Introduction

The Intersectional Problem of Race and Gender Discrimination in Science Research Careers

Women of color remain underrepresented in science fields in both the UK and the USA at the research level compared to their prevalence in the overall population. Data suggests that such individuals are at risk of being doubly disadvantaged in science research careers: in 2013 Black women constituted 0.1% of professorial jobs in the UK (Advance HE, 2019), while representing 1.5% of the UK population (Office for National Statistics, 2013). While still under-represented, White women represented 22.9% of professorial jobs while representing 51% of the UK population. Over-represented were White men, who represented 67.5% of professorial jobs while representing 43% of the population. In the USA, minority women continue to be under-represented in receiving science and engineering degrees relative to their population (National Science Foundation, 2019). Black women are less likely than White women to progress from undergraduate to masters’ degree (52% vs 60%) and less likely to progress to a doctorate (9% vs 13%) (National Science Foundation, 2019).

The reasons for the continuing underrepresentation of women of color in science careers continues to be under-researched. Existing literature on discrimination in science careers continues to focus on gender rather than race (Bruning et al., 2015). A color blind and androcentric viewpoint ignores the combination of racial and gender challenges that act as a “double bind” (L. E. Malcom & Malcom, 2011; S. M. Malcom et al., 1976) for women of color scientists. The limited literature on women of color in science careers identifies challenges which combine both racial and gender prejudices, such as being held to higher standards than men or White women (J. C. Williams et al., 2016a, 2016b), exclusion from White or male-dominated mentoring networks (Blackwell et al., 2009; E. O. McGee & Bentley, 2017), tokenism (Villanueva et al., 2019), hypervisibility when challenging inequalities (Alexander & Hermann, 2016), discrimination in recruitment based on both race and gender (A. C. Johnson, 2011a, 2011b), the lack of gender and race awareness from majority White, male colleagues (W. H. Robinson et al., 2016), and diminished confidence and stagnation due to repeated rejection (Opara, 2017). Science research organizations have invested in diversity strategies aimed at improving the retention and promotion of minorized scientists (Gomez et al., 2021).

For the purposes of this study, the term “strategies” includes both organizational interventions, such as recruitment strategies, as well as strategies developed and enacted by individual women of color scientists to counter gendered racism.

Intersectionality Theory

An intersectional framing offers a corrective to the conflation of gender and race, recognizing the simultaneous racialized and gendered experiences of women of color in research contexts. Whereas many diversity initiatives focus on a single dimension of identity, such as the UK’s Race Equality Charter (Equality Challenge Unit, 2016), an intersectional response acknowledges that one cannot divorce one’s race from one’s class, gender, sexuality, or other characteristics (Crenshaw, 1989). Research suggests that organizational diversity policies which address gender alone privilege White women (Torres, 2012), whereas those that address race alone have been found to privilege non-White men (Guyan & Oloyede, 2020). The term “intersectionality” originated from Black feminists, seeking to understand how power operated through both race and gender (Hooks, 1987; Smith et al., 1974), although the concern with connections between different forms of subordination and domination have long been part of sociological analyses of institutions. Organizational scholars have argued that the complexity of contemporary work environments requires more nuanced explorations of how unequal access to social and institutional resources occurs (J. K. Rodriguez et al., 2016), based on how individuals are positioned within categories of race and gender, as well as how they position themselves within institutional and organizational hierarchies (Acker, 2006).

Applying Intersectionality Theory to Science Research Careers

In seeking to understand how to apply intersectional theory to power relations in work and society, intersectional models of analysis have evolved (Hill Collins & Bilge, 2016). These models allow for the understanding of how different social spaces within a society interact to produce multiple, interlocking forms of oppression. Floya Anthias (2013) conceptualized social spaces as consisting of four arenas: organizational, intersubjective, representational, and experiential. The value of Anthias’ concept of social arenas to this scoping review is twofold; firstly, social space becomes a lens through which the literature reviewing strategies to counter gendered racism can be analyzed. Furthermore, it becomes possible to analyze whether organizational or individual strategies are prevalent or absent from each social arena, thereby leading to recommendations for future research directions. The four social arenas as they relate to the topic of this study are described below.

Firstly, the organizational arena will be discussed. This social arena which relates to how researchers are positioned in relation to the research organization through processes of recruitment, retention, promotion. The organizational arena also contains organizationally sanctioned practices such as mentoring and training. Researchers’ positioning in relation to wider society is described by this category, including their family life, class position, and nationality. It would be expected that strategies led and conducted by organizations dominate in the literature relating to this space. Individual strategies would be largely absent, as an individual researcher is rarely able to direct institutional policies and practice.

Secondly, the representational arena is that where the images, texts, and documents of an organization reside. This arena includes photography and works of art which depict ideals of scientists, marketing and job recruitment material, research papers, data collection and analysis on race and gender, and published works. It would be expected that strategies led and conducted by organizations would prevail in the literature relating to the representational space, as data collection, marketing, and publishing practices are institutionally sanctioned. However, individual strategies may also occur, as women of color narrate their lived experience through journal articles.

Intersubjective (practices): how scientists relate to each other in a lab or in the field, and how scientists develop coping strategies. It would be expected that organizational policies would prevail in the literature relating to this social space, as organizations are expected to provide and enforce norms of behavior in its social spaces. A prevalence of individual practices in this space would unfairly put the onus on individual women of color to develop coping and resistance mechanisms to deal with gendered racism.

Experiential (narratives): individual identity narratives, such as the formation of a science identity, a racial, and/or gender identity. The development of psychological characteristics, such as self-efficacy and career values, are included in this category. Organizational strategies would be expected to support women of color in developing their confidence and identity as scientists. A prevalence of individual strategies in this space would point to the absence of institutional support for women of color as knowledge producers and scientific experts.

This study goes on to identify the presence (or absence) of literature within each social space. Anthias’ lens adds value in allowing us to evaluate which types of strategy may be over-represented within a type of space, and which are under-represented. For example, a prevalence of individual strategies in a social space would indicate that organizational strategies are lacking or ineffective with the result is that women of color are forced to develop strategies themselves to survive in a hostile environment. Where strategies appear to be under-represented in a particular arena, the study goes on to make research recommendations later in the paper.

