A Goal Congruity Model of Role Entry,Engagement,and Exit: Understanding Communal Goal Processes in STEM Gender Gaps

Motivational responses

The effects of goal incongruity experiences on motivational engagement and disengagement depend on the extent of the incongruity experience. For short-term experiences of incongruity, the result is likely to be greater activation of the motive; for longer term experiences of incongruity, the result is likely to be disengagement from the motive (Custers & Aarts, 2007; Sheldon, 2011). In the short term, goal incongruity may heighten goal accessibility as individuals try to meet these goals. A fundamental aspect of motivation is that valued goals will be pursued until fulfilled. Long-standing research, dating back to the Zeigarnik (1927) effect, finds that unmet goals gain in strength and accessibility over time (e.g., Fishbach & Ferguson, 2007; Masicampo & Baumeister, 2011).

However, with repeated experience, goal incongruity may reduce goal accessibility: Sustained experience in a goal-incongruent environment leads people to improve their goal pursuit strategies (in which case the environment becomes goal congruent) or to disengage from goals. In part, this disengagement may occur through nonconscious processes resulting from the particular environments that people occupy. Individuals tend to activate and pursue goals that are associated with the immediate physical context (Aarts & Dijksterhuis, 2003) and social context (Aarts, Gollwitzer, & Hassin, 2004). Even if goals are explicitly endorsed by an individual, they may be automatically inhibited if perceived to conflict with the goals of the environment (e.g., Shah, Friedman, & Kruglanski, 2002). Furthermore, the negative affect resulting from failure to fulfill one’s valued goals may lead to nonconscious disengagement from these goals (Aarts, Custers, & Holland, 2007). In addition, experience with goal failure leads individuals to modify their expectations and behavioral strategies for goal pursuit (Carver & Scheier, 1998). Individuals who successfully disengage from unattainable goals experience greater well-being (Wrosch, Scheier, Miller, Schulz, & Carver, 2003).

If continued experience in a social role suggests that a valued goal will not be met, then the key question is whether to disengage from the goal or to disengage from the role. Goal engagement or disengagement then cycles back to feed into the motives that influence future role selection (as reflected in the top arrow in Figure 1). If goal disengagement is unlikely because the goals are fundamental, then role disengagement is favored as a strategy; if role disengagement is unlikely because role exit or reconstruction is not possible, then goal disengagement is favored as a strategy.

Role responses

Role decisions to seek or maintain goal congruity can take a number of forms. Most basic is that individuals can opt either to stay in a social role or to exit a social role: If the role is not perceived as affording valued goals, then individuals might leave the social role for another one that affords the goal (e.g., equifinality; Kruglanski et al., 2002). The literature on exit, voice, loyalty, and neglect in organizations (Rusbult, Farrell, Rogers, & Mainous, 1988; Withey & Cooper, 1989) and in relationships (Rusbult & Martz, 1995) suggests that people exit their roles when satisfaction is low, commitment is low, they do not see possibilities for change, and they have a more attractive alternative role. Individuals are more likely to speak up actively about the role problems (i.e., voice) when satisfaction is low, commitment is high, and they see a possibility for change.

However, exit decisions can be both psychologically and financially costly, and a range of possibilities exist for reconstructing or reconstruing the social role. Reconstruction involves behaviorally changing the role: For example, employees might negotiate different working conditions that afford valued goals, or organizations might structure opportunities for employees to do so. Reconstrual involves cognitively changing the content of the role: For example, students might mentally connect their activities to their broader goals. In short, individuals who are seeking congruity might try a range of strategies in their efforts to align their social roles with their valued goals, and the social structures they are embedded in might support or restrict these actions. These experiences then in turn cycle back to inform their next role selection decisions (as reflected in the bottom arrow in Figure 1).

As noted earlier, perceived goal affordances may or may not be accurate. Like any stereotype, even a perception that accurately captures a group average may be inaccurate when applied to any particular exemplar (McCauley, Jussim, & Lee, 1995). What is important is that either accurate or inaccurate goal affordances can influence social role selection. For the question of recruitment into social roles, then, the mere existence of a goal affordance belief (and anticipated goal incongruity) can shape approach to or avoidance of the role. For the question of retention into social roles, the accuracy of goal affordance beliefs (and experienced goal incongruity) is important: Individuals who enter into a social role with the expectation that they will fulfill valued goals, but find that they cannot, may select out of that social role and into a different one—one that they think will afford their valued goals.

Patterns of representation

The underrepresentation of women in STEM is acute within the United States. Although women hold nearly half (48%) of all jobs in the U.S. economy, they hold under a quarter (24%) of STEM jobs (Beede et al., 2011). In the United States, STEM occupations are particularly projected to grow over the next decade, relative to other occupations, but STEM degrees granted have declined relative to other degrees (U.S. Congress Joint Economic Committee, 2012). Women’s underrepresentation in STEM is particularly puzzling when considered alongside the fact that women are nearing equal representation in degrees earned in other challenging, male-dominated fields, such as medicine (49%) or law (46%; T. D. Snyder & Dillow, 2011). Since the 1970s, women have remained highly represented in occupations characterized more by engaging with people than with things, even as women have entered into high-status occupations, such as physician or lawyer (Lippa, Preston, & Penner, 2014). Female college students have strongly increased their interest in medicine and law (Astin et al., 2002): Indeed, the occupations of lawyer and physician were considered nontraditional for women in the 1980s but are not any longer (U.S. Department of Labor, 2006). In contrast, even talented women continue to opt out of many STEM fields, particularly computer science and engineering (American Association of University Women [AAUW], 2015). The key question is not simply why women are not choosing STEM fields but also what careers women are choosing instead (e.g., Jacobs, Davis-Kean, Bleeker, Eccles, & Malanchuk, 2005).

Considerable variability in gender representation exists within the domain of STEM as well (see Cheryan, Ziegler, Montoya, & Jiang, in press, for a review). Women and men within STEM tend to earn different types of degrees: Among male STEM degree holders, 48% hold engineering degrees and 31% physical and life sciences degrees; among female STEM degree holders, 18% hold engineering degrees and 57% physical and life sciences degrees (Beede et al., 2011). Within the science and engineering workforce, women are most prevalent in social science (53%) and biological and medical sciences (51%), and less prevalent in engineering (13%) or computer/mathematical sciences (26%; National Science Board, 2012). Female STEM degree holders tend to work in non-STEM occupations more than their male counterparts do. Of the 2.5 million female STEM degree holders, only 26% held a job in a STEM occupation, compared with 14% in education and 19% in health care. In contrast, of the 6.7 million male STEM degree holders, 40% held a STEM occupation, with 6% in education and 10% in health care (Beede et al., 2011). Women in STEM may thus be drawn to professions that use science and technology to directly benefit others. Women are well represented in biology and chemistry (50% and 59% of bachelor’s degrees, respectively; National Science Board, 2012), and these degrees can lead to careers in medicine and pharmacy (e.g., women are 48% of medical school graduates; Jolliff, Leadley, Coakley, & Sloane, 2012; women are 54% of pharmacists; U.S. Department of Labor, 2013). The disproportionate representation of female STEM workers in education and health care aligns with the communal goal perspective that careers that more directly afford opportunities to help others are more attractive to women.

