3

Educational Interventions to Improve Recruitment and Retention¹

The analysis draws substantially from the research paper by Drs. Evava Pietri, Leslie Ashburn-Nardo, Corinne Moss-Racusin, and Jojanneke van der Toorn, which was commissioned for this study. The full research paper can be found online at: nap.edu/catalog/25585.

Recruiting and retaining more women in science, technology, engineering, mathematics, and medicine (STEMM) fields will require changes to the status quo in STEMM education, both in terms of the way students are taught and the experiences they have with faculty, role models, and mentors. Though White women are well represented among degree earners in certain STEMM fields (e.g. life sciences, chemistry), women remain particularly underrepresented in math-intensive STEMM disciplines such as engineering, computer science, and physics as early as the undergraduate level. Further, women express waning interest and self-efficacy in these fields at even earlier educational stages, despite the fact that there are no differences in average math performance between girls and boys in K-12 education or women and men in college math performance (see Chapter 2). Women of color remain underrepresented among undergraduate degree earners in all STEMM fields, including those disciplines in which White women are well represented. Given the national need for a greater number of STEMM professionals in many disciplines (particularly computer science and engineering), it is critical to identify strategies to improve recruitment and retention of women in educational programs in these fields. Fortunately, research offers a picture of the strategies educators and administrators can use to improve recruitment and retention of girls and women in STEMM education.

In the sections below we review the current research on interventions that can serve to promote recruitment and retention of women in STEMM with a

___________________

¹ This chapter builds on the significant contribution of the Committee on Understanding and Addressing the Underrepresentation of Women in Particular Science and Engineering Disciplines.

particular focus on effective educational strategies used throughout K-12 and undergraduate STEMM education and the positive impact of role models and mentors (see Chapter 4 for a discussion of the important role of sponsors). Many of these interventions are effective because they challenge stereotypes about who can be a successful scientist, engineer, or medical professional, and about the nature of work in STEMM (see Chapter 2 for a discussion of “Cultural Association Between Masculinity and STEMM”) in ways that can mitigate biases against women and improve self-efficacy, belonging, and performance in these disciplines. Interestingly, research shows that many of the interventions described in this chapter, such as growth mindset interventions, active learning, and communicating the societal impact of STEMM, can serve to make these fields more attractive to both women and men and can benefit a range of additional underrepresented groups in STEMM, including underrepresented minority men and first generation college students (i.e., whose parents did not attend college).

It is worth noting, however, that much of the research presented in this chapter has not taken an intersectional approach; rather, it has tended to examine gender and race as distinct identities. While limitation in sample sizes may explain the paucity of reported research on women of color, it is difficult to evaluate the efficacy of specific interventions on women of color without these disaggregated data. The committee also acknowledges that there is not currently research on each of these interventions in the context of every STEMM discipline. However, as noted in Chapter 1, many of these interventions will likely be efficacious in a range of STEMM disciplines, as the research presented in this chapter does, for the most part, demonstrate similar positive outcomes associated with these intervention across different STEMM disciplines, including STEMM disciplines with very distinct cultures (e.g. biology vs. computer science). For those readers particularly interested in the current state of knowledge on the impact of specific interventions in the context of a specific discipline, the report offers a table that appears in Appendix A that provides an extensive review of interventions that have improved the recruitment, retention, and advancement of women in STEMM. This table provides detail on whether the intervention has been tested in STEMM and if so, in which disciplinary contexts.

EDUCATIONAL INTERVENTIONS IN STEMM CLASSROOMS

Active Learning

Reorganizing STEMM courses to incorporate active learning exercises (i.e., having students work in groups, using clickers) generally improves learning among all students (Freeman et al., 2014; Handelsman et al., 2007), and is particularly beneficial for women in STEMM. As one example, in a traditional lecture-based biochemistry class there was an achievement gap between male and female students, and incorporating active learning exercises alleviated this grade

disparity (Gross et al., 2015). Moreover, when women students took an introduction to computer science class with multiple group activities, African American, Hispanic, and White women persisted longer in the computer science major than those who took a traditional lecture-based introductory course (Latulipe et al., 2018). Thus, ensuring STEMM courses integrate active learning is one strategy to help retain women in STEMM majors throughout college. Research has also found that active learning can decrease the achievement gap between educationally and/or economically disadvantaged students (predominantly students of color) and advantaged students (predominantly White) in introductory biology courses (Haak et al., 2011).

