Understanding and Interrupting Bias

People who identify as women and/or members of minoritized groups continue to face bias in the scientific workplace that impedes their recruitment, advancement, and success. Interrupting bias is essential to diversify science, broaden participation, and improve scientific innovation, productivity, and impact. But what exactly is bias and how do we interrupt it?

What is bias?

Bias, simply put, is a preference or tendency to respond one way versus another. Bias consists of prejudice (i.e., biased emotional reactions, such as a gut feeling or general like or dislike), stereotypes (i.e., biased beliefs about group characteristics), and discrimination (i.e., biased behavior) toward a person or group based merely on their group membership. Bias can arise from any social category, including gender, race, ethnicity, age, ability status, sexual orientation, gender expression or identity, weight, attractiveness, socioeconomic status, education, hometown, and any other dimension on which we categorize people. These categories intersect to create unique experiences of bias that are often compounded for individuals who belong to multiple minoritized groups. Overall, bias can distort our perceptions of people and inadvertently lead us to make poor decisions, treat people unfairly, and miss opportunities to build positive relationships and collaborations with others.

Barriers to detecting and interrupting bias

Several factors make bias difficult to detect and interrupt without deliberate intervention. First, bias often operates without our intent or awareness, fueled by unconscious stereotypes that work behind the scenes to guide how we process information, make decisions, and interact with others (Cundiff 2020). Second, bias can operate subtly and ambiguously when plausible explanations unrelated to bias are available to justify differential outcomes in hiring, promotion, or pay, giving the illusion of objectivity. As a result, bias is often only visible when looking at patterns across multiple cases rather than single incidents (Crosby et al. 1986). Third, bias often occurs as minor events that are disregarded as unimportant. However, small inequities can accumulate over time to negatively impact career trajectories and create large disparities (Shandera et al. 2021, Valian 1998).

Certain circumstances can make us more vulnerable to bias. Ambiguous situations, such as when candidates’ qualifications are unclear or when evaluation criteria are ill-defined, provide ripe opportunity for stereotypes to unwittingly fill in the blanks and distort our perceptions (Dovidio and Gaertner 2000, Uhlmann and Cohen 2005). We are also more likely to rely on stereotypes when we don't actively monitor our decisions and behaviors for potential bias, such as when we are mentally exhausted, under time pressure, or unmotivated to do so because we think we are immune or think that bias does not exist (Martell et al. 1991, Régner et al. 2019). In fact, people who believe they are least vulnerable to bias are, ironically, most vulnerable (Uhlmann and Cohen 2007).

Strategies for interrupting bias

Good intentions are not enough to avoid bias; instead, we need effective strategies. I came together with Asia Eaton, Ti'Air Riggins, and a working group at the AIBS 2021 IDEA Conference. The AIBS IDEA conference is an effort to enable scientific societies to create inclusive, diverse, equitable, and accepting (IDEA) environments. The purpose of the conference is to directly engage scientific society leadership in working together to understand the components of IDEA cultures, describing the current culture in biology, identifying barriers to enabling IDEA environments, creating action plans to overcome the barriers, and collaborating in mutual commitment to move forward in changing the culture of biology. During the first conference experts and scientific society members convened to discuss evidence-informed strategies that scientists and scientific societies can adopt to interrupt bias.

Track data

Bias is most visible when looking at patterns aggregated across multiple cases rather than single incidents. One study showed that evaluators were 16% more likely to correctly identify gender bias in pay when presented with salary data from multiple cases all at once (aggregated) than when the information trickled in case by case (Crosby et al. 1986). Each individual case could be explained away by some factor other than gender, such as individual differences in education or seniority, but these idiosyncratic differences could no longer explain away the pattern of inequity revealed by the aggregated data. Organizations must collect data—and lots of it—to identify patterns of inequity and track progress on measurable goals related to hiring, salary, promotions, awards, publications, honors, society membership, harassment, and other areas potentially affected by bias. Because bias can be compounded for individuals who belong to multiple minoritized groups, data should be collected and reported in ways that allow for intersectional analyses.