Research Questions and Paper Structure

Two research questions can therefore be formulated.

Research question one: Which strategies across literature findings have been employed to address gendered racism in the arena of science research careers?

Research question two: What value does Anthias’ model add in understanding the differences between organizational and individual strategies?

The paper is structured as follows: firstly, a positionality statement and description of methods will be provided. The high-level findings will then be covered, followed by a detailed description of the findings, organized by Anthias’ model of social spaces, and exploring the organizational and individual strategies found in the literature. The research questions will be revisited in the “Discussion” section of this paper and suggestions for further research are provided. A brief conclusion summarizes the findings of the study.

The paper provides only a limited assessment of the extent to which a strategy can be recommended and does not recommend certain strategies above others. This is due to the narrow range of countries included, the wide variation of organizations and scientific disciplines covered, and the difficulty in defining a successful strategy. For example, coping strategies help women of color remain in the workplace, which would be a successful outcome for an organization, but may be detrimental to the individual woman’s mental health. This paper therefore remains a scoping review with indications for where further research could be conducted but does not provide policy and practice recommendations.

Positionality Statement

The author is a South Asian cis-gender, heterosexual, middle-class woman, whose childhood, education, and career experiences have been in predominantly White, male environments. When developing this research project, the author was a temporary staff member of an all-White researcher group. The resulting experience of isolation and misunderstanding while writing this paper provided a personal impetus to understand the policy implications of the literature. The combined personal and professional experiences of the author have driven the desire to understand the full range of personal and institutional strategies to address gendered racism.

Methods

Given the extensive scope of the STEM subjects (science, technology, engineering and mathematics), a scoping review was selected as the most appropriate form of literature, following the definition of Grant and Booth (2009). A scoping review provides an initial assessment of the extent of available research literature and in this case, a preliminary assessment of the literature on strategies to address race and gender discrimination. The scoping review method therefore offers some elements of replicability and transparency, while allowing for the author to provide a critical analysis of the strategies. The scope of careers included undergraduate, graduate, postdoctoral, early entry, midcareer, and leadership career positions. Given the lack of empirical studies which address both race and gender discrimination, the decision was taken to review not only empirical papers but also theoretical papers and literature reviews. As a result, tests for reliability were only performed on empirical quantitative papers. The decision was taken to include literature from multiple countries (albeit published in English) to review the widest possible range of inclusion strategies.

This scoping review complies with the recommendations for preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement (Page et al., 2021).

Developing the Inclusion Criteria

Given the decision to include qualitative and theoretical papers, the inclusion criteria were formulated through aspects of the SPIDER question format tool which was developed for searching qualitative and mixed methods reviews (Cooke et al., 2012). For this review, the “sample” and “phenomenon of interest” criteria were addressed so as to fulfill the aim of characterizing the literature in this area, rather than providing an exhaustive analysis more appropriate for a structured literature review and more feasible given a single author. In line with the iterative, flexible nature of the scoping review process, the “sample” was revised during the study selection process. Initially defined as “scientists who are women of color,” this was refined to “women who are working as scientists in a science research organization” in order to include organizational and individual strategies. The “phenomenon of interest” was defined as “strategies to address race and gender discrimination.” “Strategies” in this article include formal organizational interventions as well as individual responses.

Identifying Relevant Studies

The search strategy was designed to identify formally published, peer-reviewed articles, and selected secondary literature, including conference proceedings and book chapters. The search strategies consisted of keywords and subject headings related to gender, race, strategies, and STEM. Google Scholar was not used due ongoing concerns with the lack of transparency in its search engine and ensuing lack of replicability (Haddaway et al., 2015). Table 9 gives full search strategies for each database. The database searches were supplemented by conducting backward and forward citation checking.

Study Selection

Study selection followed a two-step approach: firstly, screening of the title, abstract, and keyword screening followed by a full text screening. As the work is sole authored, no inter-rating was possible, which is a limitation of the study.

Data Extraction and Management

Data from the full texts were extracted onto an Excel table named “characteristics of included studies,” including country, authors, publication type, publication date, publication name, study design, population of interest (sample characteristics including gender and race), type of strategy discussed (e.g., mentoring, training), thematic relevance (e.g., to race, to gender, or both), and a mapping to one or more of Anthias’ social arenas.

Thematic Analysis

The results of the 86 included studies were synthesized using a basic narrative synthesis approach, applying principles of thematic analysis. An initial coding pass generated 120 codes, which were then collated into potential themes. The themes were then checked in relation to the original dataset, as well as the structuring mechanism of Anthias’ four datasets. The themes are presented in this study as subdivisions in Table 4 below.

Quality Assessment

The Mixed Methods Appraisal Tool (MMAT) (Hong et al., 2018) was used to assess the quality of included empirical studies. This tool is designed for quality assessment in systematic reviews that include quantitative, qualitative, and mixed-method studies. The author assigned articles one point for each criterion that was fulfilled and zero points for each criterion that was unfilled, or where it was not possible to assess. The points were totaled to produce a quality index which ranked studies on the proportion of total criteria met. Four studies were excluded from the synthesis due to a low-quality score using the MMAT tool. The identification, screening, and inclusion stages are described in Fig. 1 below.

Fig. 1
figure 1

Adapted from Page et al. (2021)

PRISMA flow diagram.

Full size image

The stages followed in the PRISMA flow diagram in Fig. 1 above are explained below, in Table 1. Additional stages have been added based on guidelines produced by Robinson and Lowe (2015) with additional recommendations from Harari et al. (2020).

Table 1 Scoping review stages
Full size table

Limitations of the Study

The limitation of the study is that only peer-reviewed academic journals, reports, book chapters, or conference proceedings were included in order to guarantee a level of validity and reliability. Therefore, policy-oriented research strategies for race and gender discrimination, such as those published by professional organizations or government departments, were not included in this study. The author speaks English, Russian, French, and Sinhala; however, only English language articles were selected as no articles were found in the other languages spoken by the author. Additionally, the paper contains the limitations inherent in a scoping review method (M. J. Grant & Booth, 2009), namely the limited reproducibility of results, an absence of quality assessment, and, given the author’s positionality, the potential for bias.