Peer-led team learning (PLTL), where students work in small groups to solve course-related problems with a peer mentor (a student who has previously been successful in the course), is another active learning strategy that improves outcomes for women in STEMM (Dennehy and Dasgupta, 2017). Incorporating PLTL improves learning outcomes generally in STEMM classes (Streitwieser and Light, 2010; Wilson and Varma-Nelson, 2016), and is particularly beneficial for students that have been underrepresented in STEMM (i.e., women and underrepresented minorities; (Horwitz et al., 2009; Thiry and Hug, 2012). For example, when PLTL was implemented in introductory STEMM courses, it improved the completion rate of all students and specifically enhanced Latinx students’ completion rate (Hug et al., 2015; Thiry and Hug, 2012). Providing additional evidence, across eight universities, Horwitz and colleagues (2009) found that relative to female students who took a traditional lecture-based introduction to programming course, those who took a class with PLTL were more likely to enter, persist, and earn higher grades in computer science majors. PLTL also may encourage students to participate in helpful research experiences. In particular, Gates et al. (2015) examined the effectiveness of PLTL across primarily Hispanic serving institutions in introductory computer science classes, and found that PLTL not only improved students’ problem-solving skills, but also increased the likelihood of students assisting with computer science research (Gates et al., 2015).

Aside from the pedagogical benefits of active learning, working together on a task (via active learning exercises) can promote social connection with other students, engagement with a task, and belonging in the STEMM environment (Carr and Walton, 2014). The benefits of working in groups also has been demonstrated with pre-school children. Relative to pre-school children working on a STEMM task alone, children who worked in a group showed higher engagement and interest in the task (Master and Walton, 2013; Master et al., 2017). This research with pre-school children demonstrates that interventions to recruit women into STEMM majors and careers can be implemented early in the educational system. Indeed, one large-scale strategy to spark girls’ interest in STEMM disciplines where they are least represented (e.g., computer science, physics, engineering) is ensuring that girls are exposed to classes dispelling masculine

STEMM stereotypes in the fields early in their educational development (Cheryan et al., 2017) (see Chapter 2 for a discussion of stereotypical associations between STEMM and masculinity).

Growth Mindset

With regard to STEMM ability, students can either have a growth mindset/incremental mindset (i.e., have beliefs that they can improve and get better) or a fixed/entity mindset (i.e., have beliefs that their ability is fixed and cannot change) (Dweck, 1995). Across many years of research, Dweck and colleagues have demonstrated that having a growth mindset increases academic performance among middle school, high school, four-year college students, and community college students (Dweck, 2006; Yeager and Dweck, 2012). For example, Chen and Pajares (2010) found that middle school students with a growth mindset (as opposed to a fixed mindset) had higher self-efficacy and learning-focused goals, and that middle school boys were more likely to have growth mindsets than girls. Indeed, the more female high school students believe they have the capacity to be successful after setbacks, the more likely they are to major in physics, engineering, mathematics, and computer science in college (Nix et al., 2015). Moreover, compared to those who have a fixed mindset, women college students with a growth mindset about math ability indicated higher belonging in math, reported more attraction to math careers, and earned higher grades in math classes (Good et al., 2012).

Researchers also have demonstrated the benefits of implementing short mindset interventions, which provide evidence that ability is not fixed and can improve. For instance, middle school students who took part in a workshop discussing how the brain is malleable and intelligence is not fixed had increased motivation in math and improved math grades relative to students who did not complete the workshop (Blackwell et al., 2007). This mindset intervention was effective because it encouraged students to value learning and effort, and respond more positively to challenges (Blackwell et al., 2007). In another example, female seventh grade students who were mentored by college students promoting a growth mindset performed better on standardized math tests compared to students who did not receive this growth-focused mentoring (Good et al., 2003). Thus, having female middle school students undergo a growth mindset intervention may be one way to recruit them into STEMM majors.

Aside from women students, mindset interventions help other students who traditionally have been underrepresented in STEMM. For example, relative to those in a no intervention control condition, Black college students who underwent a growth mindset intervention had higher academic motivation and grade point averages (Aronson et al., 2002). Specifically, Aronson et al. (2002) found that after three sessions of advocating the malleability of intelligence, African American study participants were found to have “created an enduring and ben-

eficial change in their own attitudes about intelligence.” Further, this change resulted in improvements in their academic profile. As compared to control group participants, African American participants reported more enjoyment and value in their academics and received higher grades. While the intervention also positively impacted White students, the results were not as striking. The authors noted that over time, “African American students appeared to become more convinced of the expandability of intelligence, the White students’ attitude change did not persist” (Aronson et al., 2002).