Use objective evaluation criteria

Research demonstrates that when hiring criteria is ill-defined, evaluators will shift the criteria to match the specific credentials that a desired candidate happens to have; however, committing to criteria prior to viewing applications eliminates the bias (Uhlmann and Cohen 2005). Therefore, evaluation criteria should be clearly defined—using specific, measurable, and relevant metrics—prior to reviewing applications. In addition, evaluators should standardize ratings by using rubrics and should provide justification for their ratings. For example, when one organization tweaked their evaluation process by replacing vague criteria with standardized rating forms that measured specific job-relevant competencies backed by evidence, they saw a vast improvement in evaluations of White women and People of Color (Williams et al. 2021).

Confront bias

Many people like to think they would confront bias, but in reality, they shy away (Kawakami et al. 2019). Confrontation can be scary, but it doesn't have to be. How you confront matters: Research suggests that, to be effective, confronters should avoid hostile, aggressive, threatening, and extreme comments (Monteith et al. 2019). Instead, share your thoughts and feelings, encourage perspective taking, provide information that counters stereotypes, ask questions, and engage in respectful conversations that assume good will and appeal to shared egalitarian values (Byrd 2018, Moors et al. 2022). Confronting works: In one research study, participants who were confronted about their stereotypic comments were less likely to rely on stereotypes in a subsequent task than were those who had not been confronted (Czopp et al. 2006; see also Chaney and Sanchez 2018, Burns and Granz 2021, Chaney et al. 2021).

It's also important to confront bias in yourself by being mindful (Oyler et al. 2022): Question your initial reactions, pause and reflect on how unconscious bias could be affecting your judgments, and consider how to respond without bias. Social psychologist Margo Monteith and colleagues (2015) called this the self-regulation of prejudice: When people become aware that their behavior conflicts with their own egalitarian values and they reflect on how to respond without bias, they can successfully inhibit biased responding in the future. Of course, we don't always notice our own biased behavior, and we need others to bring it to our attention. When others call out your bias, apologize and thank them for making you aware.

Foster belonging

I have heard countless stories from talented young women who have opted out of science, technology, engineering, and math (STEM) careers not because they didn't have what it takes or weren't interested in STEM but because they felt they were not valued, were not accepted, or did not belong. One student summed it up: “I don't want to have to become one of the guys.” A sense of belonging is a fundamental social need that drives human behavior. When individuals feel valued and connected to others, they experience increased interest, motivation, and confidence. One of the most influential experiences in my career was attending a 2-week residential program during graduate school with students from other institutions. The program catalyzed long-lasting social connections with my peers and garnered a sense of belonging that has helped me persist in my career. Indeed, research studies show that providing opportunities for budding scientists to build authentic relationships increases their belonging, confidence, and motivation to persist in science (Thiem and Dasgupta 2022). Institutions and professional societies can fulfill belonging needs by supporting affinity groups, peer mentoring programs, summer institutes, networking events at conferences, and other opportunities for scientists to build relationships and community with their peers.

These types of opportunities can be particularly effective when they connect underrepresented scientists with role models who share their gender and racial or ethnic identity, interests, and experiences—that is, when role models are relatable (Dasgupta 2011). Relatable role models send a powerful signal to minoritized groups that people like them are valued and welcomed in science and help buffer against negative stereotypes and bias. Strategies for increasing exposure to diverse, relatable role models include creating programs that connect faculty to same gender and same race or ethnicity mentors at other institutions, inviting underrepresented scholars to serve as panelists and speakers on campus and at conferences, featuring achievements of diverse scientists in textbooks and course content, and using a cohort approach to faculty hiring (Sgoutas-Emch et al. 2016). Male and White allies can say no to serving on all-male or all-White panels, and conference organizers can use resources such as the Request a Woman Scientist tool at 500womenscientists.org, the Databases of Databases of Diverse STEM Speakers, and the Conference Diversity Distribution Calculator to assess and diversify speaker lists (Witze 2019, Perkel 2020).

In addition to creating opportunities for social connection, institutions and professional societies can facilitate belonging by promoting additional cues that signal identity safety, such as diversity statements that value all-inclusive multiculturalism (Stevens et al. 2008, Good et al. 2020), nondiscrimination policies (Godbole 2021, Howansky et al. 2022), emphasizing that success depends on professional growth rather than inherent ability (Emerson and Murphy 2015), and demonstrating how STEM careers can benefit communities in meaningful and tangible ways (Dasgupta et al. 2022). For instance, undergraduates reported greater motivation to pursue biomedical research careers after learning about specific and concrete ways that biomedical research directly helps others (Brown et al. 2015).