Findings

This section will provide an overview of the list of journals and conference proceedings, number of articles and impact factor, the incidence of articles over time, their country of origin, and demographic details of the population studied.

This review identified 86 studies describing strategies to address race and gender discrimination. The complete list of journal articles and their impact figure is provided in Table 2 below.

Table 2 List of journals
Full size table

As Table 2 indicates, educational journals predominate (11 journals), with journals focused on their scientific sub-discipline coming second (8 journals). Once the content of the articles has been included (the Bibliography in Appendix 1 at the end of this article provides the titles), much of the literature is confined to minority women scientists who are in higher education, either as students or faculty staff. There is a lack of literature relating to minority women scientists working in private industry or in the government sector. Furthermore, there is only one paper from a journal focused on race and ethnicity, indicating that strategies relating to gendered racism are studied and published mostly within the scientific discipline or within higher education more widely. Surprisingly, only one management journal is included, suggesting that HR professionals, diversity and inclusion professionals, and organizational leaders are not exposed to organizational strategies to counter gendered racism within STEM.

The full list of conference proceedings reviewed is provided in Table 3 below.

Table 3 List of conference proceedings
Full size table

As Table 3 indicates, engineering and computer sciences predominate, with a lack of publishing in the conference proceedings of other scientific disciplines. Conferences relating to gender and race do not appear to have published papers relating to STEM disciplines, suggesting that, for the timescale included in this study, there was a lack of interest in gendered racism within the STEM context. As long as the onus remains on STEM-centered conferences to address the issue of gendered racism, the potential from learning from other disciplines and industries more widely will be missed.

The incidence of articles published over time is provided in Fig. 2 below.

Fig. 2
figure 2

Incidence of race and gender-related articles over time

Full size image

As can be seen in Fig. 2, the date range of the final sample is 44 years. There are long gaps between the first three publications which may refer either to the lack of online archived material during this time or, more saliently, to the lack of mainstream interest in gendered racism. The majority of papers were empirical, with 59 papers employing quantitative methods, 24 papers employing qualitative methods, and 3 papers employing mixed methods. The remaining papers consisted of one literature review, two symposium reviews, and two theoretical papers.

The final sample contains papers from only four countries: the USA, Brazil, the UK, and Denmark. The USA is represented in 82 papers. There are only four papers which assessed race and gender data from outside the USA: these include two papers from Brazil (Ferrari et al., 2018; Ribeiro & Schlege, 2018), one from the UK (Heward et al., 1995), and one from Denmark (Andersen, 2001). Given the relative cultural similarity between these countries, which are all part of the Global North, the decision was taken to include these four papers.

African-Americans were the most studied racial group (60 papers), then Latinx/Hispanic (47), Asian-Americans (38), White (29), American Indian/Native American (24), mixed race (12), Native Hawaiian/Pacific Islander (6), Black (not American) (5), Arab American (1), and Asian British (1). Thirty-two papers referred to under-represented minorities (URM), people of color (POC), or non-White people, but did not further detail their racial heritage. Several papers excluded other racial minorities from analysis due to the small sample size.

Both men and women were studied in 59 papers, 22 papers studied only women, and 2 papers studied only men. A further three papers did not specify the gender under analysis. Only three papers recruited transgender populations in addition to traditional binary gender choices, and all three subsequently excluded the transgender population due to the small sample size. The most vulnerable and under-represented racial and gender groups remain unstudied in quantitative research, a methodological choice which has been claimed to exacerbate problems of under-representation in a field (Ginther, 2018).

Detailed Findings

This section is structured according to Anthias’ four social arenas in Table 3 below, with further granularity shown in Table 4. Some articles (marked with an asterisk) appear in more than one category, as they discussed strategies which crossed more than one social arena.

Table 4 List of articles identified in the literature search categorized according to Anthias’ intersectional model of social arenas
Full size table

Within each of the four categories above, the findings are further sub-divided, as outlined in Table 5 below.

Table 5 Subdivision of the four main arenas
Full size table

The paper will now go on to discuss the detailed findings by each of the four arenas described in Table 5 above.

Arena 1: The Organizational

Before discussing the organizational strategies in detail, Table 6 shows the distribution of papers across topic areas.

Table 6 List of papers organized by types of organizational strategy
Full size table

As would be expected, the majority of interventions discussed in the literature took place in the organizational arena. Much recent scholarship focuses on the responsibilities of organizations and their policies on recruitment, mentoring, retention, and training practices. This is unsurprising, given that such official interventions are easier for institutions to plan, record, and assess than solutions developed and enacted by individuals.

Recruitment Policies

Recruitment was the most popular type of strategy mentioned in the organizational arena: changes to recruitment policy were recommended in 33% of articles in this arena. Strategies were presented across two broad approaches: those that specifically addressed gender and race in recruitment and those that did not, but which recommended gender- and race-blind interventions.

Papers which proposed tailoring science recruitment campaigns to appeal to women of color, consisted of favoring the social value of science. Ferrari et al. (2018) suggested targeting recruitment to the geographical areas of Brazil which were more likely to contain indigenous and poorer minorities in a paper that considered class along with race and gender. Elsewhere, suggestions were made for changing traditional degree-entry requirements which focus on academic grades at the exclusion of students’ socio-economic backgrounds (Antón et al., 2018; R. McGee et al., 2012; Wood et al., 2016) and providing more accessible explanations of academic practices, such as fellowships, during recruitment campaigns (A. A. Rodriguez & Anderson-Rowland, 2012). The creation of more part-time degree and job options was recommended to attract minority women (Cipher et al., 2017). Increasing financial incentives for minority women to apply was a popular recommendation (Jeffe & Andriole, 2011; J. Johnson & Bozeman, 2012; S. M. Malcom et al., 1976; Newton et al., 2011; Turner et al., 2018).