In a larger-scale experiment involving 90 percent of first-year college students attending a public university, researchers found that compared to students in a control group, a mindset intervention increased the grades of Latinx students and reduced the achievement gap between Latinx and White students. Testing the effectiveness of this intervention across multiple academic environments (i.e., at a high school, public university, and selective private university), Yeager et al. (2016) also demonstrated that this growth mindset intervention improved the academic performance of first generation and underrepresented minority students relative to those who did not complete the intervention (Yeager et al., 2016). Taken together, this research provides compelling evidence that mindset interventions are scalable (i.e., can be implemented across multiple academic contexts) and have the potential to be beneficial for women with multiple negatively stereotyped identities in STEMM.

Communicate to Students the Societal Impact of STEMM

Steinberg and Diekman (2018) found that encouraging students to introspect on why (e.g., improving society) as opposed to how (i.e., running experiments) scientists conduct research in STEMM increases beliefs that STEMM careers broadly satisfy communal ambitions and enhance both male and female students’ positive attitudes toward those careers. Illuminating one such intervention, STEMM classes can incorporate helping-focused projects to encourage beliefs that STEMM fields value communal aims (Belanger et al., 2017). Both male and female students are more likely to believe that engineering classes that have a service learning component (i.e., during which students use what they learn in class to help their local communities) fulfill communal goals and in turn are more interested in taking these classes (Belanger et al., 2017). Incorporating service learning projects in STEMM classes, therefore, helps promote perceptions that STEMM fields advance communal goals, which can serve to recruit women into STEMM classes. In another study, explicitly describing biomedical research as aiming to improve lives sparked students’ motivation, among both women and men, to conduct biomedical research (Brown et al., 2015). Similarly, when class lectures are structured to emphasize how STEMM research and careers help others, female first year college students believed that STEMM careers advance communal goals and expressed more interest in these careers (Fuesting and Diekman, 2017).

Educators can also personally emphasize how their work in a STEMM field satisfies their communal motives (Chen and Pajares, 2010; Dweck, 2006; Emerson and Murphy, 2015; Good et al., 2012; Yeager and Dweck, 2012). As one example, watching women scientists present on the altruistic aspects of their research increases adolescent girls’ interest in science (Weisgram and Bigler, 2006). Learning about how a scientist’s daily tasks involve working with and helping others (as opposed to working alone) also encourages women college students’ attraction to STEMM fields (Diekman et al., 2011). Finally, women STEMM majors report being more interested in working with a faculty mentor who values communal goals compared to a mentor who values agentic goals (Fuesting and Diekman, 2017). Taken together, this research provides compelling evidence that presenting STEMM fields (including biomedical sciences and computer science) as communal enhances interest in STEMM and encourages recruitment of women in STEMM. It is noteworthy that many of these interventions enhanced interest in STEMM among both women and men (Brown, 2015; Steinberg and Diekman, 2018), suggesting that these strategies benefit any student who values helping others and do not inadvertently dissuade men from entering STEMM careers.

Instructor connections with students are also a critical predictor of whether women will feel welcome and will be successful in STEMM classes. Students generally are more engaged in active learning and earn higher grades in STEMM classes when they trust their instructor (i.e., believe their instructor cares about and accepts them) (Cavanagh et al., 2018). That said, even though encouraging trust and good relationships with students promotes engagement, it is important that STEMM faculty still work to challenge students. Compared to those with a growth mindset, math instructors with a fixed mindset are more likely to employ comfort strategies (e.g., assigning less work) for students with low math ability (Rattan et al., 2012). Comforting rather than challenging students leads students to believe that their instructors have low expectations for their success in math and harms their math motivation (Rattan et al., 2012).

A recent large-scale study further demonstrated the benefits of instructors with growth mindsets, and examined the performance of students across 634 STEMM courses (Canning et al., 2019). Relative to students who took classes with an instructor who had a fixed mindset, those who took classes with an instructor with a growth mindset were more likely to believe the instructor emphasized learning and development, were more motivated to do their best work, and, importantly, earned higher grades in the course. Moreover, the achievement gap between White and underrepresented minority students was twice as large in classes with fixed mindset instructors than in classes with growth mindset instructors (Canning et al., 2019).

Across another series of studies, Fuesting et al. (2019) found that when students believed their STEMM instructors have a growth mindset compared to a fixed mindset, they are more likely to believe that STEMM environments afford communal goals, which ultimately relates to higher interest in STEMM majors and careers. Finally, relative to those with a fixed mindset, instructors with a growth mindset also are more likely to adopt active learning exercises in their courses (Aragón et al., 2018), and growth mindset interventions are less effective in classes when teachers have a fixed rather than a growth mindset (Schmidt, 2015). Taken together, multiple studies suggest that training STEMM instructors to have a growth mindset will improve the performance of all students (not just women), and specifically will help recruit female students from STEMM classes into STEMM majors and careers.