Even after earning a STEM degree, women more than men select out of applying for research-intensive academic positions: Across six scientific disciplines, the proportion of female degree holders was larger than the proportion of female applicants for research-intensive tenure-track positions (National Research Council, 2010). However, although men are more likely to be hired into faculty positions, women who are hired tend to be retained at similar rates to men, except in mathematics (as shown in an analysis of approximately 3,000 faculty from 14 universities; Kaminski & Geisler, 2012). The general pattern of similarity in retaining female and male faculty led Kaminski and Geisler to conclude that “if women are hired, they will likely persist and that recruitment is a more pressing issue than retention in achieving gender parity” (p. 866). Within academic science, the key place to increase the representation of women appears to be at recruitment, rather than at retention.

Furthermore, women who persist in academic STEM careers enact these roles differently with regard to leadership and entrepreneurship. According to one study (Niemeier & González, 2004), women chaired departments in only 2.7% of engineering departments, 5.9% of math or physical science departments, and 12.7% of life science departments. In contrast, women chaired in 23.4% of business departments and 31.5% of humanities departments. Even when accounting for cross-discipline discrepancies in the number of senior women faculty eligible to stand for chair, women were underrepresented as chairs of STEM departments. Furthermore, even though women and men earn advanced degrees at similar rates in the life sciences, women engage in fewer commercial practices related to science such as patenting (Ding, Murray, & Stuart, 2006) or disclosure of inventions (Thursby & Thursby, 2005), even though these analyses found the scientific work of men and women to be of similar quality. At a time when need for innovation is high, the loss of female innovators’ potential is of critical importance. Understanding the reasons why women—even those retained in STEM—do not engage with leadership or entrepreneurship to the same extent as men is thus critical for fostering economic and scientific growth and for individual professional development.

Consequences of women’s underrepresentation

These patterns of different participation of women than men in STEM fields have several consequences. First, society incurs economic and scientific losses because of the failure to recruit and engage fully the most qualified pool of potential STEM practitioners. The shortage of STEM workers in the United States has gained national political and public attention (e.g., U.S. Congress Joint Economic Committee, 2012). Lower numbers of women in STEM result in a narrower range of inquiry and progress in those fields; fields that have experienced increases in diversity also witness an increase in the range of topics pursued (e.g., greater representation of women in legislative bodies leads to more policy advocacy for topics such as education or health care; Swers, 2001, 2002, 2013).

A second consequence of women’s underrepresentation in STEM is the loss to women as a group because women are not accessing these highly valued occupations. Greater entry of women into STEM pathways could be a meaningful inroad to gender equality in occupations generally, because STEM occupations tend to be well paid and relatively protected from unemployment (U.S. Congress Joint Economic Committee, 2012). For example, economic analyses show that convergence in men’s and women’s college major choices in the 1980s led to reductions in the wage gap (Eide, 1994). As noted in economic analyses (Beede et al., 2011), women in STEM earn about 33% more than women in other occupations. Increasing the proportion of women in STEM occupations can thus narrow the gender wage gap (14% in STEM occupations, compared with 21% in non-STEM occupations). In particular, college-educated women earn a higher wage premium in STEM occupations than do college-educated men (20% wage premium for women vs. 11% for men).

A third outcome is that the lack of women in STEM fields can carry forward in constraining opportunities for future generations. The underrepresentation of women in fields such as computer science and engineering can self-perpetuate if few women enter because there are few women in these fields. Low numerical representation of women in STEM fields can create environments that cue social identity threat for women (Murphy, Steele, & Gross, 2007). For example, greater proportions of female faculty in economics departments were shown to foster greater numbers of female economics students (Hale & Regev, 2014). Women in STEM can attract future generations of girls and women to STEM, because the presence of similar models fosters entry into and persistence in careers (e.g., Riegle-Crumb & Moore, 2014; Stout, Dasgupta, Hunsinger, & McManus, 2011; Young, Rudman, Buettner, & McLean, 2013). More generally, broadening the pool of STEM talent would help address the shortage of qualified STEM educators (U.S. Congress Joint Economic Committee, 2012). In this way, more students may be introduced to and excited about STEM topics.

“Can they?” Explanations related to actual ability

Much previous research about gender-differentiated STEM pursuits has focused on ability—that is, do girls and women have the same mathematical and scientific talents as boys and men? The documentation of gender differences in underlying cognitive abilities is challenging, and differences vary in nuanced ways across methods, samples, ages, or culture (D. I. Miller & Halpern, 2014). We limit our review here to meta-analytic evidence, which provides effect size estimates across multiple studies.

Meta-analytic evidence, based on over 1 million U.S. students tested between 1990 and 2007, estimated the effect size in math performance to be d = .05, indicating a slight female advantage (Lindberg, Hyde, Petersen, & Linn, 2010). Similarly, a meta-analysis of gender differences in teacher-assigned school grades based on 369 samples (Voyer & Voyer, 2014) found a similarly slight female advantage for math courses (d = .07) and a female advantage for science courses (d = .15). The small female advantage in grades is apparent at all levels, except for elementary school math (d = −.01) and college science (d = .01; data not available for graduate-level science). Although Hedges and Nowell (1995) found boys to be especially overrepresented among the most talented students (i.e., in the right tail of the performance distribution), more recent analyses find nearly equal variability in male versus female math performance scores (variance ratio = 1.08; Lindberg et al., 2010). This convergence of the previous gender gap in mathematics performance emerges even on complex mathematical problems (Hyde, Lindberg, Linn, Ellis, & Williams, 2008) and among the most talented students (Hyde et al., 2008; Hyde & Mertz, 2009). Data collected from 1990 to 2009 show a fair amount of gender similarity in advanced mathematics and science courses in high school (National Science Board, 2012). High school girls and boys take advanced mathematics courses at generally equivalent rates (e.g., 73.5% for boys and 77.7% of girls for Algebra II; 29.4% of boys and 30.3% of girls for trigonometry and statistics/probability). Furthermore, high school girls are more likely than high school boys to take advanced classes in biology (49.9% of girls and 39.4% of boys) and chemistry (72.4% of girls and 66.7% of boys), and nearly as likely as boys to take advanced classes in physics (41.5% of boys and 35.9% of girls) or engineering (5.6% of boys and 1.1% girls). Boys and girls thus appear to be performing at generally similar levels of ability in mathematics and science.