Recognizing that incentives for minority women to apply were insufficient, recruitment quotas were recommended as far back as 1976 (S. M. Malcom et al., 1976). Positive discrimination in the sense of favoring the recruitment of minority women scientists continues to be recommended (Casad et al., 2018; Mathur et al., 2018a, 2018b; Mathur et al., 2018a, 2018b; R. McGee et al., 2012; symposium, 2011). Other authors recommended increasing the number of women and minorities when recruiting staff or students, but stopped short of recommending quotas (Jackson et al., 2019; O’Herrin et al., 2018; Oh & Lewis, 2005). As will be seen throughout the findings, some strategies put the onus on women of color and their communities, suggesting that certain ethnic backgrounds discouraged women from entering science (Hill et al., 1990; Parent & Oliver, 2015; Stokes et al., 2015).

Strategies that recommended race-blind strategies included Wylie and Gorman (2018), Chen and Simpson (2015), and Hill et al. (1990), who suggest avoiding gendered and raced descriptions of scientists in recruitment campaigns, instead focusing on the enthusiasm, sociability, and curiosity required in a scientist.

In a rare exception to the narrative that diversity is of universal benefit, Myers and Fealing (2012) noted that increasing the participation rate of minority women reduced the number of job roles available for White men, White women, and minority men. Their solution was to increase the overall number of scientific jobs available, rather than changing existing recruitment policies to favor women and minorities.

Gaps in the literature suggest that while recruitment strategies may be effective in the short term, they fail to address retention and career development problems once minority women have entered the organization. Furthermore, it is possible that the popularity of recruitment strategies is due to the reluctance of organizations to engage in strategies which may incur conflict between employees. Such organizational squeamishness in addressing gendered racism directly is a theme that recurs through the findings.

Mentoring

Following recruitment, mentoring (25% of articles) was the next most popular recommendation from within the organizational arena. Most studies described race- and gender-blind mentoring programs. Such strategies had benefits for the individual, including confidence (K. A. Robinson et al., 2018), increased sense of belonging (San Miguel & Kim, 2015), professional opportunities (S. N. Williams et al., 2016a, 2016b), enthusiasm for science as a career (Hill et al., 1990; Sax et al., 2018), and scientific identity (Cain & Trauth, 2015; Hodari et al., 2014). Only one study examined the benefits to an organization (Windeler & Riemenschneider, 2013), finding that increased successful mentoring programs increased organizational commitment and loyalty, thereby reducing employee turnover. Three studies actively discouraged gender- and race-specific mentoring (Parent & Oliver, 2015), (Hernandez et al., 2017), and (Sax et al., 2018), instead recommending matching mentors and mentees based on shared values.

The limitations of gender- and race-blind mentoring schemes are noted, even by those studies which recommended mentoring as a solution. These studies comment that science culture valorizes objectivity, which can then manifest as racial and gendered color blindness, whereby students are asked to ignore their racial and gendered experiences (Taylor, 2018). Resentment and misunderstandings were created by gender and racial mismatches between mentor and mentee (Hodari et al., 2014; W. H. Robinson et al., 2015, 2016).

Such limitations of traditional mentoring schemes were addressed by studies which matching a mentee with a mentor by both race and gender (Bowen, 2009; C. Grant et al., 2010; W. H. Robinson et al., 2015). Morales et al. (2019) suggest that where gender and race matching is not possible, mentors should be exceptionally committed to promoting professional opportunities in order to mentor women and non-White people. Crawford (2015) provides a detailed consideration of how to alleviate the failure of traditional mentoring schemes to benefit women of color: she suggests network mentoring, whereby a mentee has a number of mentors who they will select based on their individual requirements.

There are a few exceptions to the overwhelmingly positive literature on mentoring. Buzzanell et al. (2015) do not recommend mentoring as a solution, due to the emotional toll from mentees’ subject positions as women, junior, and people of color. A class-based critique is offered by Bowen (2009) who finds that middle-class women of color are more likely to express dissatisfaction with mentoring schemes than working-class women of color, theorizing that their higher expectations of a mentor enable them to more confidently critique their experience.

The literature suggests that mentoring, as a strategy to counter gendered racism, can on occasion ignore the reality of unequal power dynamics between mentors and mentees. The prevalence of such critiques suggests that, in the case of women of color in STEM research careers, mentoring is of doubtful effectiveness as an organizational strategy.

Retention Policies

Despite the importance of retention in ensuring that women of color remain the workplace, retention policies were less represented in the literature (31%) and exhibited less consistent set of strategies.

Gender-blind strategies included improving the work-life balance for women scientists (Cech & Blair-Loy, 2019; Cipher et al., 2017; Turner et al., 2018) including part-time working and paid childcare (S. M. Malcom et al., 1976; Turner et al., 2018). Such policies were found to be of limited benefit to women of color scientists, who are less likely than White women, minority men, or White men to have children (Ong et al., 2011; symposium, 2011).

Within organizations, Tao (2018) recommends eliminating the wage-gap between racialized women and other groups of scientists in order to establish equity. Windeler and Riemenschneider (2013) suggest that, until women and minorities are included in senior level roles, these groups should be actively involved in strategic decision-making, as feeling included in organizational-level decision improves staff retention. While such policies would be of limited use to women of color scientists, they would do little to address the isolation and harassment which drive many women away from science careers.

In an important, but rare intervention, Clancy et al. (2017) recommend enforcing existing policies on bullying and harassment, rather than creating new retention strategies for racial and gendered harassment. This recommendation builds on the literature which suggest that safe and equitable cultures within a department, laboratory, or team were seen as essential to retaining minority women (Gibbs & Griffin, 2013; Gibbs et al., 2014; Hodari et al., 2014; Payton et al., 2018). The role of leaders was considered vital in establishing an inclusive culture (Clancy et al., 2017). Such interventionist leadership would be of vital importance in using their authority to sanction abuse as well as in role-modeling appropriate behavior.

