Portal de Periódicos da CAPES

Advancing Science: How Bias Holds Us Back

2018; Cell Press; Volume: 99; Issue: 4 Linguagem: Inglês

10.1016/j.neuron.2018.07.045

1097-4199

Maria Asplund, Cristin G. Welle,

Career Development and Diversity

As scientists and engineers, we must recognize the overwhelming evidence that we each harbor bias that influences our professional decisions. Yet, solving today's increasingly complex public health challenges requires diverse perspectives from multidisciplinary teams. We all have the opportunity to actively promote a more representative scientific community; let's harness the power of collective action to build diverse teams that deliver the most innovative science. As scientists and engineers, we must recognize the overwhelming evidence that we each harbor bias that influences our professional decisions. Yet, solving today's increasingly complex public health challenges requires diverse perspectives from multidisciplinary teams. We all have the opportunity to actively promote a more representative scientific community; let's harness the power of collective action to build diverse teams that deliver the most innovative science. If you ask a 5 year old to draw a scientist, roughly half of the drawings will feature a woman (Miller et al., 2018Miller, D.I., Nolla, K.M., Eagly, A.H., and Uttal, D.H. (2018). The development of children's gender-science stereotypes: a meta-analysis of 5 decades of U.S. draw-a-scientist studies. Child Dev. Published online March 20, 2018. https://doi.org/10.1111/cdev.13039.Google Scholar). Ask those same kids at age 10 and the percentage of female scientists drawn drops to 25% and continues to decline, for both males and females, throughout the teenage years (Figure 1A). And it sticks with us as we get older: extensive research shows that we carry our childhood perception of what a scientist looks like into adulthood. The Harvard implicit association test (IAT; https://implicit.harvard.edu/implicit/), which uses reaction times to assess subtle associations between groups of words, reveals that both adult women and men have a strong association of the word "male" with STEM (science, technology, engineering, and math) topics, while "female" is associated with words related to humanities and family (Nosek et al., 2009Nosek B.A. Smyth F.L. Sriram N. Lindner N.M. Devos T. Ayala A. Bar-Anan Y. Bergh R. Cai H. Gonsalkorale K. et al.National differences in gender-science stereotypes predict national sex differences in science and math achievement.Proc. Natl. Acad. Sci. USA. 2009; 106: 10593-10597Crossref PubMed Scopus (560) Google Scholar). This example demonstrates the pervasive influence of implicit bias, defined as subconscious associations that occur without awareness, intention, or control. We clearly don't "grow out" of the associations we learn in childhood about who is a scientist. Our culturally ingrained perceptions of STEM individuals indeed play a role in our professional decisions, as we act to reinforce stereotypical roles and maintain homogeneity in working environments. While those of us in STEM would like to presume that we judge each other solely on scientific merit, data suggest that we reflect our unconscious biases around gender and race in our assessments of ourselves and others. When STEM faculty are asked to evaluate a job application for a position in the lab, their evaluation score depends on whether the name at the top of the (otherwise identical) application is "John" or "Jennifer" (Moss-Racusin et al., 2012Moss-Racusin C.A. Dovidio J.F. Brescoll V.L. Graham M.J. Handelsman J. Science faculty's subtle gender biases favor male students.Proc. Natl. Acad. Sci. USA. 2012; 109: 16474-16479Crossref PubMed Scopus (1747) Google Scholar). John is more likely to get offered a position and is offered a starting salary of 14% more, even when performance scores indicate that Jennifer is more likely to fulfil the job requirements (Reuben et al., 2014Reuben E. Sapienza P. Zingales L. How stereotypes impair women's careers in science.Proc. Natl. Acad. Sci. USA. 2014; 111: 4403-4408Crossref PubMed Scopus (357) Google Scholar). Similar findings were reflected in an analysis of over a thousand evaluation letters written for postdoctoral candidates, which finds lower scores for all female applications, and an emphasis on "grindstone" adjectives, like hardworking and diligent, while male candidates received "standout" adjectives, such as superb and brilliant (Dutt et al., 2016Dutt K. Pfaff D.L. Bernstein A.F. Dillard J.S. Block C.J. Gender differences in recommendation letters for postdoctoral fellowships in geoscience.Nat. Geosci. 