Ability-focused explanations also fail to explain why women would enter some math-intensive fields but not others. For example, women earn a substantial proportion of higher degrees in mathematics: In 2008-2009, 43.26% of bachelor’s degrees in mathematics and statistics were awarded to women, as well as 41.20% of master’s degrees and 31.01% of doctoral degrees (Snyder & Dillow, 2011; Table 323). These numbers are certainly well beyond what might be expected if women were deficient in higher mathematical ability. Moreover, analyses of nationally representative samples, in three cohorts ranging from 1982 through 2004, found that gender differences in high school achievement did not account for the gender differences in declaring a college major in physical sciences or engineering (Riegle-Crumb, King, Grodsky, & Muller, 2012), even across multiple aspects of high school achievement (i.e., grade-point average [GPA] or standardized tests) and when examining either the gender differences in average achievement or gender differences at the upper levels of the distribution. Overall, the patterns of convergence over time, along with girls’ excellent performance in mathematics courses, have led various reviewers to conclude that men and women show similar aptitude for mathematics (Halpern et al., 2007; Spelke, 2005).

“Do they want to?” Motivational explanations

Given the accumulating evidence that girls and women show similar aptitude for mathematics and science as do boys and men, it is time for a critical shift in the focus of explanatory theory: Rather than asking whether women and girls can perform in STEM fields, we need to ask whether girls and women want to perform in STEM fields—and if not, why not? Despite gender similarities in mathematics performance, gender differences in mathematics attitudes and affect persist, with boys expressing more positivity toward mathematics (Else-Quest, Hyde, & Linn, 2010; Hyde, Fennema, Ryan, Frost, & Hopp, 1990). Ceci, Williams, and Barnett (2009) conclude their comprehensive review by positing gender differences in interests and lifestyle preferences as promising explanations for the gender gap in STEM pursuits. Understanding such gender differences in motivation, even in the context of gender similarities in mathematical and scientific aptitude and performance, is precisely the focus of the communal goal congruity perspective.

Moreover, the time is ripe to consider explanations that focus on ability or perceived ability along with motivational explanations because these beliefs and motives are inextricably linked. The breadth of abilities among girls and women contributes to their STEM decisions (Valla & Ceci, 2014). Young women are more likely to score highly on both the quantitative and the verbal sections of the SAT than are young men, and individuals with this profile are less likely to select into math-intensive fields (Wang, Eccles, & Kenny, 2013). Women’s decisions to enter into non-STEM educational pathways thus reflect that they have more choice, rather than less ability. We consider two clusters of explanations—gender differences in encountered stereotyping and prejudice, and gender differences in self-efficacy and identification—before elaborating the case for examining communal goal processes.

Women divert from STEM pathways because of gender stereotypes and prejudice

The evidence for continued gender stereotypes related to STEM is abundant. For example, stereotypes that pair men with math have been demonstrated at implicit levels; individuals may be unaware that they hold these associations, or that these associations can influence their behavior (Nosek, Banaji, & Greenwald, 2002). Implicit stereotypes that pair boys with math have been shown in children as young as 7 years old (Cvencek, Meltzoff, & Greenwald, 2011; Cvencek, Meltzoff, & Kapur, 2014). Moreover, both male and female science faculty have shown gender bias in preferring male over female applicants for a lab manager position, even when qualifications were matched experimentally (Moss-Racusin, Dovidio, Brescoll, Graham, & Handelsman, 2012; but see Williams & Ceci, 2015, for evidence of reverse preference by faculty among highly qualified applicants for a hypothetical faculty position). Scientific abstracts that were ostensibly authored by a male versus female scientist were rated as higher in quality (Knobloch-Westerwick, Glynn, & Huge, 2013). Implicit stereotypes pairing men with math/science predicted how much individuals in an “employer” role overestimated men’s performance on a math task and underestimated women’s performance (Reuben, Sapienza, & Zingales, 2014). For all of these reasons, beliefs about ability might shape motivational differences to enter into and persist in STEM fields, even given increasing female achievements in mathematics and science.

Particularly important is that girls’ and women’s intrinsic motivation may be undermined by widespread gender stereotypes about math and science. For example, stereotypic beliefs can function as self-fulfilling prophecies (Geis, 1993), leading girls to gain less exposure to math and science. Although the gender gap in formal math exposure (e.g., number of courses) has narrowed (Hyde et al., 1990), boys tend to receive greater exposure to math and science through informal means, such as hobbies, games, and sports. Differential exposure to video games, for example, has been tied to gender differences in spatial cognition, which can be reduced with specific training (Terlecki & Newcombe, 2005). Differential exposure to technical material can facilitate greater interest in technical skills among boys than girls, thus leading to boys’ advantage on measures of technical aptitude; exposure can drive interest and knowledge gain within a specific domain, given gender similarities in general mental ability (Schmidt, 2011). Moreover, cultural stereotypes are reproduced, sometimes without awareness or knowledge, by parents and teachers. An observational study of boys’ and girls’ interactions with science exhibits at a children’s museum found that even though boys and girls approached the exhibit at equal rates and asked equal numbers of questions, parents spent more time explaining scientific principles to boys than to girls (Crowley, Callanan, Tenenbaum, & Allen, 2001). Despite the growing evidence about similarity in mathematical and scientific abilities of boys and girls, strong gender stereotypes still influence socializers’ attitudes and behaviors. Differences in exposure and encouragement can foster the continuing gender differences that are observed in attitudes and interest in mathematics and science, even as achievement and course performance move toward gender similarity.

Women divert from STEM pathways because of lower self-efficacy or identification

Given robust cultural stereotypes that pair men with math, women and girls may perceive themselves as less skilled in math and science than do men and boys. One’s perceived self-efficacy in a given academic domain clearly predicts positivity toward the academic domain and persistence in the domain and related career fields (Bussey & Bandura, 1999; Eccles, 1994; Lent, Brown, & Hackett, 1994). Expectancies about success are related to self-evaluations of skill: For example, Beyer (1990) found that women held lower expectancies than men of their performance on male-stereotypic tasks (i.e., politics and sports-related tasks), which were in turn related to their self-evaluations of their performance. In contrast, men did not hold lower expectancies than women of their performance on female-stereotypic tasks (i.e., fashion and celebrities). In a task where candidates projected their own performance on simple math tasks (where men and women perform similarly), men were more likely to overestimate and women were more likely to underestimate their performance (Reuben et al., 2014). Beliefs that a field requires “genius”—innate talent rather than hard work—are more prevalent in male-dominated fields, such as physics (Leslie, Cimpian, Meyer, & Freeland, 2015), and dampen women’s motivations to pursue such fields (Smith, Lewis, Hawthorne, & Hodges, 2013). Women’s career choices are thus in part driven by their self-assessments of being lower in math ability than men (Correll, 2001, 2004). Indeed, because individuals gauge their talents in part by comparing across performance in different dimensions (Möller & Marsh, 2013), excellent performance in one domain (e.g., verbal) might diminish the impact of very good performance in another domain (e.g., math).