Address harassment

Professional societies are uniquely positioned to define the discipline's core values and cultural expectations by educating, training, codifying, and reinforcing professional standards for ethical behavior. The National Academy of Sciences, Engineering, and Medicine's report on sexual harassment (2018) recommended adopting codes of conduct that redefine harassment and discrimination as scientific misconduct and set clear norms and expectations for respectful behavior (Williams et al. 2017, Marin-Spiotta et al. 2022). Codes of conduct and antiharassment policies should be implemented not only in academic institutions, but also at training sites, field stations, scientific meetings, and conferences (see the Ada Initiative for example policies). Sanctions for misconduct should be implemented swiftly, justly, and consistently; examples include the expulsion of members, redacting awards and honors, and removing research support. Societies should provide support and guidance for members targeted by harassment, including formal and informal reporting mechanisms and zero-tolerance policies against retaliation. Societies should also support bystander intervention training, which promotes a culture of support rather than silence. The ADVANCEGeo Partnership is one example of how professional societies can collaboratively create culture change to address harassment and improve workplace climates. Since 2018, they have implemented over 50 bystander intervention trainings to over 1500 scientists with positive improvements to participants’ knowledge and intentions to implement strategies (Berhe et al. 2020).

Change the reward system

Decenter traditional metrics and methods. Relying on narrow views of scholarly impact, such as citations and journal impact factors, can underestimate the contributions of scientists who belong to minoritized groups and impede their career advancement. Universities and professional societies can broaden the view of scholarly impact by recognizing and valuing activities that powerfully contribute to the mission of higher education beyond the ivory tower, such as community engagement, mentoring, pedagogy, sponsorship, career preparation, educational training, and diversity and inclusion work. For example, Seattle University recently revised their promotion guidelines to recognize and reward an inclusive, comprehensive range of faculty activities, resulting in a fourfold increase in successful promotions per year. Similarly, Worcester Polytechnical Institute saw a 50% increase in the proportion of promotions awarded to women faculty after implementing more inclusive guidelines. Professional societies can affirm and legitimize broad views of scholarly impact by establishing grants, awards, and journals that reimagine what counts as academic excellence.

Increase access to resources

Level the playing field by sponsoring workshops and institutes focused on issues that disproportionately affect people who identify as women and/or members of minoritized groups. Ensure equitable access to conferences, workshops, and other networking and training opportunities by providing funding for travel and dependent care. Offer virtual options for those who can't travel. Develop policies and professional guidelines that ensure that internships are paid rather than voluntary positions (for examples, see the National Council of Architectural Registration Boards and the American Institute of Architecture Students). Sponsor events that allow diverse groups of junior scientists to network with influential scholars. Develop mentoring programs that support diverse scholars across career stages. As one example, the Association for Information Systems recently launched a cohort-based mentoring program for midcareer women faculty designed to springboard career advancement and foster connections among often isolated faculty.

Develop equity-minded leaders and mentors

Institutions and professional societies should engage STEM leaders in equity-minded, inclusive practices that increase the retention and success of scientists who belong to minoritized groups. One example is the American Geophysical Union's Leadership Academy and Network for Diversity and Inclusion in the Geosciences, which provides cohort-based professional development, networks of support, and a community of practice for aspiring DEI (diversity, equity, and inclusion) champions and leaders in the geosciences. Professional societies can also develop resources and sponsor workshops that train mentors how to effectively mentor diverse groups of students and junior colleagues. Mentor training programs can draw on evidence-based practices, such as the Entering Mentoring curricula (Pfund et al. 2015), which has demonstrated improved cultural competencies and mentoring skills among trainees (Pfund et al. 2006, Byars-Winston et al. 2023).

Innovative, productive, and impactful science requires diversity, equity, and inclusion. Implementing strategies to interrupt bias and promote equity and inclusion will help science realize the full potential of a diverse talent pool.

To learn more about bias and how to interrupt it, see Cundiff (2018), Williams (2021), and Stewart and Valian (2018).

Acknowledgments

The AIBS IDEA Conference was supported by the National Science Foundation's LEAPS program (award no. 2134480) and the Doris Duke Charitable Foundation.

Author Biography

Jessica L. Cundiff was a participant in the American Institute of Biological Sciences (AIBS) IDEA conference and is an associate professor of psychological science at Missouri University of Science and Technology, in Rolla, Missouri, in the United States.