2016; 9: 805-808Crossref Scopus (104) Google Scholar). Likewise for medical students applying to residency, black candidates are more likely to be described as "competent," while white candidates are bestowed with many standout adjectives (Ross et al., 2017Ross D.A. Boatright D. Nunez-Smith M. Jordan A. Chekroud A. Moore E.Z. Differences in words used to describe racial and gender groups in Medical Student Performance Evaluations.PLoS ONE. 2017; 12: e0181659Crossref PubMed Scopus (121) Google Scholar). For all of these studies, there was no substantial difference in evaluation if the evaluator was male or female, reinforcing the idea that implicit bias is a cultural force, shared by men and women in STEM. These studies, and a growing number of other similar research efforts, indicate that the underrepresentation of women and minorities in STEM topics can be at least partially attributed to implicit bias. Merit alone does not predict our judgement—instead we are influenced by our unconscious set of expectations and stereotypes. This means candidates who are not in line with the stereotypical idea of a "high-performer" find themselves working against an invisible barrier. Across all sorts of evaluations, definitions of merit are corrupted by underlying cultural bias. Implicit bias contaminates our peer-review procedures, from grant reviews (Pohlhaus et al., 2011Pohlhaus J.R. Jiang H. Wagner R.M. Schaffer W.T. Pinn V.W. Sex differences in application, success, and funding rates for NIH extramural programs.Acad. Med. 2011; 86: 759-767Crossref PubMed Scopus (157) Google Scholar), acceptance of papers in high-ranking journals (Holman et al., 2018Holman L. Stuart-Fox D. Hauser C.E. The gender gap in science: How long until women are equally represented?.PLoS Biol. 2018; 16: e2004956Crossref PubMed Scopus (317) Google Scholar), to our own perception of our academic performance (Cooper et al., 2018Cooper K.M. Krieg A. Brownell S.E. Who perceives they are smarter? Exploring the influence of student characteristics on student academic self-concept in physiology.Adv. Physiol. Educ. 2018; 42: 200-208Crossref PubMed Scopus (56) Google Scholar). While we expect that positions, awards, and grants in academia are based on objective measures, we must realize that these measures are systematically influenced by implicit bias. Are we working in a meritocracy or are we just maintaining this illusion? We consistently apply our biases throughout daily life—in our hiring decisions, scientific reviews, and even in casual interactions. Any person allowing their implicit bias to get the better of them is, in principle, a participant in discrimination. Nevertheless, it might seem harsh to use such a strong word for a single action which at the time might seem unintentional and not meant to cause harm. For example, we have often experienced our colleagues complaining to us about other female professionals, describing these women as "too aggressive" and "hard to work with." However, we, on a closer look, find our male and female colleagues equivalent across these attributes. Examples like these may seem minor in insolation, but collectively send a message about how women in our field are expected to act in order to avoid being excluded from collaborations with peers, a critical element for success in a scientific career. That small individual biases can result in large collective biases was demonstrated in the Nobel Prize winning theory of Thomas Schelling, and elegantly visualized online by Vi Hart and Nicky Case in the Parable of the Polygons (https://ncase.me/polygons). In this simulation, triangles and squares in a neighborhood each hold a slight preference to be near neighbors of the same shape. Rearranging the shapes within the neighborhood until each shape is happy with its neighbors leads to a highly segregated community (this may make more sense if you try the simulation yourself). Likewise, within our STEM community, our slight individual biases lead to collective lack of diversity. So what is the consequence of collective bias for STEM? When we limit diversity in our lab and networks, we end up limiting our science. Analysis of more than 2.5 million publications shows that increased ethnic diversity in the author team correlates with increased citations and impact factor (IF) of the journal (Freeman and Huang, 2015Freeman R.B. Huang W. Collaborating with people like me: ethnic co-authorship within the US.J. Labor Econ. 2015; 33: S289-S318Crossref Scopus (145) Google Scholar). Research in business has shown a clear case for diversity in terms of increased productivity, innovation, and financial benefit (https://www.mckinsey.com/business-functions/organization/our-insights/delivering-through-diversity). Interestingly, increases