Beyond the organization, retention could be improved through external career coaching in order to broaden networks for women of color scientists (Kameny et al., 2014; Mathur et al., 2018a, 2018b; Mathur et al., 2018a, 2018b). Creating paid positions in professional organizations, which are largely run by volunteers, could explicitly cater for the needs of raced and gendered scientists (Antón et al., 2018 and Ong et al., 2011).

The literature on retention policies suggests that the active enforcement of norms of respect and civility would be helpful to minoritized women scientists. However, strategies stopped short of recommending an accelerated promotion of minority women to senior positions in their employing organizations or in scientific societies.

Training

The literature discussed above points to an aversion to challenging racist or misogynistic staff members; therefore, it is not surprising that training strategies were the least represented in the organizational arena (16%).

Psychological approaches largely discounted structural race or gender discrimination as a factor in why women and minorities are under-represented in the sciences, preferring instead to focus on personality traits. Gwillian and Betz (2001) and Casad et al. (2018) recommend training to improve a perceived misalignment of women and minorities’ personality type to a scientific career. Gwilliam and Betz (2001) suggest that improving the self-efficacy of minorities would improve their ability to participate in science and mathematics academic programs. Casad et al. (2018) suggest that women and minorities in science would benefit from training to improve their competence, belonging, and sense of purpose. The further implication is that the problem lies with the women of color, rather than institutional gendered racism. Training and professional development strategies aimed at increasing persistence in STEM among those from racially underrepresented groups and women have been rooted in a deficit model. Such a model aims to “fix” individuals through programs such as improving confidence in applying for grants or socializing them into a STEM culture, but do not seek anti-racist or gender-based reform at the lab, departmental, or institutional levels. Efforts rooted in a deficit model may advance individuals, but fail to alter perceptions of STEM as a welcoming place for all, instead indicating that women of color scientists should acclimate to a largely unchanging STEM climate (i.e., hyper-masculine, cis-gender, White, and often middle class and heterosexual) or fit in by hiding core parts of their social or cultural identity.

In contrast, authors who take a social justice position argue that scientists require training in order to support colleagues from different races, classes, and gender backgrounds. These authors note that science careers are characterized by a survival-of-the-fittest culture (E. O. McGee & Bentley, 2017) whereby success is attributed to individuals’ science abilities rather than to the race and gender privilege (Benore-Parsons, 2006; Cantú, 2012).

Maya’s frustrations associated with being a woman of color at her institution were unacknowledged. When she talked to her advisor, a White man, about the lack of women of color in the lab and the uneasiness she endured in that space, he told her, “Just be the person you need to be and don't worry about the color of your skin and your sex. Just worry about the engineering. (E. O. McGee & Bentley, 2017, p. 276).

Anti-racist training is recommended to introduce scientists to concepts of class, race, and gender (Bowen, 2009; Clancy et al., 2017). Such training is rooted in concepts of social justice (Wilkins-Yel et al., 2018), rather than apolitical models such as unconscious bias training. Such training would discuss explicitly sociological and political concepts (Corley et al., 2019): power, White privilege, anti-Blackness, the model minority myth, institutional racism, and internalized oppression (E. O. McGee & Bentley, 2017).

A theme running through the literature is that organizational strategies largely fail to address the need to create career-enhancing relationships between powerful White men and women of color. Such strategies also fail to increase the agency of women of color, instead reinforcing negative self-image. As will be revealed, effective strategies, whether in the organizational or other arenas, require both race and gender to be explicitly addressed.

Arena 2: The Intersubjective

Before discussing the intersubjective strategies in detail, Table 7 shows the distribution of papers across topic areas.

Table 7 List of papers organized by types of intersubjective strategy
Full size table

Of all the arenas within the science research environment, the intersubjective arena provided the fewest number of strategies: 14 papers (see Table 7), as compared to the 57 papers in the organizational arena, 24 in the experiential arena, and 18 in the representational arena. This result is disappointing, as research science is a collaborative endeavor, requiring trust between colleagues. The need for such strategies which address relationships between individuals is pressing:

The nature of “interpersonal relations”—including isolation, racism, sexism, being racially/ethnically identifiable, and relationships with faculty and other peers—caused more difficulty for women of color than structural barriers such as financial aid, recruitment practices, composition of the faculty body, tutorial and counseling support, and teaching or research assistantships. (Ong et al., 2018, p. 207).

The nature of intersubjective relations is that they can take place outside the official gaze of organizations, occurring between people in unsupervised situations. The intersubjective arena is where the microaggressions, exclusions, and stereotyping occur that makes daily life difficult for women of color scientists. Although organizational policies are intended to cover all aspects of working life, the reality is that women of color are reluctant to invoke formal sanctions for fear or repercussions from the aggressor (who may be a line manager) or of being labeled “difficult.” Strategies offered by papers focusing on intersubjective relations are those that can be forged by groups of individuals, often outside the formal, organizational arena.

Peer Networks

A slight majority of papers (64%) in the intersubjective arena recommended peer networks, formed between women of color scientists, sometimes across institutions and outside professional bodies. Such networks provided alternative forms of support which avoided reliance on unhelpful advisors (Alfred et al., 2019) and microaggressions from White, male peers (Bruning et al., 2015). Peer networks offered graduate students the opportunity to help junior graduates navigate the politics of graduate school (Solem et. al., 2009; Thakore et al., 2014) and to provide advice about study, research, and publishing (Clancy et al., 2017). Women of color from working-class backgrounds found peer networks particularly helpful in offering social and cultural confidence:

One component, out of many, of being an immigrant—in particular, an immigrant whose parents don’t have access to a lot of resources—is that you don’t come with a lot of social capital… And so, I didn’t have, kind of access to this whole set of social capital at that time so [the peer network] made a difference. (Hodari et al., 2014, p. 86)

Such shared personal experience with other women of color was considered more important than other forms of support by one participant:

Her major concern at this point in her life was to find a space to be with other people like herself…she needed other people who were racially similar to herself. (Bruning et al., 2015, p. 18).

Inclusive physical and social spaces are crucial in establishing supportive peer networks (Wood et al., 2016).