A 2020 report by the National Task Force to Elevate African American Representation in Physics and Astronomy (TEAM-UP), which examined the reasons for the persistent underrepresentation of African Americans in these fields (AIP, 2020), also found that the characteristics of faculty influenced recruitment and retention. Specifically, the task force found that faculty behavior could influence the development of “physics identity,” defined as “how one sees oneself with respect to physics as a profession.” The task force noted that how students perceive themselves with respect to physics is predictive of achievement and retention in the field and identified faculty encouragement, recognition, and

representation as key aspects of fostering physics identity among students. The task force recommended that to build physics identity, departments should be strategic in determining whether departmental activities are supportive of physics identity and assess the efficacy of activities and the diversity of faculty across race/ethnicity/gender and other social identities. The task force also found that African American student retention in physics improved when faculty “recognize and respond to students as unique individuals with a wide range of intersecting social identities and acknowledge their experiences of being minoritized in physics and astronomy department may impact their academic performance.” Other recommendations to address underrepresentation of African Americans in physics, included ensuring that teaching, mentoring, and advising include a focus on African American student success (AIP, 2020).

Instructors also play an important role in retention of post-traditional students, who make up the majority of students in this country (see Box 3-1 for more information). Packard and Jeffers (2013) demonstrated that women from community colleges are more likely to persist in STEMM fields if they are given

intentional and proactive academic advising from an individual who has knowledge of particular STEMM fields and STEMM transfer possibilities (Packard and Jeffers, 2013). This advising can help to alleviate confusion in signing up for coursework and in selecting a major (Packard et al., 2012). It may be helpful if faculty and transfer advisors work together to share this information, as many community college women do not have extended time to navigate campuses to receive transfer and career advice (Wang et al., 2017). In a study examining women’s experiences in STEMM community college transfer pathways, Packard et al. (2011) identified several facilitators to women’s success, including inspirational professors, effective transfer advising, and academic resources. Students reported that finding a helpful professor or advisor boosted student belongingness and contributed to their persistence in STEMM.

Group Composition

The group composition of classes or small activity groups in class (for female students), and working groups (for female scientists) may also play an important role in recruiting and retaining women in STEMM. For instance, female students perform worse on a math test when they are in a setting with majority male students as opposed to majority female students (Inzlicht and Ben-Zeev, 2000). Women students anticipate less belonging and are less interested in attending a conference that has majority male students versus gender parity (Murphy et al., 2010). Women established in STEMM also anticipate less belonging and are less interested in an academic conference when nearly all of the attendees are men (Richman et al., 2011), and women working in STEMM environments where they are outnumbered by men experience the highest level of gender identity threat² compared to men and to women who are not outnumbered by men (van Veelen et al., 2019).

Although women may benefit generally from active learning, being in women-majority activity groups may create the most welcoming and inspiring STEMM classroom environments (Springer et al., 1999). For example, Springer, Stanne, and Donnovan (1999) conducted a meta-analysis of the effects of small-group learning on undergraduates in STEMM courses and found that this is an effective approach for promoting greater academic achievement, improving attitudes, and increasing persistence for women (Springer et al., 1999). The authors noted that the positive effect of small-group learning on students’ achievement was significantly greater for groups composed “primarily or exclusively of African Americans and Latinas/os (compared with predominantly White and relatively heterogeneous groups).” In a separate study, Dasgupta et al. (2015) found that students were more likely to participate and feel less anxious in women majority groups compared to male majority groups in an engineering class.

___________________

² Gender identity threat is the fear that their gender identity will be devalued (Steele et al., 2002).

The women students in the women majority groups also indicated higher STEMM career aspirations and confidence (Dasgupta et al., 2015). Beyond gender, related work has demonstrated that underrepresented minority students in STEMM also benefit from environments with other underrepresented minority students (Gates et al., 2011; Hurtado et al., 2007; Johnson et al., 2019). Hurtado et al. (2007), for example, found that among first year underrepresented minority college students in the sciences, several factors positively and significantly shaped their sense of belonging. These included interacting with a graduate student or teaching assistant, receiving advice from a junior or senior, receiving academic advice from a freshman, and interacting with peers of diverse racial backgrounds. Similarly, the authors noted that significant positive influence of cross-racial interactions on underrepresented minority students’ sense of belonging further supports the benefits of diversity on college campuses.