in conflict within a group, as a consequence of diversity, lead to greater learning for all group members and better solutions. These few findings, among many, point out that when we collectively fail to offer equal opportunities, we miss out on the best talent and the most productive teams. Therefore, we argue change is crucial for science and suggest that this should be driven on two levels: the institutional and the individual. Coordinated efforts to combat implicit bias and improve diversity must come from the top and be designed to impact critical junctures that predict success in the scientific pathway. Universities, funding agencies, and scientific journals can all drive important progress in encouraging fair and diverse representation in science. These key players have power and resources to form a framework where diversity is not only encouraged but requested. Universities must continue to thoughtfully evaluate their admissions processes, faculty hiring, promotion, and tenure policies. Distribution of resources across faculty should be reviewed ensuring they do not present barriers to a diverse faculty (National Academy of Science, https://www.nap.edu/resource/11741/bias_and_barriers_summary.pdf). Data regarding the diversity of students and faculty across all levels of the academic hierarchy, the metrics of success, and the allocation of resources should be collected, publicized and used as the basis of institutional policies. Funding agencies should ensure that their review processes are equitable, for instance, through broader recruitment for peer review panels. Journals, who serve as gatekeepers of the prize of publication, should ensure diversity of reviewers and editorial boards and equity among authors of invited papers. Those who believe this is already the case might want to think about why women, at present, have systematically lower rates of R01 (an important U.S. NIH grant) renewal, despite the use of more standout adjectives in the written comments (Kaatz et al., 2016Kaatz A. Lee Y.-G. Potvien A. Magua W. Filut A. Bhattacharya A. Leatherberry R. Zhu X. Carnes M. Analysis of national institutes of health r01 application critiques, impact, and criteria scores: does the sex of the principal investigator make a difference?.Acad. Med. 2016; 91: 1080-1088Crossref PubMed Scopus (99) Google Scholar)? And why does internal review of practices at major publishers reveal that gender imbalance in journals is still pervasive (https://www.nature.com/news/gender-imbalance-in-science-journals-is-still-pervasive-1.21348)? When considering how to increase diversity, the Parable of the polygons simulation shows us another important principle—it is not sufficient to be unbiased. Even when the individual triangles and squares living in a segregated neighborhood lose their preference for similar neighbors, the segregation of the community doesn't change. Instead, each shape needs to actively seek out increased diversity in order for their neighborhoods to become less segregated. Ultimately, diversity is personal. In a survey of 22 companies that excelled in maintaining a diverse workforce, each had leaders who championed increasing diversity as an integral part of their leadership strategy (https://www.mckinsey.com/business-functions/organization/our-insights/lessons-from-the-leading-edge-of-gender-diversity). The CEO of the notably diverse tech-company Slack, Stewart Butterfield, sends an inspirational message when he explains why Slack does not have a Head of Diversity. He wants these efforts to be something "everyone is engaged in" and not "shunted off to a designated specialist," a statement that raises the bar for all of us (https://www.theatlantic.com/technology/archive/2018/04/how-slack-got-ahead-in-diversity/558806/). Promoting diversity can, and should, be implemented by every member of the organization. Therefore, we must ask ourselves as individual scientists how we can actively promote diversity in our own spaces. As we manage other scientists, participate in admissions committees, serve as grant and paper reviewers, plan conferences, and write letters of recommendation, we need to pay attention to our own bias. Who we promote and more importantly—why? Each of these situations presents an opportunity to actively address bias and promote diversity. Moreover, the collective action of like-minded individuals is a powerful force for change—so it is crucial that we as individuals work together to improve diverse representations in science. The predictive models of authorship make it clear that, at their current rate, STEM fields are many decades (or even centuries) away from gender parity (Figure 1B; Holman et al., 2018Holman