Safe Spaces

In addition to the virtual spaces of peer networks, the literature defines safe spaces as gender- and race-balanced physical or social places. A large minority of papers (42%) in the intersubjective arena recommended the creation of such spaces within an organization, to enable women of color could work, study, and interact without fear of gendered racism (Alfred et al., 2019; Bruning et al., 2015; Cantú, 2012). Such physical spaces are non-hierarchical, which are argued to be more conducive to learning (Buzzanell et al., 2015), support the mental health of junior researchers, and are a crucible for role modeling a collaborative academic environment (Ong et al., 2018).

The prevalence of sexual and racial harassment in the field, in the laboratory, and at conferences has led to women of color avoiding White-, male-dominated spaces, with resulting damage to their personal and professional development. The lack of safe spaces is a causation factor for minority women to drop out of STEM fields. Authors blame a culture of complacency within science which has tolerated “a significant failure in the [scientific] community to create safe working conditions for all scientists.” (Clancy et al., 2017, p. 1619).

Lab spaces (Solem et al., 2009) and study groups (Rice et al., 2013) were considered the most important environments for women of color scientists. Societies, clubs, and informal mentoring relationships also create safe spaces, whereby marginalized individuals can other see “brown faces, build a sense of community and security of belonging to the campus and, by extension, to science” (Ong et al., 2018, p. 233). In addition to physical spaces, co-created social spaces where activism took place were considered crucial to generate self-respect and a sense of agency.

Activism

A minority of papers (35%) in the intersubjective arena recommended activism as way for minoritized women to exercise agency (Seron, 2016) while refusing to passively accept hostile cultures (Alfred et al., 2019). Given that students join the sciences in order to improve the world around them, such activism is well aligned with altruistic motives (Sax et al., 2018). The prevalence and severity of gendered racism were found to be a catalyst for activism among women who have no prior experience of political activity (Hodari et al., 2014). Those who embraced activism found it had not harmed their career, but had instead clarified their intersectional identity and validated their feelings of exhaustion in constantly unpicking the nature of discrimination:

Activism helped me understand that [Race and gender] aren’t separate in me. I am always black and female. I can’t say, ‘Well, that was just a sexist remark’ without wondering would he have made the same sexist remark to a white woman. So, does that make it a racist, sexist remark? And that takes a lot of energy to be constantly trying to figure out which one it is. (Ko et al., 2013, p. 223).

Authors who suggest intersubjective strategies such as peer networks, safe spaces, and community activism argue that such “counterspaces” are a major contributory factor to the retention of women of color in science. Creating welcoming areas where individuals can relate to each other with respect and agency offers sanctuary from the daily microaggressions and isolation that is the norm of a White, male-dominated workplace. The development of academic and social safe spaces allows underrepresented students to validate their experiences, vent frustrations, and challenge deficit notions of women of color. More so than formal institutional initiatives, counterspaces are argued to establish a collegiate racial climate which assists with science retention.

Arena 3: The Representational

Before discussing the representational strategies in detail, Table 8 shows the distribution of papers across topic areas.

Table 8 List of papers organized by types of representational strategy
Full size table

The representational arena refers to images, texts, and information flows relating to intersectional identities within an institutional framework (Anthias, 2013). The representational relates, in this study, to how minority women scientists describe themselves and are represented visually within their organizations and in wider society. The representational arena contained only 18 papers, the second smallest number out of the four arenas. Given the importance of data and images to understanding gendered racism in the sciences, more research in this arena is required.

Dissemination of Research on Gendered Racism

The majority of articles in this study, including those in the representational arena, point to the lack of published research on the experiences of minority women. Yet only 2 out of the 18 articles in the representational arena (11%) explicitly recommend publishing research into gendered racism in the sciences. Since Malcom et al. (1976) first recommended that the experiences of minority women scientists should be understood more widely, journal editors and book publishers continue to reject articles which examine race and gender discrimination in science:

We found many, many dissertations. When I asked my researcher to find out how many had been published, what they had published, the answer came back as zero. There’s not a knowledge gap. It’s a serious gap in publishing, in being able to get the word out (Ong et al., 2011, pp. 156–157).

Given that policy-makers and organizations state the importance of improving diversity in the sciences, the publication and dissemination of research into gendered racism in academic journals are essential.

Disaggregated Data

A further data issue is the lack of disaggregated data which allows race and gender to be analyzed separately. Disaggregated data has been considered a problem since the 1970s (S. M. Malcom et al., 1976) when the dangers of treating a group, such as women of color scientists, as monolithic was first pointed out. Assuming that such a group has a single identity fails to account for differences in class, ethnicity, sexuality, and country of origin, perpetuating common misconceptions and excluding the voices of subgroups. Crenshaw’s concept of intersectionality was formulated to highlight the problem of treating all women as monolithic.

Providing legal relief only when Black women show that their claims are based on race or on sex is analogous to calling an ambulance for the victim only after the driver responsible for the injuries is identified. But it is not always easy to reconstruct an accident: Sometimes the skid marks and the injuries simply indicate that they occurred simultaneously, frustrating efforts to determine which driver caused the harm. In these cases the tendency seems to be that no driver is held responsible, no treatment is administered, and the involved parties simply get back in their cars and zoom away (Crenshaw, 1989, pg. 149).

In Crenshaw’s metaphor of a car accident, it is not easy to understand whether the harm suffered by a Black woman is due to her race or to her gender, and failing to understand that distinction means that woman’s injuries remain untreated. Disaggregating data allows us to, metaphorically speaking, better reconstruct the accident in order to determine who is responsible.

Subsequently, calls have been made for employers to collect data on minority scientists that can be disaggregated by race, gender, and ethnicity (Ribeiro & Schlege, 2018). A minority of papers (28%) in the representational arena call for disaggregation of data by race and gender, with Leggon noting that “data collection is crucial to assess equity projects in the sciences” (2006, p. 331). Yet even when disaggregated data is collected, the paucity of minority women scientists leads to problems of anonymity (Antón et al., 2018; L. E. Malcom & Malcom, 2011). Minority women are less likely to share their data because their small numbers in a sample mean that their stories will not remain confidential.