When it is not possible to have women majority groups, it also may be helpful to address the biases of the male students in STEMM classes. Women and underrepresented minority STEMM majors report facing unwelcoming environments in their STEMM class from fellow students (Hurtado et al., 2007; Robnett and Thoman, 2017; Steele et al., 2002). Particularly relevant to group activities, Meadows and Sekaquaptewa (2013) found that when working in groups in engineering courses, male students tended to take on active roles (e.g., talk more, present group work), whereas women tended to be in technical roles (e.g., note takers). Thus, bias literacy interventions may not only be beneficial when implemented among STEMM faculty, but may also promote more inclusive STEMM classroom environments when targeted toward students in those classes (Becker and Swim, 2011, 2012; Kilmartin et al., 2015). As one example of the benefits of classroom bias literacy interventions, in an experiment by Bennett and Sekaquaptewa (2014), introduction to engineering courses were randomly assigned to receive presentation on the importance of egalitarian social norms (i.e., intervention classes) or receive no presentation (i.e., control classes). Relative to those in the control classes, White male students who underwent the intervention at the beginning of the course, reported valuing diversity more and higher intentions to speak out against discrimination (Bennett and Sekaquaptewa, 2014). In addition, regardless of receiving the intervention, racial/ethnic minority males did not differ on their attitudes toward diversity in engineering. The authors attributed this finding to the limited sample size or the fact that this population’s attitudes may have already been more positive compared to White males.

Another successful intervention for students employed videos to demonstrate equitable classroom interactions (Lewis et al., 2019). Specifically, researchers assigned STEMM majors to watch a video of mixed gender groups conforming to gender stereotypes (i.e., male students speaking more than female students), or acting nonstereotypically (i.e., female students talking more than male students). The STEMM majors then completed a group task, modeled after typical STEMM classroom activities. In the interventions group, female and male students spoke

equal amounts, whereas in the nonintervention teams, male students spoke more than female students (as revealed from both self-report data and video footage of group interactions) (Lewis et al., 2019).

THE IMPORTANCE OF ROLE MODELS

Researchers have found that exposure to a woman scientist role model (i.e., a scientist that women feel similar to and aspire to be like (Gibson, 2004); enhances female students’ identification with and interest in STEMM (Ramsey et al., 2013; Stout et al., 2011), can change their personal beliefs about STEMM fields, and break stereotypical associations between men and STEMM (Young et al., 2013). Aside from changing perceptions of STEMM, however, when women interact with scientist role models, they also picture themselves becoming the scientists in the future (altering their possible future selves or their representations of who they could become in the future) (Lockwood and Kunda, 1997; Markus and Nurius, 1986; Markus and Kitayama, 2010). As a result, multiple theories have highlighted the benefits of role models for encouraging women’s attraction to STEMM.

For example, the Motivational Theory of Role Modeling (Morgenroth et al., 2015) posits that, because individuals aspire to be like successful similar others, role models act as inspiration to encourage individuals to value certain domains and be attracted to those fields (Paice et al., 2002). This model further asserts that by identifying with role models, individuals view role models as evidence that it is possible to succeed in a given area, and feel self-efficacious (Lockwood and Kunda, 1997). Stereotype Inoculation Model further argues that when women feel similar to scientist role models, the role models inoculate against threatening stereotypes about women in STEMM and indicate that women will be valued and belong in STEMM environments (Dasgupta, 2011; Stout et al., 2011). For example, research has also shown that it is important for women at community colleges to identify role models from similar backgrounds who have successfully completed the transfer process and are currently enrolled in a STEMM baccalaureate degree (Wang et al., 2017). Having these role models helps to dismiss self-doubting notions that these students are not capable or will not receive support if they pursue a baccalaureate STEMM pathway. Critical to both theories is that women must identify with the role model for the role model to be inspirational. In general, women are more likely to identify with and feel more inspired by female than male role models (Lockwood and Kunda, 1997).

Consequently, even brief exposure to a woman scientist role model enhances female students’ identification with and interest in STEMM (Ramsey et al., 2013; Stout et al., 2011). As one example, relevant to the early stages of recruitment, instructing middle school girls to reflect on and write about a role model they interacted with during a summer science program enhanced their sense of fit in STEMM relative to students who wrote about their best friends (O’Brien et al.,

2017). In another experiment at the college level, researchers randomly assigned women engineering majors to learn about successful men engineers, women engineers, or innovative discoveries in engineering (i.e., control information). Relative to students who learned about male engineers or control information, those who read about female engineers indicated higher self-efficacy and career motivation in STEMM. This finding is pertinent to recruiting female engineering majors into the STEMM workforce, as well as retaining female engineers from college into STEMM careers. With regard to retaining female scientists after college, having supportive role models in their workplace also encourages belonging among women established in their STEMM careers (Richman et al., 2011).

