alleging sexual harassment or gender discrimination against another woman; or a Black complainant alleging racial discrimination against another Black employee. The idea that these types of claims do not exist, or that they are inherently implausible and unreasonable is a common misconception. An understanding of implicit bias research allows us to not be as surprised by these types of claims.

Honing our Investigatory Skills

Knowledge of implicit bias can also allow us to probe these allegations more effectively. An increased contextual knowledge of the ways in which bias manifests in different situations, and the modes through which these processes have been studied and analyzed, will enable us to hone our interviewing skills, and ask questions more adequately aligned with the nature of the allegation.

Interviewing similarly situated people within the workplace (even if they are not necessarily direct witnesses to the alleged conduct); asking follow-up interview questions relating to potential patterns of bias; and analyzing aggregate workplace demographics and similar data are all useful techniques for assessing elements of implicit bias, beyond the more overt aspects of the allegations.

Again, this type of evidence can support or contradict a claim of discriminatory behavior. For example, if other women on a particular team experience gender-based microaggressions from the respondent (similar to those alleged by the complainant), the likelihood that the respondent’s conduct occurred and was gender-based will increase. Asking the complainant questions regarding why they felt or believed that something was discriminatory or harassing will also help probe these allegations more deeply and effectively. And looking at workplace demographics overall (such as how many other women are on the team, how many are in positions of power, and how many report to the same supervisor) will provide further insight into the reasonableness of the complainant’s perception, and even potentially inform our investigation plan (by suggesting other witnesses to interview).

Implicit Bias Evidence in Litigation

As I’ve discussed above, knowledge of implicit bias can help us better tailor our investigation plan and processes; and can help us assess the “reasonableness” of a person’s claim (which will ultimately lend itself to that person’s credibility). In making our final determination, however, it is also important to understand the limitations of implicit bias evidence.

In her book, Nordell discussed two legal cases where implicit bias evidence was insufficient in holding an employer liable for discrimination: Ellen Pao v. Kleiner Perkins Caufield & Byers LLC (2015), which was tried in the San Francisco County Superior Court, and eventually dropped before appeal; and Wal-Mart Stores, Inc. v. Dukes (2011), which was tried in the United States Supreme Court.

In both cases, the plaintiffs argued that they had been discriminated against based on their gender by alleging numerous instances of implicit bias. First, in the landmark 2011 United States Supreme Court case Wal-Mart Stores, Inc. v. Dukes, Justice Scalia, writing the majority opinion for the U.S. Supreme Court, held, essentially, that these daily biases did not add up to a large enough harm. The lawsuit was filed on behalf of 1.6 million female Walmart employees, and alleged that Walmart engaged in a series of practices that led to an indirect, disparate impact on women, including denying women promotions, paying them less than male workers, and steering them to low wage positions. As evidence of their alleged discrimination, the plaintiffs cited aggregated data showing gender disparities in pay and promotions across the company; and provided statements from 120 women alleging instances of bias or discrimination, including being told to “doll up,” doing the work of a manager without the title or recognition, and being penalized more heavily than male employees for the same or similar infractions.

In this case, Scalia determined, however, that these myriad incidents could not be attributed to a single manager. Scalia concluded, in essence, that individual managers could have never been able, in the aggregate, to produce the kind of pay and promotion gaps seen at Walmart; there must have been other factors (such as male and female employees having different qualifications) that gave rise to these disparities. In other words, Scalia found that the connection between the aggregate effect on female employees and the actions of individual managers (who would confer liability onto the employer, Walmart) was too attenuated; there were too many other potential reasons for the disparities that had not been successfully disproved and whittled down.

Then came Ellen Pao. Pao was a junior partner at Kleiner Perkins Caufield & Byers, a powerful venture capitalist firm. In 2012, Pao was terminated and she immediately sued the company for gender discrimination, alleging, specifically, that the company had blocked Pao and other women from being promoted and paid fairly, and blocked opportunities for them to advance professionally and succeed. She alleged, for example, that she had been asked to take notes at a meeting and excluded from a high-profile networking event because women “kill the buzz.” Pao further alleged that her ideas and contributions were dismissed and overlooked; that she was often criticized for either being “overly opinionated” or “needing to speak up more”; and that her male coworkers took credit for her work, among other things. Similarly, the jury in this case found that gender had not played a substantial role in her termination.