Narratives from Minority Women Scientists

Turning to qualitative research, 28% of papers in the representational arena highlight the importance of using explicitly sociological or political lenses in analyzing narratives written by minority women scientists. Structural analytical frames uncover the limited opportunities minority women have for displaying agency in White male-dominated environments (Buzzanell et al., 2015; Cantú, 2012; Ginther, 2018; Heward et al., 1995). Highlighting the complex positionality of authors who are also women of color, such theoretical lenses reject the existence of an objective, simplistic narrative. Individual narratives have the power to highlight the interconnected nature of race and gender within an organization as experienced by the individual:

Specifically, our three women engineers did not separate their stories of mentoring from other aspects of their organizational lives. (Buzzanell et al., 2015, p. 16).

Narratives by minority women scientists are written to value their lived experience and as a way of regaining control of how they are perceived (King Miller, 2017).

Increased Visual Representation of Minority Women Scientists

There were few studies recommending increased visual and media representations of minority women scientists to counter-effect racialized gender stereotypes. Of the four papers to do so, Rosenberg-Kima et al. (2010) point to the persistence of racial stereotyping in science, recommending that the visual effect of senior minority women would improve a sense of belonging in their junior counterparts. Antón et al. (2018) suggest educational and recruitment videos, delivered by minority scientists, to increase public perceptions of minorities as scientists.

Strategies in the representational arena require funding for inclusive marketing campaigns, media resources (S. M. Malcom et al., 1976). Intersectional research publishing and collecting intersectional data require a change of strategic direction from publishers, research funders, and employers (Ginther, 2018). These strategies require organizational influence to invest in buying art and advertising space, which is beyond the reach of most minority women scientists.

Turning to the experiential arena, the paper goes on to investigate individual, rather than institutional strategies, where minority women are able to exert more influence.

Arena 4: The Experiential

The experiential arena relates to the lived experienced of intersectionality, including the affective, the emotional, and the body. This arena also examines how individuals make sense of their identity, develop strategies to further their identity, or create coping strategies to resist othering. Before discussing the organizational strategies in detail, Table 9 shows the distribution of papers across topic areas.

Table 9 List of papers organized by types of experiential strategy
Full size table

Developing a Science Identity

A minority of papers (38%) in the experiential arena recommend developing a science identity. The concept of science identity was developed by Carlone and Johnson (2007) in a rare theoretical addition to the literature of gender discrimination in the sciences. Race and gender identity are posited as being central to the formation of a scientist, as highly visible aspects of a scientist’s identity which affect the extent to which the scientist is socialized into their immediate scientific community.

Science identity is composed of three interrelated concepts: performance, consisting of ways of talking, and using scientific tools; recognition, which involves being recognized and recognizing others as a scientist, and competence; and knowledge (not always publicly demonstrated) required to understand science content. The importance of teams and the prevalence of communal working (in labs and fieldwork) magnifies the isolation of minority women scientists who do not perform science in the same ways as their White, male peers (D. R. Johnson, 2011a, 2011b), and who may not be offered the same opportunities to demonstrate their competence.

Historically, Black colleges and universities (HBCU) were found to offer supportive climates for women scientists, particularly in allowing them to demonstrate their scientific skills and tools in open and welcoming lab settings (Ong et al., 2011). Remembering their passion for science enabled minority women to retain scientific identities (Alfred et al., 2019) and to remain in science careers (S. M. Malcom et al., 1976). First-generation immigrants were more likely to have a strong science identity than second- or third-generation immigrants due to their majority racial status in countries of origin and ensuring confidence in their self-image as a scientist (Lung et al., 2012; S. M. Malcom et al., 1976). The literature suggests that science identities are not formed in isolation but require supportive contexts to reinforce minority women’s self-belief in themselves as scientists (L. E. Malcom & Malcom, 2011).

Pride in a Racialized Gender Identity

A minority of papers (28%) in the experiential arena suggest that minority women take pride in their racialized gender identity (Beoku-Betts, 2008; Puritty et al., 2017). Such women prized their racial and gender difference, developing science identities which foregrounded independent thinking (Bruning et al., 2015; Trauth et al., 2012), giving back to their community (Cantú, 2012), and passing on their racialized and gendered knowledge to junior scientists (Hodari et al., 2014; Wood et al., 2016).

Developing Coping Strategies

A small minority of papers (14%) in the experiential arena recommend developing coping strategies to resist stereotyping. Race theorists have defined “whitely” behaviors are those which (re)produce whiteness as a normative racial identity, as well as maintaining the invisibility of White privilege (Tate & Page, 2018). Within science, adopting whitely behaviors (Bowen, 2009) and masculine behaviors (Remich et al., 2016) allows minority women to temporarily align with racial and gender norms:

When I am at school, I act white. I dress conservatively. I avoid speaking Spanish. I can’t change that I am a female, but I can make them stop assuming I am a dumb Mexican girl. (Bowen, 2009, p. 39)

Avoidance was another coping mechanism, whereby minoritized women avoided going into physical workplaces where they might encounter perpetrators of abuse (E. O. McGee & Bentley, 2017). Coping strategies, while dealing with immediate dangers, are maladaptive in the long term as women suffer stress from adopting inauthentic identity positions and end up leaving mainstream scientific networks (W. H. Robinson et al., 2015).

Improving Self-Efficacy, Outcome Expectations, and Persistence

Finally, 33% of papers in the experiential arena recommend self-improvement as a strategy for women of color to remain in the science workplace, including developing the personal traits identified by social cognitive career theory (SCCT; Lent, Brown, and Hackett 1994, 2000). SCCT is a race- and gender-blind psychological theory which posits that cognitive and behavioral factors influence the development of career interests, choices, and behaviors. Only one paper provided an intersectional analysis of SCCT, Byars-Winston and Rogers (2018), found that SCCT’s concepts such as self-efficacy are of limited value in relation to understanding the career intentions of minority women. All other papers reported results for women and minorities separately (Chen & Simpson, 2015; Lee et al., 2015; Lent et al., 2018; Sheu et al., 2018). Strategies focus on improving traits found in successful scientists: women and minorities are encouraged to improve self-efficacy (Kvasny et al., 2011), intention to persist (Lee et al., 2015), and mastery of science tools. Such self-improvement strategies are aligned with the predominance of deficit culture within STEM and the assumption that women of color are unable to advance due to their lack of skills. Self-improvement recommendations in the literature come close to recommending that women of color behave more like White men, or indeed, to denying that gendered racism has any bearing on the absence of women of color in science.