However, requiring women scientists to act as role models may create extra service requirements (i.e., by having them serve on panels or give guest lectures) and harm their research productivity (Guarino and Borden, 2017). Consequently, another strategy to ensure women see relatable role models without burdening women working in STEMM is featuring women STEMM professionals in movies and television shows. The “Scully effect” is one demonstration of benefits associated with women scientists’ representation in popular media. Specifically, researchers found that girls who consistently watched the X-Files television series in middle school and were exposed to the character, scientist agent Dana Scully, were more likely to express interest in STEMM, major in a STEMM field, and work in a STEMM profession compared to girls who did not watch the X-Files (Geena Davis Institute on Gender in Media, 2018). In a related study, researchers found that watching 10 short television clips featuring men and women scientists encouraged both male and female adolescents to picture themselves becoming scientists in the future (Steinke et al., 2009).

It is important to note that there are certain characteristics that result in role models being more or less effective for recruiting women into STEMM. For instance, researchers found that when female students who were not computer science majors interacted with a stereotypical female computer scientist (i.e., had masculine traits), these students reported lower self-efficacy and sense of belonging in computer science relative to students who interacted with a male or female counter-stereotypical scientist or who did not interact with a scientist (Cheryan, 2012; Cheryan et al., 2011a). Specifically, female students felt less similar to the stereotypical scientist than to the counter-stereotypical scientist, which feeling in turn, correlated with lower success beliefs and sense of belonging (Cheryan et al., 2011a; Cheryan et al., 2013). It is therefore unsurprising that female scientists who clearly value communal goals (as opposed to those who value agentic goals, such as being competitive, determined, and aggressive) are more likely to spark female students’ interest in STEMM (Clark et al., 2016; Fuesting and Diekman, 2017). However, the role of counter-stereotypical feminine role models depends in part on students’ perceptions of STEMM. Because of the masculine stereotypes associated with STEMM, being a feminine scientist may seem highly unattainable, particularly among female students with low STEMM iden-

tification³ (Lockwood and Kunda, 1997). Supporting this possibility, Betz and Sekquaptewa (2012) found that feminine STEMM role models reduced middle school girls’ current interest in math, self-rated abilities, and success expectations as compared with gender neutral scientist role models, particularly for girls who already disliked STEMM. Finally, women relate better to women scientists when they believe the scientists have had similar experiences and past challenges as themselves (Asgari et al., 2012; Pietri et al., 2018a).

It is also important to take into account that White women scientists may not be effective role models for women with multiple stereotyped identities because the identity of being a woman may not be the most salient of their intersecting identities with respect to STEMM (Pietri et al., 2018b). As one illustration of this possibility, because many Black women are more sensitive to the possibility of racism than sexism (Kirk and Olinger, 2003; Levin et al., 2002), they may identify more strongly with a Black male or female scientist than a White female scientist (Johnson et al., 2019; Pietri et al., 2018b). As a result, Black women anticipate more belonging in a STEMM company (Pietri et al., 2018b) or School of Science and Technology (Johnson et al., 2019) when they learn about a Black female or male scientist at the company or school compared to when they learn about a White female scientist. White female scientists, therefore, may not function as role models to recruit Black women in STEMM (Johnson et al., 2019; Pietri et al., 2018b).

With regard to retention, researchers found that among Black female college students majoring in STEMM, having Black women role models related to higher belonging in STEMM, whereas having White women role models did not predict belonging in STEMM (Johnson et al., 2019). Future research should continue to explore if White female scientists do or do not act as role models for women with multiple negatively stereotyped identities in STEMM. Nevertheless, these initial studies suggest that when presenting women with role models to broaden their future selves and spark self-efficacy, belonging, and interest in STEMM, it will be important that these interventions feature female scientists with multiple identities aside from gender.