L. Stuart-Fox D. Hauser C.E. The gender gap in science: How long until women are equally represented?.PLoS Biol. 2018; 16: e2004956Crossref PubMed Scopus (317) Google Scholar). Even so societal changes, such as more diverse representation in prominent STEM positions, might fade out our gender-bias over time, faster solutions require action. A number of successful approaches to increase diversity have been identified in recent literature. Increasing diversity on leadership committees is shown to produce more diverse selections (Nittrouer et al., 2018Nittrouer C.L. Hebl M.R. Ashburn-Nardo L. Trump-Steele R.C.E. Lane D.M. Valian V. Gender disparities in colloquium speakers at top universities.Proc. Natl. Acad. Sci. USA. 2018; 115: 104-108Crossref PubMed Scopus (114) Google Scholar). Providing lists of qualified women speakers (for example, Anne's List, https://anneslist.net) to conference organizers can promote gender parity (Figure 1C; Klein et al., 2017Klein R.S. Voskuhl R. Segal B.M. Dittel B.N. Lane T.E. Bethea J.R. Carson M.J. Colton C. Rosi S. Anderson A. et al.Speaking out about gender imbalance in invited speakers improves diversity.Nat. Immunol. 2017; 18: 475-478Crossref PubMed Scopus (57) Google Scholar), as can matching speaker diversity to the rate of representation in the field (for example, as calculated on https://biaswatchneuro.com). Acknowledging that merit-based evaluations still contain bias (Dutt et al., 2016Dutt K. Pfaff D.L. Bernstein A.F. Dillard J.S. Block C.J. Gender differences in recommendation letters for postdoctoral fellowships in geoscience.Nat. Geosci. 2016; 9: 805-808Crossref Scopus (104) Google Scholar, Ross et al., 2017Ross D.A. Boatright D. Nunez-Smith M. Jordan A. Chekroud A. Moore E.Z. Differences in words used to describe racial and gender groups in Medical Student Performance Evaluations.PLoS ONE. 2017; 12: e0181659Crossref PubMed Scopus (121) Google Scholar), and thus advocating during admissions and hiring processes for candidates with slightly less traditional backgrounds, or slightly less glowing recommendation letters, in the interest of acknowledging that these applicants have experienced bias throughout their career that other applicants have not (http://www.bain.com/publications/articles/what-stops-women-from-reaching-the-top.aspx). Avoiding casual bias, such as inequitable use of professional titles (Files et al., 2017Files J.A. Mayer A.P. Ko M.G. Friedrich P. Jenkins M. Bryan M.J. Vegunta S. Wittich C.M. Lyle M.A. Melikian R. et al.Speaker Introductions at Internal Medicine Grand Rounds: Forms of Address Reveal Gender Bias.J. Womens Health (Larchmt.). 2017; 26: 413-419Crossref PubMed Scopus (151) Google Scholar), can reinforce that scientists are valued equally. Last, but not least, everyone is responsible for ensuring that efforts to diversify STEM fields are not accompanied by social penalties (Hekman et al., 2016Hekman D.R. Johnson S.K. Foo M.-D. Yang W. Does diversity-valuing behavior result in diminished performance ratings for non-white and female leaders?.AMJ. 2016; 60: 771-797Crossref Scopus (117) Google Scholar). Don't underestimate the importance of mentorship and advocacy. Simple interactions, such as when a mentor nominates a woman or underrepresented minority to give a talk in their stead, or to meet with influential guest speakers, can accumulate into a major impact. Suggesting the names of women and minority colleagues for a grant or paper review, or as a member of a conference planning committee can increase visibility and help foster critical networks. As a mentor to women and underrepresented minority trainees, ensure that those trainees get equal distribution of your time and resources. In STEM PhD programs, women and minority students are less likely to publish a paper than their white, male peers (Figure 1D; Mendoza-Denton et al., 2017Mendoza-Denton R. Patt C. Fisher A. Eppig A. Young I. Smith A. Richards M.A. Differences in STEM doctoral publication by ethnicity, gender and academic field at a large public research university.PLoS ONE. 2017; 12: e0174296Crossref PubMed Scopus (37) Google Scholar). Likewise, male faculty, and particularly those with HHMI or National Academy of Science membership, train fewer female graduate students and postdocs than their colleagues (Sheltzer and Smith, 2014Sheltzer J.M. Smith J.C. Elite male faculty in the life sciences employ fewer women.Proc. Natl. Acad. Sci. USA. 2014; 111: 10107-10112Crossref PubMed Scopus (223) Google Scholar), serving as de facto gatekeepers to academic faculty jobs for trainees. Ensure that such a disparity does not exist in your lab or graduate program. Finally, aggressively promote and support the transition to independent research careers for trainees, arguably the most critical step toward diversifying STEM research. There is an interesting conundrum that comes when discussing implicit bias—it can actually increase bias! By acknowledging that bias exists, people can paradoxically feel like they have license to act on their bias—"everyone is doing it!" (Duguid and Thomas-Hunt, 2015Duguid M.M. Thomas-Hunt M.C. Condoning stereotyping? How awareness of stereotyping prevalence impacts expression of stereotypes.J. Appl. Psychol. 