In the experiential and representational arenas, solutions are developed by women of color who create psychological techniques and political identities to counter exclusion. When compared with the organizational arena, such strategies are far from the traditional training and mentoring techniques so popular with HR managers. Instead, these representational techniques reinforce the cultural self-affirmations created in the safe spaces of the interpersonal arena. Yet, both representational and interpersonal strategies expend time and energy for women of color, above what is required by their scientific careers.

Discussion

The first research question aimed to provide a scoping review of intersectional strategies (both organizational institutional and individual) to address race and gender discrimination within science careers. The literature demonstrates that organizational strategies employed by institutions generally address either race or gender (Hancock, 2007), or are entirely race and gender blind, such as improving work-life balance. Additionally, there is a lack of representational strategies by organizations to gather disaggregated data and publish research into gendered racism, which required to develop effective countering strategies. The literature demonstrates that race discrimination and gender discrimination combine within science institutions to form mutually reinforcing axes of discrimination for the individual woman of color. The woman of color scientist finds it impossible to disentangle her experience of gender discrimination from her experience of race discrimination (Cantú, 2012; King Miller, 2017). The over-representation of individual strategies to cope with gendered racism in the intersubjective and experiential spaces suggests that the onus remains on women of color to adapt themselves to their environment (National Academies of Sciences Engineering & Medicine, 2019). The corresponding lack of institutional support structures echoes recent findings that organizations tend to expect the individual scientist to change, rather than the institution (Gomez et al., 2021).

This paper echoes long-standing calls that strategies to address the under-representation of women of color in science should address both race and gender discrimination (Clancy et al., 2017; S. M. Malcom et al., 1976; E. O. McGee & Bentley, 2017). Policy-makers and diversity professionals should therefore develop interventions which address the effects of both race and gender.

The second research question relates to the value of Anthias’ intersectionality model as an analytic tool. The benefit of Anthias’ model is that policy strategies adopted in the organizational arena can be compared with adaptive strategies by individuals which are adopted in more intimate and informal arenas. The effects of organizational strategies which fail to confront the reality of institutional gendered racism play out in the intersubjective arena (of interpersonal relationships) and the experiential arena (of the lived experience of women of color scientists). Anthias’ model adds value through providing a lens (the social space) for the analysis of responses to gendered racism, the range of places and spaces in which intersectional interventions can occur. The categories are also valuable in allowing organizations and individuals to select from a range of strategies, from resource-intensive training, and mentoring programs to those which can be implemented at no cost by an individual or by a team. One weakness of Anthias’ model is that it fails to address the “cultural” space: the set of social practices, often unspoken, which pervade a given area. In the case of science research careers, this would include definitions of academic excellence that are raced and gendered (Gabriel & Tate, 2017), a middle-class environment which isolates non-White working class women (Rickett & Morris, 2021), an alcohol-based social culture, which excludes Muslims and pregnant women. Conceptualized by Bourdieu as “habitus” (Bourdieu, 1977; Eagleton & Bourdieu, 1992), the “cultural” space incorporates an assumed way of being, which often excludes non-White, female bodies. Future theorizing based on Anthias’ model could pay attention to the ways in which such a culture reproduces over time and generates methods to exclude individuals who do not adopt its rules.

Further Research

The majority of papers discussed strategies in the organizational arena (57 papers) with the fewest number in the intersubjective arena (14). More strategies, and further studies, in the arena of relations between individuals are required, particularly given the prevalence of microaggressions (Payton et al., 2018; Wilkins-Yel et al., 2018), sexual assault (Clancy et al., 2017), and isolation (Ong et al., 2011).

Methodologically, the majority of papers were based on quantitative data or interviews. Creative methods such as autoethnography (Liu, 2018), creative writing in the form of plays (Dar, 2018), poetry (Biehl-Missal, 2015), and other forms of writing have been used to humanize the lived experience of gendered racism. Deploying different methodological strategies would broaden our thinking about how to build more equitable organizations (Brewis & Williams, 2019). A further methodological barrier to overcome is that quantitative studies exclude the most vulnerable of minority women scientists, who exist in small numbers. Policy-makers should be willing to introduce strategies based on qualitative data or to commission action research, which would account for, but not exclude, the small population of minority women scientists.

Finally, more research is required which recognizes the political origin of intersectionality and its radical potential to change social practice: only 12 papers mentioned the concept of intersectionality explicitly. Research which acknowledges the centrality of structural inequality to intersectional theory (J. K. Rodriguez et al., 2016) would fruitfully analyze race and gender discrimination within science in relation to class, colonialism, and capitalist structures of exploitation.

Conclusion

In conclusion, the literature on gendered racism in science careers implies that, while there is clearly disagreement as to which strategies are the most effective (L. E. Malcom & Malcom, 2011), cross-arena strategies would incorporate policy-type interventions with individual efforts. Anthias’ model is found to be a valuable structuring tool in analyzing solutions to gendered racism across a range of social arenas, as “no one solution alone will sufficiently improve the hostile workplace climate” (Clancy et al., 2017, p. 1620). Most papers (56%) recommend strategies which fall into more than one arena, suggesting that anti-discrimination requires multi-level, far-reaching, and complex strategies. Anthias’ model adds further value in illuminating the interdependence between arenas: for example, developing a science identity is not only a task for the individual woman in the experiential arena; her organization is also required to create an inclusive environment which supports the identity of women of color. Ultimately, none of the strategies examined can be said to be entirely successful, as the number of women of color in science remains low and, in certain areas such as computing and engineering, continues to decline.