CREATING INCLUSIVE RELATIONSHIPS THROUGH MENTORING

Scientists can act as role models and impact women’s perceptions of STEMM fields, even when women lack direct contact with the scientists. In contrast, scientists only function as mentors when women have consistent interactions

___________________

³ According to the 2018 National Academies’ study The Science of Effective Mentoring in STEMM, science identity is defined as “an identity that is connected strongly to science, including three overlapping dimensions—competence in one’s own mind and as judged by others, performance in terms of having the skills and opportunities to act like a scientist, and recognition by oneself and meaningful others.”

with the scientists, during which the scientists provide guidance and support (Gibson, 2004). Consequently, creating positive relationships is more important for mentoring than the previous described interventions, and encouraging these meaningful connections is critical to retaining women in STEMM majors. Indeed, having mentors during college is one of the best predictors of women’s reported involvement in their STEMM major (Downing et al., 2005; Hernandez et al., 2017), and, as discussed in Chapter 2, lacking mentors is a challenge for women in engineering, chemistry, and mathematics (Herzig, 2002, 2004a; Marra et al., 2009; Newsome, 2008). Once women graduate from college, continuing to build mentoring relationships is essential for the success of their career in STEMM (Allen et al., 2004; Eby et al., 2008). In light of the positive impacts of mentoring, there are a variety of professional organizations devoted to mentoring women in STEMM. Box 3-2 offers several examples.

Mentors help women in academic science grow and thrive in their careers by connecting women with potential collaborators and supporting both research and teaching (Misra et al., 2017). Thus, positive mentor relationships help women advance in their STEMM career (i.e., receive promotions, be successful in research and mentoring). Sponsor relationships also are useful interventions for advancing women in STEMM. Differing from mentorship, sponsorship does not involve emotionally supportive relationships, but rather is focused on providing opportunities to help women excel in their career by suggesting them for leadership positions and awards (Helms et al., 2016). Sponsorship is discussed further in Chapter 4.

Beyond providing support and career advice, mentors can also help enhance female students’ interest in STEMM by providing valuable research opportunities in STEMM laboratories. Research experiences in general encourage students (particularly those from underrepresented groups) to both enter, persist, and advance in STEMM majors (Graham et al., 2013; Gregerman et al., 1998; Imafuku et al., 2015; Junge et al., 2010; Linn et al., 2015). For instance, Jones and colleagues explored how research experiences in biology impacted students who were interested in a biology major at University of California, Davis. Compared to those who did not take part in research, those who participated in research persisted for longer in the biology major, were more likely to graduate with a biology degree, and earned higher grades in their biology courses (Jones et al., 2010b). A qualitative study conducted at a primarily Hispanic serving institution additionally found that students who were a part of affinity research groups (ARGs) in computer science reported that these groups helped them grow as researchers and professionals, and promoted their integration into the larger computer science community (Villa et al., 2013). Importantly, these ARGs were strategically designed to create a sense of community in research labs via team-building activities, which suggests that labs should carefully construct inclusive and welcoming research opportunities to enhance interest in STEMM.

Research laboratory environments can be structured to counteract masculine stereotypes and demonstrate how STEMM research can fulfill communal goals (Allen et al., 2018; Thoman et al., 2017). As one example relevant to recruitment, Thoman and colleagues (2017) surveyed a large sample of undergraduate research assistants across STEMM laboratories, and found that when lab culture values using science to helps others, underrepresented minority (URM) research assistants expressed more interest and motivation in STEMM. Mentors therefore may play a vital role in ensuring the success of women in STEMM; however, because of the pervasive masculine stereotypes in STEMM, both male and female STEMM faculty may be less interested in mentoring female students than male students (Moss-Racusin et al., 2012). Thus, it is important that interventions work to motivate scientists to act as mentors for women as well.

Mentorship for women of color is particularly important, as underrepresented students in STEMM are less likely than well-represented students to receive formal mentoring (Felder, 2010; Gayles and Ampaw, 2011; Johnson, 2015; King et al., 2018; Thomas, 2001). However, mentoring women of color requires attention to the identity-related challenges that their mentees may have, as well as skills to develop the talent of these mentees, while recognizing the racial and ethnic contexts they face (NASEM, 2019b). The majority of mentors in STEMM are likely to dismiss the idea that social identity shapes the experiences of mentees and approach mentoring with a color-blind approach (Brunsma et al., 2017; McCoy et al., 2015; Prunuske et al., 2013). However, in a study of primarily White mentors and undergraduate mentees from underrepresented groups, mentees were more likely than their mentors to want to discuss cultural diversity matters in the mentoring relationship. (Byars-Winston et al., 2019). In addition to mentees wants, research shows that culturally responsive mentoring helps validate students’ experiences, reinforces their self-efficacy in their fields, and increases the likelihood of succeeding in STEMM fields (Byars-Winston et al., 2010; Thomas et al., 2007; Vaccaro and Camba-Kelsay, 2018). Therefore, all mentors, regardless of background, should be culturally responsive to their mentees.