2015; 100: 343-359Crossref PubMed Scopus (122) Google Scholar). Behavioral experiments show that clearly identifying the behavior as undesirable, and emphasizing better alternatives, can reduce unconscious bias. It deserves to be mentioned that there indeed are some "diversity promotion" strategies that are counterproductive. We've encountered many strategies that provide peripheral accommodation for women or minorities, without addressing overarching evidence of systemic bias. For example, we've attended conferences that require a "diversity speaker," but don't strive for equivalent representation of female and underrepresented scientist presenters. This makes it seem as though diversity is a box to be checked, rather than a goal for the community. Although it might seem obvious, it deserves to be repeated: measures to seek equality can never be a platform for legitimizing prejudice! We have personally witnessed how generalizing statements such as women's lack of "necessary confidence or interest in power" or the female "high-pitched voice" are to blame for why women fail to advance. These statements must be challenged and exposed as the prejudice they are. Let's change the system, rather than teach those at a disadvantage to compensate for a system that fails them. In this era of "big science," we need interdisciplinary and multi-scale teams to create innovative solutions to our biggest public health challenges. Major scientific initiatives promoted by NIH, DARPA, and the ERC call on complex coordination across scientists, engineers, clinicians, and industry to achieve bold goals. This presents a unique and timely opportunity for improving diversity. For example, thanks in part to the BRAIN Initiative, the field of neural engineering is growing rapidly to build tools for the forefront of neuroscience research and clinical implementation. The term "neural engineering," which in 2005 generated no more than 12 PubMed results, produces greater than 10-fold more results in 2017. As a direct consequence, this field, which previously was dominated by engineering topics, and predominantly white men, has become increasingly interdisciplinary and now presents the opportunity to grow gender and ethnic diversity in our research teams, conference attendees, peer review panels, and editorial boards (Box 1). It is imperative that neural engineering, and other STEM fields, use this changing landscape to actively seek out diversity.Box 1Personal ExperienceRecently, we experienced the changing scene of Neural Engineering first hand, while participating at the Gordon Research Conference on Neuroelectronic Interfaces, where we led a session on implicit bias. This serves as good example on how institutional efforts (a funding scheme requesting that diversity is addressed) and individual efforts (all that actively participated) can be brought together to drive change. This brainstorming session was well attended by both women and men. Most importantly, it resulted not only in vivid discussions but also in solid action. After the session, participants drafted and signed a letter to major journals in the field requesting diversified representation on their editorial boards, highlighting the importance of the collective action of individuals. Recently, we experienced the changing scene of Neural Engineering first hand, while participating at the Gordon Research Conference on Neuroelectronic Interfaces, where we led a session on implicit bias. This serves as good example on how institutional efforts (a funding scheme requesting that diversity is addressed) and individual efforts (all that actively participated) can be brought together to drive change. This brainstorming session was well attended by both women and men. Most importantly, it resulted not only in vivid discussions but also in solid action. After the session, participants drafted and signed a letter to major journals in the field requesting diversified representation on their editorial boards, highlighting the importance of the collective action of individuals. To summarize, each of us needs to play an active role in order to change our culture in STEM that currently allows bias to flourish. We