In addition, even individuals who view themselves as capable in math and science may not identify with math and science. The threat of confirming cultural stereotypes about women’s incompetence at math has far-reaching negative consequences, including the sense that one’s social identity is devalued in a particular field (Steele, Spencer, & Aronson, 2002). Stereotype threat—the fear of confirming negative cultural stereotypes about one’s group—has been shown to decrease working memory (Beilock, Rydell, & McConnell, 2007), learning (Rydell, Shiffrin, Boucher, Loo, & Rydell, 2010), motivation (Davies, Spencer, Quinn, & Gerhardstein, 2002; Davies, Spencer, & Steele, 2005), and interest in careers or tasks (Davies et al., 2002). These social identity threats lead to STEM domain disidentification through changes in achievement orientation, reduced sense of belonging, and reduced intrinsic motivation. Women with stronger implicit stereotypic associations between science and men identify less with science and hold less positive attitudes toward science, whether measured implicitly or explicitly (Young et al., 2013). Women who highly identify with their gender and who implicitly associate men with math are less likely to associate math with the self, even when they have selected into math-intensive majors (Nosek et al., 2002). Moreover, high gender identification paired with high implicit stereotyping predicted weaker performance on a mathematics final exam, as well as less intent to pursue a math-related career (Kiefer & Sekaquaptewa, 2007). Overall, continued encounters with beliefs that women are less talented or committed to STEM can create conditions under which even highly qualified women might be motivated to pursue other career paths.

Communal motivation

Communal, other-oriented goals tend to be important to people. These goals are part of a broad cluster of prosocial attributes (e.g., morality) that can facilitate smooth social functioning (Diekman & Clark, 2015). The fulfillment of communal goals such as helping others and being with others can serve important needs, such as belonging (Baumeister & Leary, 1995; Fiske, 2004), relatedness (Deci & Ryan, 2000), or affiliation (Hill, 1987). Communal attributes lead to positive consequences for the self and others. One example is that individuals who value communion reported more frequently receiving and providing social support, and their endorsement of communal goals was related to positive outcomes, such as feeling more peaceful, prosocial, engaged, empathetic, cooperative, and connected to others (Crocker & Canevello, 2008). Moreover, volunteering appears to buffer the link between stress and mortality (Poulin, Brown, Dillard, & Smith, 2013), as well as stress and psychological distress (Poulin, 2014). Communal opportunities offer particular benefits within occupational roles: Collaboration with colleagues predicts scientific productivity (Landry, Traore, & Godin, 1996; Lee & Bozeman, 2005), and workers whose jobs allow them closer contact with the beneficiaries of their help report greater motivation and exhibit greater persistence (Grant, 2007; Grant et al., 2007). Thus, fulfilling communal goals benefits the community, specific others, and the self.

Consistent with the centrality of communion to social life, both communal traits and behavior are evaluated extremely positive (Abele, Uchronski, Suitner, & Wojciszke, 2008; Eagly & Mladinic, 1989; Rudman & Goodwin, 2004), and social norms include expectations that individuals will enact communal traits such as being friendly, polite, and helpful (e.g., Schwartz, 1977). For example, a child’s communal tendencies as rated by peers and teachers (e.g., “This child is helpful to peers.” “This child is kind to peers.”) positively predict subsequent social acceptance, even when controlling for aggressive tendencies (Crick, 1997). The tendency to consider the needs of others, balanced against the needs of the self, has been shown to be advantageous in career performance across a wide range of occupations and contexts (Grant, 2013). Communal information receives priority in information processing: Individuals recognize communal words faster than agentic words, and are more likely to spontaneously mention communal than agentic information when describing other people (Abele & Bruckmüller, 2011).

Although many predictions of the goal congruity model hold for communal goals generally, greater precision can be obtained through specifying the type of communal goal that is important to a particular person or in a particular situation. Communal goals can involve helping others—that is, providing emotional or material support—or by working with others—that is, acting in concert with others, either physically or psychologically (Diekman & Steinberg, 2013). In addition, collaboration and altruism may involve connecting with distal others, such as society or a broader community, or proximal others, such as a mentor or lab mates. Given that the initial step must be to begin investigating communal goals in the context of STEM at all, we focus on the broader umbrella construct of communal goals and affordances here, and return to the question of particular communal opportunities in discussing research directions in Part III.

Affordance beliefs: STEM careers are believed not to afford communal goals

The belief that STEM careers impede the pursuit of communal goals means that the gender gap in STEM careers may remain despite the erosion of other barriers, unless communal goal incongruity is addressed. In particular, understanding the sources, mechanisms, and consequences of goal affordance stereotypes can offer a route to increasing participation in STEM. As shown in Figure 3, STEM careers are believed to be less likely to fulfill communal goals than female-stereotypic careers or than other male-stereotypic but non-STEM careers. However, individuals who do see the communal opportunities within STEM tend to have more positive attitudes toward and engage more in STEM (Thoman, Brown, Mason, Harmsen, & Smith, 2015; Weisgram & Bigler, 2006). Moreover, individual differences in believing that STEM affords communal goals more strongly predict positivity toward and motivation to pursue STEM than individual differences in endorsing communal goals (Brown, Thoman, Smith, & Diekman, 2015).

The belief that STEM careers impede communal goals appears to be robust and consensual. These beliefs are endorsed by both women and men, and by STEM majors and nonmajors (Diekman et al., 2011). Likewise, perceptions of the goals afforded by physical/mathematical sciences, compared with education/social services or medicine, did not differ by gender (Morgan et al., 2001). Such beliefs might emerge in childhood, as demonstrated with other career value affordances (Weisgram, Bigler, & Liben, 2010). Moreover, in addition to explicit stereotypes about goal affordances (see Figure 3), implicit measures reflect a dissociation between STEM and communal goals. Using Implicit Association Tests (IATs; Greenwald, McGhee, & Schwartz, 1998), we found that participants less easily associated science (vs. medicine) with communal constructs such as warmth (vs. power) and together (vs. alone; Diekman et al., 2011, Experiment 1c). These implicit associations can exert influence beyond the awareness or control of individuals; for example, a female engineering major may retain nonconscious beliefs that may associate engineering with individualized work, and these nonconscious associations might direct her tendencies to approach or avoid specific engineering activities.

The dissociation between STEM and communal goals can result from the lack of direct connection between STEM practitioners and the people who benefit from their work. In an unpublished investigation both proximal and distal communal affordances, science and engineering careers were perceived as less likely to fulfill proximal communal goals (such as direct helping) than distal communal goals (such as helping society; Steinberg, 2015). Similarly, a qualitative study by the National Academy of Engineering (Committee on the Public Understanding of Engineering Messages, 2008) found that K-12 students had trouble connecting engineers with those who help. As one teenager noted, engineers are “behind the scenes helping people” (p. 59), unlike doctors who visibly influence their beneficiaries. Cunningham, Lachapelle, and Lindgren-Streicher (2005) found that among elementary school students, the majority thought that engineers repair cars, install wiring, and drive machines, whereas a minority of students reported that engineers designed things and worked as a team. Even adults may not be aware of the range of functions that engineers perform. The report from the Committee on the Public Understanding of Engineering Messages (2008) found that whereas the majority of respondents agreed that engineers create economic growth (69%) and preserve national security (59%), fewer agreed that engineers save lives (14%) or are sensitive to societal concerns (28%). In another study (Klotz et al., 2014), both engineering and nonengineering students were fairly unlikely to see engineering as related to disease, food availability, or water supply, even though these are key issues that engineers could contribute to in terms of sustainability. To the extent that students believed that engineering improved quality of life or saved lives, they were more likely to pursue engineering.