As discussed earlier in this chapter with role models, many underrepresented students also prefer to have a mentor with a similar background to their own (i.e. race, gender, ethnicity, LGBTQIA status, and more) (Blake-Beard et al., 2011; Williams et al., 2016). However, due to the dearth of mentors from underrepresented backgrounds in STEMM fields, matching underrepresented students with mentors of the same background may not be possible, and, when it is possible, could lead to unequal service burdens for the mentors (Armstrong and Jovanovic, 2017; Xu, 2008). While prior research does support a race and gender mentoring match, Blake-Beard et al. (2011) demonstrated that having shared beliefs, values, and interests is a better predictor of mentoring relationship quality. Additionally, having a mentor from a well-represented background may provide access to resources that would otherwise be difficult for underrepresented students to access. Therefore, mentors of different identities who are culturally responsive and

work to understand the experiences of underrepresented groups, may meet the needs of their mentees (Felder and Barker, 2013; O’Meara et al., 2013; Sanchez et al., 2014).

MALE ALLIES

To further alleviate the heavy service expectations for female scientists (Guarino and Borden, 2017), it is critical that male scientists play an active role in changing perceptions of STEMM (Akcinar et al., 2011). Indeed, by not conforming to agentic masculine stereotypes and by describing how their work fulfills communal goals, male scientists can inspire women’s self-reported interest in STEMM and promote the recruitment of female students into STEMM (Cheryan et al., 2011a, 2013; Clark et al., 2016; Fuesting and Diekman, 2017). Women also may feel more welcome in STEMM environments that have supportive male allies. For instance, researchers found that when Black female STEMM majors perceive having multiple ally professors and care about helping Black women succeed in STEM, they report higher belonging in STEMM (Johnson et al., 2019). This study suggests that allies may help retain female students in STEM majors, by ensuring that students feel welcome in STEMM. Additional work has found that White female participants perform better on a spatial ability task, when they think that task was created by an expert from a similarly negatively stereotyped group (i.e., a Black male expert) than by a White male expert (Chaney et al., 2018). This enhanced performance is in part explained by participants’ perceptions that the expert from a negatively stereotyped group is an ally who believes women have strong spatial abilities (Chaney et al., 2018). It is important to note that saying one is an ally and has a positive attitude toward a group may be beneficial, but not sufficient to elicit trust from members of that group (Dovidio et al., 2006; Hebl et al., 2009). Rather, one may need to perform a series of actions to signal commitment to helping that group (Ashburn-Nardo, 2018; Brown et al., 2015; Droogendyk et al., 2016), and hence researchers should continue testing how male scientists can effectively signal that they care about helping women in STEMM and are allies.

FINDINGS: CHAPTER 3

FINDING 3-1: To improve the representation of women in STEMM will require interventions to improve recruitment and retention of female students throughout their STEMM educational careers, including K-12. Women of color remain underrepresented among undergraduate degree earners in all STEMM fields, including those disciplines in which White women are well-represented (e.g. biology), and all women remain particularly underrepresented in math-intensive STEMM disciplines such as engineering, computer science, and physics.

FINDING 3-2: Reorganizing STEMM courses to incorporate active learning exercises generally improves learning among all students and is particularly beneficial for women in STEMM.

FINDING 3-3. Growth mindset interventions that impress upon students that skills and intelligence are not fixed, but, rather, are increased by learning, help all students, including those who have traditionally been underrepresented in STEMM, including women and underrepresented minorities.

FINDING 3-4. Encouraging individuals to introspect on why (e.g., improving society) as opposed to how (i.e., running experiments) scientists conduct research in STEMM increases beliefs that STEMM careers broadly satisfy communal ambitions and enhances both male and female students’ positive attitudes toward those careers.

FINDING 3-5. Even brief exposure to a woman scientist role model enhances female students’ identification with and interest in STEMM. However, requiring women scientists to act as role models may create extra service requirements for these STEMM professionals (i.e., by having them serve on panels or give guest lectures) and harm their research productivity. Consequently, another strategy to ensure women see relatable role models without burdening women working in STEMM, is ensuring women STEMM professionals are featured in movies and television shows.

FINDING 3-6. Having mentors during college is one of the best predictors of women’s reported involvement in their STEMM major. Lack of mentorship is a particular challenge for women in engineering, chemistry, and mathematics. Mentorship for women of color is particularly important, as underrepresented students in STEMM are less likely than well-represented students to receive formal mentoring.

FINDING 3-7: Male allies can promote the recruitment of female students into STEMM, however, additional research should explore the characteristics of effective allies.