As these two cases illustrate, cases involving implicit bias evidence alone are unlikely to be successful in litigation; rather, most cases that lead to liability for the employer focus on more egregious specific harms endured by specific people, or the disparate impact of a specific employer policy or practice. As Scalia found, cases that cite a diverse myriad of microaggressions occurring at different times and places and with different people are too vague and attenuated.

The author of this book takes issue with these legal holdings and seems to believe that the law should progress to incorporate a more sophisticated understanding of implicit bias: “The problem is with the way bias is typically assessed… They document reality at a particular moment. What they do not do is account for how bias is actually experienced in the real world, where individuals are not the targets of prejudice once, or twice, or three times, but experience it continuously over weeks, months and years. Just as one photograph (or even a series of photographs) cannot fully capture the trajectory of an object in motion, studies that take a snapshot of bias do not capture its actual effect on people over time.” With respect to the Walmart case, Nordell also cheekily comments, “Reading this opinion, one comes away with the distinct impression that Scalia had never had a job.”

For employers, however, these holdings are likely a relief. Indeed, if these implicit biases exist pervasively throughout our entire society, is it fair to hold individual employers uniquely responsible for them? So far, our courts don’t seem to think so.

Weighing Implicit Bias Evidence in Investigations Findings

So, why should employers be concerned with implicit bias, if implicit bias evidence is unlikely to lead to employer liability in litigation? While employers can be held liable for the unlawful conduct that transpires in their workplace after the fact, employers also have a duty to prevent unlawful conduct before it occurs. Employers are therefore required to investigate all allegations of potential discrimination, harassment, or other possibly unlawful conduct, even if the factual allegations, on their face, may not appear to rise to the relevant legal standard. Thus, if an employee makes a claim of implicit bias discrimination, the employer still has a duty to investigate the claim, and to make efforts to prevent future harm. Indeed, where implicit bias runs rampant, the likelihood of more severe conducting occurring – conduct that would likely confer legal liability on the employer in litigation – increases.

Moreover, workplace investigations – that generally function as preventative, pre-litigation processes – operate under a “preponderance of the evidence” standard, meaning that if a claim is more likely true than not, that allegation will be sustained. This standard of proof (also known as “fifty percent and a feather”) is lower than most legal standards in litigation. While litigation determines remedial action (typically in the form of monetary relief, which, in California, can equate to multi-million dollar verdicts), workplace investigations determine personnel decisions (such as employee discipline). Essentially, the stakes are lower in a workplace investigation. Thus, under this lower standard of proof, conduct that may not necessarily rise to the level of legal liability in litigation, can still lead to a sustained finding of misconduct in a workplace investigation. A workplace investigation prevents future harm (and potential future legal liability in litigation) by stopping the unlawful conduct in its tracks, and allowing employers to make fair, objective personnel decisions in response to that conduct (if its found to have occurred).

Indeed, in an investigation, we can certainly consider evidence of implicit bias and this evidence may in fact tip the scale towards a sustained finding of race-based or gender-based conduct. However, the evidence must still relate specifically to the accused. An argument that all of society and the entire workplace is permeated by implicit biases, and the respondent, therefore, was likely to have been acting on these biases, would likely not be sufficient (without further evidence) for sustaining a claim. In contrast, credible evidence of repeated microaggressions (such as biased comments) towards the complainant by the respondent is likely to lead to a sustained finding.

Conclusion

Overall, the science of implicit bias is still relatively new, and our workplace laws are always evolving. As investigators, we should stay up to date on these developments and continue assessing the best ways in which to incorporate these ideas into our practice, especially as our common-knowledge understanding of implicit bias increases, and employment attorneys (on both sides of the bar) become more sophisticated in incorporating these ideas into their legal arguments.

In On Implicit Bias: Part II, which will be published on April 15, 2025, I will discuss the ways in which investigators can reduce implicit bias in themselves, to attain greater neutrality; and the ways we can identify implicit bias in witnesses, and how this might affect their credibility.

[1] The Merriam-Webster dictionary defines a “microaggression” as “a comment or action that subtly and often unconsciously or unintentionally expresses a prejudiced attitude toward a member of a marginalized group (such as a racial minority).”