Another reflection of the noncommunal content of stereotypes about STEM comes from the finding that drawings of scientists tend to show them alone: Indeed, one study of teachers found that scientists were drawn as solitary, whereas social scientists were drawn with other people (MacDuffie, 2001). Moreover, even those in STEM majors see their fields as less connected to personal values than do students entering other fields. In a survey of college seniors (Higher Education Research Institute, 2010b), those pursuing careers in engineering were unlikely (14%) to report that it was “essential” for their career to reflect their personal values, compared with students pursuing careers in teaching (28%), law (27%), or medicine (24%).

Role-relevant experience

Like other stereotypes, beliefs about the goals afforded by careers emerge from multiple sources, including personal experience, observations of others, and media representations; these previous role-relevant experiences form the basis of affordance expectations (see Figure 2). Personal experience includes an individual’s own direct experience and history with STEM in both formal and informal settings. Unpublished data from our laboratory suggest that these communal experiences are important. Individuals who report greater experience of communal activities in their science and mathematics education (e.g., mentoring, being mentored, working in groups) are more likely to believe that STEM affords communal goals, and in turn express greater interest in pursuing a STEM career, even controlling for quantity of experience in mathematics and science (Steinberg & Diekman, 2016). These communally oriented science and math educational experiences are more frequently reported in China than in the United States and help explain cultural differences in communal goal affordance beliefs and STEM interest (Brown, Steinberg, Lu, & Diekman, 2016). Especially important is that the extent of communal experience in science and mathematics predicts beliefs about communal goal affordances, above and beyond mere quantity of experience (Brown et al., in preparation; Steinberg & Diekman, 2016). Thus, it may not be simply quantity of experience that shapes STEM interest but the particular communal quality of those experiences.

When individuals lack direct personal experience with STEM fields, observational learning becomes even more important. Through observing others’ experiences, including the consequences of their decisions, people gain important information without having to experience the same events directly (e.g., Bussey & Bandura, 1999). Drawing from the stereotype formation literature, one prediction is that at low levels of experience, stereotypic information about groups will be highly influenced by particular group exemplars; as experience increases, an abstract stereotype will form (Sherman, 1996). Thus, even brief exposure to actual STEM practitioners can change beliefs about what their careers involve. For example, Fermilab (2010) documented pictures of scientists drawn by seventh graders before and after they spent the day visiting scientists in their labs. Afterward, the pictures reflected images that scientists were more approachable, friendly, and less “geeky.” Therefore, exposure to information emphasizing the communal nature of STEM careers may be especially effective in changing STEM attitudes among those with limited experience. For example, an intervention with middle school girls (Weisgram & Bigler, 2006) asked female scientists to incorporate information about how they helped others in their careers, and girls who internalized this altruistic message increased their interest in science. Furthermore, exposure to STEM practitioners who integrate communal goals in STEM can help foster these beliefs among students. For instance, the extent to which presenters at a 1-day intervention focused on communal aspects of their scientific work predicted students’ communal affordances, their sense of fit in science and math careers, and their science career interest (Belanger, Diekman, & Weisgram, 2016).

The STEM-related experiences of close others, such as parents or peers, can inform beliefs about STEM, and thus help dispel inaccurate stereotypes. Parents can also communicate the utility of the STEM fields to their children. When an intervention increased mothers’ perceptions of the usefulness of science and math, their children also perceived science and math as more useful and took a greater number of advanced math courses than other students (Harackiewicz, Rozek, Hulleman, & Hyde, 2012). Even brief experience with computer science majors who conform to or break the mold of the “geeky” computer science major can influence interest in computer science up to 2 weeks later (Cheryan, Drury, & Vichayapai, 2013; Cheryan, Siy, Vichayapai, Drury, & Kim, 2011). In part, these patterns occur because the experiences of close others can disambiguate what is involved in the field. For example, Harackiewicz and colleagues’ (2012) intervention described many ways in which STEM is used in daily life and careers, including socializing with friends and when helping people in a range of professions.

Another important cue that might be signaled by STEM practitioners is the extent to which a communal orientation is accepted and beneficial in the field. Following from this principle, the goal congruity model offers a distinctive reason for why role models might be particularly important in recruiting women to STEM fields. In addition to providing an example of success, work/family balance, and overcoming prejudice, female STEM role models might communicate that their field allows the fulfillment of communal goals (see Diekman, Weisgram, & Belanger, 2015, for a review). In unpublished data from our laboratory (Clark, Fuesting, & Diekman, 2016), fictional male and female scientists were equally likely to cue that science affords communal goals, and communally oriented individuals especially inferred that science affords communal goals when they learn about a female scientist who is prototypic of her gender.

Finally, beliefs about the goals integral to STEM are transmitted throughout society through media representation. Social learning occurs through observation of one’s direct environment, as well as through the symbolic environment of the mass media (Bandura, 2001). In a study of middle school–aged children taking the Draw a Scientist Test, commonly cited sources of drawings were TV/films (37%), their own imagination (10%), and someone they knew (8%; Steinke et al., 2007). In general, STEM is portrayed as incompatible with communal goals. For example, a content analysis of the portrayal of scientist characters in children’s television shows (Long et al., 2010) found that scientists were portrayed as caring in a slim minority of scenes (M = 2.0% of scenes for male scientists and M = 3.9% for female scientists). In contrast, scientists were portrayed as intelligent in most scenes (M = 58.0% of scenes for male scientists; M = 64.8% of scenes for female scientists). Other research shows that these portrayals can have an impact. An illustrative example is that experimentally manipulating the representation of computer scientists as stereotypic or counterstereotypic influenced readers’ subsequent interest in computer science (Cheryan, Plaut, Handron, & Hudson, 2013). Repeated exposure to counterstereotypic role models through media might disrupt stereotypes that deter women from STEM.

Anticipated (in)congruity

A critical question, given the strong stereotypes that STEM careers impede communal goals, is whether these stereotypic beliefs might be leveraged to increase interest—that is, if the communal opportunities within STEM are emphasized, will communally-oriented people show greater positivity to these careers? Even the mere expectation that goals will be fulfilled can influence persistence: For example, new volunteers’ expectations about the emotional consequences of their volunteering activities predicted their persistence in the volunteer role over time (Barraza, 2011). The expectations about what STEM roles will offer are thus extremely important.