highlight five key actions that every scientist can take to work towards overcoming implicit bias and creating a more diverse, more effective scientific community (Box 2).Box 2Actions for Every ScientistSeek DiversityRecognize that we, as individuals, have implicit bias. We also have a responsibility to actively advocate for diversity in our hiring practices, grant reviews, conference planning, and mentoring. Let's make sure that all of the best and the brightest of all genders and races can rise to the top. Why? Because the future of science depends on the highest level of innovation.Make a Conscious Effort to Be Inclusive of Those Who Are UnderrepresentedBeing in the right spot at the right time can have a significant impact on a career, but it often depends on being invited. When minority applicants are excluded, question the motives of yourself and others—are there really no qualified women or underrepresented minority candidates around? Within similar fields, this has been empirically proven not to be true (Klein et al., 2017Klein R.S. Voskuhl R. Segal B.M. Dittel B.N. Lane T.E. Bethea J.R. Carson M.J. Colton C. Rosi S. Anderson A. et al.Speaking out about gender imbalance in invited speakers improves diversity.Nat. Immunol. 2017; 18: 475-478Crossref PubMed Scopus (57) Google Scholar).Share in the ResponsibilityEveryone plays a role in promoting diversity. It's not the job of a single woman on a search committee to ensure that adequate numbers of women are interviewed; it is the job of every committee member to do so. We need to stand together to bring change.Resist and Reduce the Effects of Implicit BiasIn recruitment and promotion, don't forget that implicit bias impacts not only your decisions, but also the many decisions others made before you. Keep in mind that common evaluation metrics are likely to be contaminated by the bias of a community and are not an accurate predictor of future performance.See Diversity as an Investment in Good ScienceMeeting the challenges of today's public health needs requires rigorous, innovative, and team-science-driven solutions. Diverse teams produce better solutions, in part because of the inherent conflict that arises from wide-ranging perspectives. By increasing diversity in our labs, departments, committees, and institutions, we strengthen our science. Recognize that we, as individuals, have implicit bias. We also have a responsibility to actively advocate for diversity in our hiring practices, grant reviews, conference planning, and mentoring. Let's make sure that all of the best and the brightest of all genders and races can rise to the top. Why? Because the future of science depends on the highest level of innovation. Being in the right spot at the right time can have a significant impact on a career, but it often depends on being invited. When minority applicants are excluded, question the motives of yourself and others—are there really no qualified women or underrepresented minority candidates around? Within similar fields, this has been empirically proven not to be true (Klein et al., 2017Klein R.S. Voskuhl R. Segal B.M. Dittel B.N. Lane T.E. Bethea J.R. Carson M.J. Colton C. Rosi S. Anderson A. et al.Speaking out about gender imbalance in invited speakers improves diversity.Nat. Immunol. 2017; 18: 475-478Crossref PubMed Scopus (57) Google Scholar). Everyone plays a role in promoting diversity. It's not the job of a single woman on a search committee to ensure that adequate numbers of women are interviewed; it is the job of every committee member to do so. We need to stand together to bring change. In recruitment and promotion, don't forget that implicit bias impacts not only your decisions, but also the many decisions others made before you. Keep in mind that common evaluation metrics are likely to be contaminated by the bias of a community and are not an accurate predictor of future performance. Meeting the challenges of today's public health needs requires rigorous, innovative, and team-science-driven solutions. Diverse teams produce better solutions, in part because of the inherent conflict that arises from wide-ranging perspectives. By increasing diversity in our labs, departments, committees, and institutions, we strengthen our science. The authors gratefully acknowledge Prof. Anelis Kaiser (University of Freiburg) and Prof. Paul Walton (University of York) for their expert advice on the science of implicit bias. Furthermore, we wish to thank our STEM colleagues and friends, in particular Dr. Joseph Pancrazio, Dr. Takashi Kozai, Dr. Aviva Abosch, Elissa Welle, Dr. Nichole Litvinas, and Jonathan Welle for giving valuable feedback on the manuscript and for helpful discussions on the topic.