Women in engineering: a review of the 2014 literature

SWE’s assessment of the most significant research found in the past year’s social science literature on women engineers and women in STEM disciplines.

By Peter Meiksins, Kacey Beddoes, Peggy Layne, Maureen McCusker, Elsa Camargo November 30, 2015

Interest in the underrepresentation of women in engineering and other STEM disciplines continued unabated in 2014. In addition to a large amount of scholarly and professional interest in the issue, broader public attention was drawn to it by discussion of America’s competitive status in STEM fields—questioning whether the U.S. is falling behind, and if recruiting more women is the solution—and by an op-ed in The New York Times in which researchers Wendy Williams, Ph.D., and Stephen Ceci, Ph.D., made the controversial claim that sexism and discrimination were not significant factors explaining the low numbers of women in academic science. Our review of the literature covered well over 100 publications, including books, major reports, and journal articles in publications representing a half dozen or more disciplines. We searched for articles by examining major research databases and more than 70 journals that publish articles on gender and engineering. As always, the studies varied tremendously in quality and rigor; they also varied in their methodological approach, from complex statistical analyses of large data sets to interpretive studies of qualitative data.

The literature we reviewed continues to explore familiar explanations of why there are relatively few women in engineering. Some studies focus on early childhood socialization and children’s experiences in the K-12 educational system, arguing that STEM in general and engineering in particular are perceived by children as male fields, so girls either are not attracted to them or are actively discouraged from entering them. Others focus on what happens to young women who display an aptitude for math, science, and engineering when they enter university. The focus here is on whether engineering programs are supportive of female students, whether engineering curricula respond to the kinds of issues in which female students are likely to be interested, as well as on understanding why young women with strong math and science skills opt for majors other than engineering. Finally, the literature continues to recognize that engineering graduates do not always thrive or remain in the field after they enter the workplace. Studies of both academic and nonacademic settings in which female engineers and scientists work continue to investigate potential obstacles to female engineers’ career progress and factors that may explain why some female engineers "opt out."

It has become fairly common to argue that the explanation for the low numbers of women in engineering lies in all of these areas and that there is no single cause of gender imbalance in technical fields. However, an interesting aspect of this year’s literature is the increasingly lively debate over the question of whether gender bias and discrimination characterize contemporary engineering workplaces. Are these factors keeping women out and/or holding them back when they gain admission, or, as some observers now hold, are women simply choosing not to enter the field, despite real progress in eliminating bias and discrimination? If gender bias and discrimination have been reduced or even eliminated, should the focus of analysis shift to the choices women make earlier in life about whether or not to enter technical fields in the first place? And, if the key is to understand women’s choices, are those choices simply a matter of individual preferences or are they constrained by gendered realities that make certain choices more likely than others? Given the prominence afforded these questions, particularly by Williams and Ceci’s New York Times op-ed piece, we have paid particular attention, in this year’s literature review, to contributions that bear on the issues of bias and the nature of women’s choices.

A historical overview

One of the most substantial publications we reviewed this year provides us with the opportunity to look at female underrepresentation in engineering in historical context. Amy Sue Bix’s (2014) Girls Coming to Tech! A History of American Engineering Education for Women surveys the experiences of female engineering students at elite institutions from the late 1800s to the 21st century. Based on extensive use of archival materials, including the SWE archives, Girls Coming to Tech tells the story of how women gained entry to engineering, with focused case studies of Georgia Tech, Cal Tech, and MIT making up a substantial portion of the book. A full-length review appeared in an earlier issue of SWE Magazine (the 2014 Conference issue), but a quick summary here will serve to introduce some of the important questions about the contemporary situation.