Existing research suggests that communal factors influence women’s career choices broadly. Meta-analytic data show that women, more than men, tend to prefer jobs that allow them to help others (d = −.35), make friends (d = −.22), and work with people (d = −.36), whereas men, more than women, prefer jobs that afford solitude (d = .26; Konrad et al., 2000). Women and more feminine individuals show a preference to work with people rather than things, and this preference predicts vocational interests (Lippa, 1998). College women, relative to their male counterparts, reported that helping others is an important work value; for both men and women, involvement with others predicted interest across a range of careers (Morgan et al., 2001). Especially interesting is Eccles’s (2007) review of evidence that women were more interested in health-related careers because women accorded greater value to people- or society-oriented occupations than did men, even when these women predicted success for themselves in science. The critical variable thus is not simply whether an individual thinks she has the ability but also whether she thinks this path will fulfill her valued goals.

Goal affordance beliefs should function like other stereotypes in the extent to which they can be changed, and in the features that make them more or less likely to be changed. Like other robust stereotypes (e.g., gender, race), it is difficult to change associations that are based on repeated experience; nonetheless, there should be some flexibility in when these stereotypes are applied, both by individuals and to particular fields or local contexts (e.g., Blair, 2002). Research with novel occupations has shown that affordance beliefs can be flexible: For example, Weisgram et al. (2010) found that novel occupations portrayed as affording altruism elicited greater interest among girls than among boys (d = −.40 for children; d = −.40 for adolescents; d = −.19 for college students).

Both collaborative and altruistic opportunities influence beliefs and attitudes about STEM careers. For instance, we (Diekman et al., 2011, Experiment 3) experimentally manipulated the opportunities for collaborative communal goals within a description of a scientist’s daily tasks. Participants were randomly assigned to read about the typical day of an entry-level scientist whose daily activities were framed to be either collaborative or independent. The essential tasks were held constant across conditions but varied in the degree to which they involved communal activities. For example, the collaborative condition included, “I usually have to communicate closely with the Operations Group (they run the high-throughput screens) to check on the status of ongoing experiments so we can go from primary to secondary characterizations,” whereas the independent condition included the task, “I usually have to check a database maintained by the Operations Group (they run the high-throughput screens) to learn the status of ongoing experiments so I can go from primary to secondary characterizations.” As predicted, the collaborative framing particularly benefited women: Women reported more positive attitudes toward the science career when it was framed to be collaborative than independent, whereas men did not. In these data, emphasizing communal activities in science attracted women without dissuading men. It is also noteworthy that the same patterns held when communal goal orientation, rather than gender, predicted science positivity.

Further work has conceptually replicated the finding that elevating communal goal affordances fosters positivity toward STEM. For instance, when biomedical research was described as helping others, as compared with a control condition, participants expressed more positivity toward biomedical research and more motivation to pursue biomedical research in the future (Brown, Smith, Thoman, Allen, & Muragishi, 2015). Moreover, women who perceived STEM courses as affording communal goals were more likely to take STEM courses in college (Stout, Grunberg, & Ito, 2016). Increasing the congruity between women’s communal goals and the perceived opportunities to fulfill those goals can thus provide a new opportunity to increase interest in STEM careers and engagement in STEM-related activities.

Affective processes

Affective processes are closely intertwined with motivational frameworks; generally speaking, motivational drive states that are unmet lead to negative affect, and motivational drive states that are met lead to positive affect (e.g., Carver & Scheier, 1998). Thus, situations that facilitate goal incongruity are generally expected to foster negative affect, particularly anxiety, whereas situations that facilitate goal congruity are generally expected to foster positive affect. Highly communal individuals who recall being in an environment with low communal opportunities (compared with high communal opportunities) are more likely to report negative affect and less likely to report positive affect (McCarty, Monteith, & Kaiser, 2014). Furthermore, negative affect may lead individuals to nonconsciously disengage from goals (Aarts et al., 2007).

Furthermore, the biosocial model of affective decision making (Kitayama & Tompson, 2015) suggests that when behavioral conflicts exist (e.g., I am not enjoying this class that I need for my major), individuals search for positive incentives to direct behavior. Applied to the communal goal congruity model of STEM decisions, the prediction would be that when individuals experience conflicting affect within a particular role, the presence of communal goal pursuit opportunities will be particularly impactful. When the positive incentive of communal opportunity is present, individuals will seek those opportunities and produce cognitive narratives that support these behavioral tendencies.

Reconstruing the role

Individuals might achieve goal congruity by reappraising their major or field as affording communal goals. One way in which this might happen is to consider STEM fields from a more abstract perspective. Because STEM fields are seen as beneficial to society but unlikely to involve direct helping, highlighting these distal opportunities for communal goal fulfillment may be one way of increasing the perceived communal impact of STEM. In an experimental demonstration of this phenomenon (Steinberg, 2015), participants who adopted a high-level construal of a STEM career (i.e., focused on the “why” of STEM) rated the career as more likely to fulfill communal goals relative to those who adopted a low-level construal of the career (i.e., focused on the “how” of STEM). Similarly, connecting a task with its abstract communal benefits increased persistence and performance (Rodriguez et al., 2013; Yeager et al., 2014). Therefore, promoting an understanding of STEM careers within the broader societal context may be a way to provide role-relevant experience aligned with abstract communal goals. For example, a scientist might focus on the parts of her job that include contributing to the understanding of societal problems, or an engineering student might focus on the end result of robotics initiatives that will ultimately aid people in disadvantaged communities.

Another strategy that might be useful is to make plans to fulfill communal goals. The effects of such plan making can be achieved before the plans are acted upon, given evidence that making plans to fulfill goals reduces the negative effects of unfulfilled goals. For example, participants who wrote about plans to fulfill goals showed protection of executive function and self-regulatory ability, compared with participants in control conditions (Masicampo & Baumeister, 2011). Thus, the mere act of considering how to meet communal goals may relieve some of the negative effects of goal incongruity.

Reconstructing the role

When individuals in STEM fields find that their work life is not fulfilling communal goals, they may seek out opportunities to fulfill these goals both in and outside of work life. In college settings, students might pursue communal goals in curricular domains (e.g., classes), cocurricular domains (e.g., informal faculty advising), extracurricular domains (e.g., student clubs), and nonacademic domains (e.g., family or friend roles). However, these communal pursuits may take time and effort away from demanding intensive STEM coursework. In addition, cocurricular, extracurricular, or nonacademic pursuits may indicate to others or to the individual that he or she is not interested or does not “have what it takes” to succeed in STEM. Communally oriented behavior in particular may be perceived as irrelevant to or even impeding success in STEM. The ideal situation may be spending time on activities that integrate STEM work with communal goal pursuits, such as mentoring or study groups. Such integration of communal goal pursuits (e.g., spending time with or helping others) with agentic goal pursuits (e.g., achieving success in academics) can occur in different ways. For example, some individuals seek reassurance from friends during periods of academic doubt, others study with academic role models in order to bring their actual selves in line with their ideal academic selves, and others spend time socializing as mood repair (Cantor, 1994).

Indeed, communal individuals may create environments where they can fulfill communal goals: For instance, in new roommate relationships, communally oriented individuals both provided and received more social support than less communally oriented individuals (Crocker & Canevello, 2008). In particular, individuals with higher levels of proactive personality are more likely to shape environments to be consistent with their own values, and in turn, to meet with both subjective and objective success in their work roles (Seibert, Michael, & Kraimer, 1999). However, as we note below, individual ability to effect change is likely to be constrained by organizational norms, and thus individual-level role change will more likely succeed if supported by the broader organizational culture.

Exiting the role

The problem fundamentally motivating this research perspective is why women (more than men) opt out of the STEM pathway at some point, even when they have performed well in science or mathematics courses. When individuals do not see possibilities for fulfilling valued goals within their current role structure, they will modify that role structure. At the most extreme endpoint, this modification can result in leaving the STEM role for another role that is perceived to meet valued goals. As shown in Figure 2, the decision to exit one role leads back to a decision to enter another role. Indeed, career decisions are not made in isolation; individuals who opt out of the STEM pathway are always opting in to a different field, and so the relative benefits of each role need to be considered. Given the average breadth of women’s abilities (e.g., Wang et al., 2013), women in particular may have the option to pursue communal goals through a variety of career roles, and this breadth of choice may undermine commitment to STEM roles if communal goals are thwarted in those roles. Understanding how STEM roles are perceived to, and actually do, fulfill or fail to fulfill communal goals, is key to understanding these decisions in fuller context.

Navigating the opportunity structure of social roles

Although individuals pursue goals within many contexts, psychological theory and evidence about the pursuit of multiple goals across multiple roles are underdeveloped. We articulate two principles to form predictions: the primacy principle and the complementarity principle. In the primacy principle, individuals have a particular goal that takes precedence over others, and if this goal is not fulfilled, the individual will continue to seek to modify or substitute the specific social role until it is satisfied. Individuals who are extremely communally oriented, for example, might require communal activities in more than one role to meet their communal needs.

In the complementarity principle, an individual will prefer a new social role within a current decision set if it provides an opportunity to fulfill goals that are currently unfulfilled in the individual’s role system. Individuals will thus seek complementarity, rather than redundancy, in pursuing valued goals across the array of current or prospective roles. This principle leads to the prediction that even communally oriented individuals can meet a motivational threshold at which roles that meet other needs will be preferred. Further, specific types of communal goals can have distinct impact: For college students, highlighting the altruistic aspects of biomedical research fostered motivation to pursue such research, and highlighting the collaborative aspects of biomedical research did not further enhance motivation (Brown, Smith, et al., 2015). In general, the goal congruity framework predicts that individuals will opt for roles that afford opportunities to goal pursuit, particularly if those goals that are not afforded elsewhere.

These two principles offer a framework to understand how different types of communal goals might be important, but the question remains as to when certain goal opportunities will be particularly impactful. Considering the different goals that might be valued and the different roles that might provide goal pursuit opportunities leads us to delineate these five dimensions that will influence the attractiveness of goal pursuit opportunities:

Generalizing the goal congruity framework to other groups and roles

The goal congruity framework can inform the understanding of entry, engagement, and exit in many other educational or occupational pathways. Depending on the field, different goals may be central. For example, understanding worker shortages in teaching or nursing might benefit from examining whether these professions are increasingly thought to impede agentic goals of authority, independence, and status. Because agentic goals are valued by both men and women, these perceived agentic affordances might be especially important over time. Moreover, the basic logic of communal goal congruity—that pathways seen as impeding communal goals are less desirable, and pathways seen as facilitating communal goals are more desirable—has been demonstrated in business (McCarty et al., 2014) and politics (M. C. Schneider, Holman, Diekman, & McAndrew, 2015). This framework can also be invoked to explain factors that inhibit men’s pursuit of some occupations: Men’s lower endorsement of communal goals has been shown to underlie their lesser interest in health care and education occupational roles (Block, 2013). Applying the framework depicted in Figure 1 to different occupational domains will enable greater understanding of how goal congruity processes operate similarly and differently across fields. In the political realm, women’s communal emphasis can foster avoidance of conflict; when political roles are framed as serving the community rather than engaging in conflict, women express greater interest (M. C. Schneider et al., 2015).

The power of focusing on the psychological mechanism that underlies the gender gap in STEM pursuit—that is, the perceived communal opportunities available in a field—is that this framework can then be applied to other individuals and groups. It is not simply belonging to a gender group that influences STEM interest but the psychological tendency to value communal goals. Interventions that highlight communal opportunities in STEM can thus attract communally oriented people generally, including members of both represented and underrepresented groups. In particular, communal goals may be especially highly endorsed by specific social groups that are currently underrepresented in STEM (e.g., Native Americans, Latinos, first-generation students). For example, Native American students endorse communal work goals more highly than individualistic work goals; moreover, communally oriented Native American students in STEM fields were especially unlikely to feel a sense of academic belonging, and in turn less likely to report intrinsic motivation, intention to persist, or high levels of performance (Smith, Cech, Metz, Huntoon, & Moyer, 2014). Furthermore, Native American and Latino undergraduate research assistants who initially believed biomedical research afforded altruistic goals reported more psychological involvement in their research 10 to 12 weeks later, and this greater involvement was associated with enhanced interest in pursuing a science research career (Thoman et al., 2015). First-generation students are particularly likely to endorse other-oriented interdependent values (e.g., “Give back to my community”) compared with continuing-generation students (Harackiewicz et al., 2014), and these values are posited to misalign with the independent values of elite academic institutions, leading to lesser belonging among first-generation students (Stephens et al., 2012). The communal goal congruity model suggests that communally oriented activities in academic contexts can help these students to find greater fit within institutions. For example, when first-generation students believed that working in science allowed for helping others and giving back to the community, they expressed more intrinsic science interest (Allen, Muragishi, Smith, Thoman, & Brown, 2015).

A critical question, however, is whether individuals who are marginalized in STEM have the latitude to engage in communal pursuits, even if they personally find these pursuits valuable. When communal activities are perceived as irrelevant or even detrimental to STEM, those individuals who might most benefit from these activities may be least able to engage in them. For this reason, authority support for communal opportunities in STEM is particularly critical, as we detail in Policy Implications.

Given the robust prediction of positivity toward STEM from greater communal affordances, an essential question is where these affordance beliefs come from, and how they might be changed. Further examination of these processes would allow for more effective interventions. For instance, socialization agents (such as textbooks, classroom structure, teachers, and parents) likely reflect stereotypes of STEM as lacking communal opportunities, and greater to exposure to these beliefs among socialization agents will predict stronger stereotypic beliefs held by the student. Furthermore, the representation of communal exemplars within a student’s own social network is likely to influence the development of communal affordance stereotypes. Similar to processes of stereotype malleability (Dasgupta & Asgari, 2004) or stereotype inoculation (Stout et al., 2011), students who are exposed to more communal exemplars in STEM may be less likely to endorse the belief that STEM fields are noncommunal or to connect these beliefs to the self.

Varied sources of fit

The goal congruity model highlights a different way to achieve fit within academic and organizational contexts. Within social psychology, this sense of fit has primarily been investigated through research on social belonging; students benefit when they feel a sense of social connection to others in the academic environment (e.g., Walton & Cohen, 2011). The goal congruity framework highlights that fit can also emerge when an individual feels that his or her occupational role will serve valued goals. Other sources of fit might have to do with cognitive style (e.g., regulatory fit; Higgins, 2006), or generally with whether one’s attributes are congruent with role requirements (e.g., role congruity; Diekman & Eagly, 2008; Eagly & Karau, 2002). These various sources of fit might operate independently or be related: An individual might feel a sense of fit stemming from goal congruity, whether or not he or she feels a sense of social belonging in a particular setting. An additional benefit of considering goal congruity as a cue to fit is that it does not rely on demographic cues—and thus a wide array of role models from different backgrounds could potentially signal belonging to members of underrepresented groups. A broader conceptualization of psychological fit that draws from multiple sources could provide more elaborated understanding of why people choose and persist in their social roles.

The sense of fit due to goal congruity might counteract other belonging threats, such as social exclusion. We predict that a sense of fit from goal congruity would buffer against threats to belonging in STEM and promote recovery following such threats. Conversely, the lack of fit resulting from goal incongruity may contribute the tenuous sense of belonging experienced by women and other underrepresented groups (e.g., Good et al., 2012; Harackiewicz et al., 2014). Thus, increasing opportunities to meet both proximal and distal communal goals may have important implications for the social experience of people in the STEM fields.

Summary

The goal congruity perspective provides a generative framework to ask and answer questions about how individuals navigate their role system to meet multiple valued goals. Although this review focused on the impact of communal opportunities within STEM, the broader principles and predictions of the framework provide the chance to understand the role actions of a wide range of individuals and groups.

Institutional examples

Some STEM programs have adopted strategies that integrate communal opportunities in STEM. Although these initiatives range in content, strategy, and scope, they provide examples of the benefits of (a) increasing societal relevance of STEM, (b) facilitating collaborative opportunities in STEM, and (c) deemphasizing or disrupting stereotypic expectations about preparation or talent in STEM.

Numerous programs highlight the ways in which STEM fields are relevant to benefiting society. For example, an engineering program for middle school girls focused on hands-on activities along with lessons about the altruistic impact of engineering projects (Colvin, Lyden, & León de la Barra, 2013). Girls reported greater perceptions that engineers fulfilled communal goals, and they increased enrollment in subsequent engineering activities. Farrell (2002) recounted a strategy of highlighting social relevance of engineering in college-level courses to bring more women into the field. Several schools have changed traditional curricula, so that students have more hands-on or collaborative experience in the first years of college, instead of waiting for higher level courses. One indicator of the appeal of engineering projects that directly help others is that the National Engineering Design Challenge attracted more than 40% female participants when the challenge was to design a device to help people with disabilities.

Institutional support for collaborative communities also reaps benefits. Cuny and Aspray (2001) detailed strategies that a panel of experts in computer science and engineering developed to recruit and retain female graduate students. They emphasize creating a peer community for female students (either at the home institution or across institutions) to provide opportunities for interaction with fellow students. The panel acknowledged that certain stereotypes of computer science and engineering might particularly deter women, including beliefs that working with computers is incompatible with working with people, or that the culture of the field will conflict with personal values. Thus, they recommend ensuring that departmental initiatives such as departmental publications or visiting speakers do not reinforce these ideas. More broadly speaking, STEM educational initiatives that focus on active learning (which often includes participation in class and group work) can reduce gender gaps in STEM achievement (e.g., Haak, HilleRisLambers, Pitre, & Freeman, 2011).

There are also clear examples of success in recruiting and retaining female STEM majors from initiatives that combine multiple strategies. For example, Harvey Mudd College (described in AAUW, 2015) successfully increased their female computer science graduates from 12% of graduates in 2000 to 38% in 2014. Their comprehensive program included various opportunities for students to consider the practical and societal applications of their research, to engage collaboratively in research, to develop relationships with faculty, and to attend conferences focused on women in computing together. The Electronics and Computing Service Scholars program at Miami University strategically integrated communal opportunities: Students are selected on the basis of their service vision, first-year coursework includes service learning, and ample opportunities for social connection and collaboration are available. Even the first year of the program showed gains in gender and ethnic diversity of those who applied and were admitted, relative to local and industry benchmarks (Brinkman & Diekman, 2016).

Yowell and Sullivan (2011) described an implementation of the rebranding effort developed by the National Academy of Engineering to improve the public understanding of engineering. At the University of Colorado—Boulder, recruitment flyers and postcards for engineering events were redesigned to show people interacting with technology rather than technology alone, and to include taglines highlighting the social impact of engineers. In response, they found that attendance at a program to introduce high school girls to engineering quadrupled, and the enrollment of women engineering students, which was previously at a 5-year average of 19%, increased to 24% in the year following the redesign.

Finally, recognition of the ways in which stereotypic expectancies might have channeled girls out of STEM pathways early suggests that a route to inclusivity is to deemphasize past experience and emphasize the potential of all students to learn relevant skills. For example, the initiative that successfully increased female computer science graduates at Harvey Mudd College included introductory courses specifically aimed at incoming students with little prior programming experience (AAUW, 2015). In addition, Cuny and Aspray (2001) suggested broadening admissions criteria to create space for people who have the ability to succeed in computer science and engineering but lack direct experience with these classes. Admission policies that emphasize past experience with computer science and engineering might disproportionately disadvantage women who became interested as a means of addressing societal problems but did not focus on the field from an early age (also see Holloway, Reed, Imbrie, & Reid, 2014).

Despite the goal congruity perspective’s potential to promote fuller understanding of women’s STEM engagement, the possibility exists that the focus on communal affordances in STEM will deflect attention from other key processes, such as prejudice against women in STEM. Much like attributions for responsibility, where explanation may be equated with exoneration (A. G. Miller, 2004), exploring relatively internal factors may be perceived as denying the contributions of relatively external factors. One implication thus may be that focusing on internal factors is seen to do a disservice to working against structural barriers that inhibit gender equality in STEM. We find this conclusion unsatisfying for a number of reasons. First, as is clear from this review, we contend that giving women’s varied motivations a primary place in our theoretical frameworks is itself necessary for progress toward gender equality. Second, truly disentangling what is “internal” and what is “external” is not easily done, given that social psychology has long documented how relatively external beliefs or roles can become internalized. As noted by Williams (1999), women’s occupational decisions include both elements of choice and elements of discrimination: “Choice concerns the everyday process of making decisions within constraints” (p. 37). Third, understanding the roots of a large-scale issue like gender gaps in STEM pursuits requires attention to multiple factors; there is not just one solution to this challenge. The elaboration of motivational processes thus serves as a call for more, rather than less, research on external factors that demotivate underrepresented groups.