Bix describes the powerful forces that have served to exclude women from engineering over the past 100-plus years. She demonstrates that women in engineering colleges often encountered conscious, deliberate efforts to keep them out as well as an unwelcoming or even hostile culture that made it difficult to feel as if they belonged and to get the help that they, like their male counterparts, needed. Various forms of harassment, from sexist jokes to stereotypical comments on women’s bodies to more serious forms of sexual misbehavior, also plagued female engineering students well into the 21st century. Bix is equally persuasive, however, in arguing that it was often unstated, implicit biases and unrecognized gendered processes and structures that blocked women’s entry or discouraged their progress. Thus, factors ranging from the absence of equal facilities for women (e.g., restrooms), to often unstated and unacknowledged perceptions of women’s inferiority, to linguistic practices ("girls" come to tech) that diminished women combined to make it difficult for them to thrive in engineering programs. She stresses the importance and consequences of conscious efforts to change attitudes and to break down the barriers that had traditionally kept women out of engineering.

In the end, the story Bix tells is one of incomplete progress. She documents the fact that the open questioning of the propriety of women’s entry into engineering has more or less died away as women have found a place in the field. SWE is singled out for praise as one of the organizations that has played a key role in effecting this change. At the same time, she notes the persistence of stereotypical beliefs about women’s abilities as engineers and scientists (she comments, for example, on Harvard President Lawrence Summers’ infamous comments in 2005 on women’s math abilities). In addition, she notes that observers have expressed concern that women’s progress may have plateaued; even if explicit biases may have weakened, many of the implicit beliefs and structural realities that kept women out in the past continue to be in place today. As the SWE reviewer put it:

Her rich and nuanced study reminds us of how far women have come, as well as how much work remains if American engineering educators hope to cultivate the potential of all who seek its rewards (Homsher 2014:61).

Bix’s historical review, then, leaves us with the question of why, despite concerted effort to increase the numbers of women in engineering, and some measurable progress in reducing conscious resistance to that effort, women remain a minority in the engineering profession.

Girls’ aptitude vs. interest vs. socialization?

One possible answer to the question of why girls are less likely to be interested in engineering lies in childhood. For whatever reason, girls are less likely than boys to express an interest in engineering, so are less likely to identify it as a career choice and to take the preparatory steps necessary to entering the field. Traditionally, this was attributed to girls’ lesser ability in the fields critical to engineering success: math and science. Educators and parents did not encourage girls to pursue engineering, and even girls themselves were not attracted to engineering because of the belief that girls lacked the skills needed to become an engineer.

Evidence has accumulated that this lesser ability is more myth than reality (for a brief summary, see Valla and Ceci 2014). But, the belief that girls lack math and science ability persists. Saucerman and Vasquez (2014) published a useful review of the literature on psychological barriers to women’s participation in STEM, emphasizing what they argue is girls’ and women’s lifelong exposure to overt and subtle messages that make them feel that their absence from STEM fields is the result of lack of ability. They note, for example, that most teachers with math anxiety are female, and that teachers transmit math anxiety to students, so girls continue to be its primary victim. Similarly, teachers continue to attribute the math success of boys (but not girls) to innate ability, while media continue to portray STEM professionals as men. They note the existence of research indicating that men actually have more egalitarian views of women’s ability in science and math than do women, perhaps reflecting the continued effects of the subtle messages to which women are exposed throughout the life course.

Scholarly attention also has turned to a different explanation of the relatively small numbers of girls expressing an interest in STEM careers. Maybe it is not a matter of perceived ability but rather that girls aren’t attracted to engineering because they lack knowledge about a field that is widely stereotyped as male and which is seen to be involved with activities in which girls typically are not interested.

Several of the studies we reviewed this year focused on this explanation for female underrepresentation in engineering and what can be done about it. Hammack and High (2014) report on a study of 68 sixth- and seventh-grade girls in the Southwest who participated in an after-school mentoring program about engineering. Prior to the program, girls viewed engineers as people who "fixed things and built stuff."

Participation in the program resulted in their viewing engineers as creative problem solvers who improve the world. Hirsch et al.’s (2014) study of 141 fourth- and fifth-graders also found that an enrichment program shifted students’ perception of engineers (this was true for both boys and girls in their study). Interestingly, they found that girls in a girls-only enrichment group were more likely to show an increased sense of self-efficacy and to depict engineers as female, something few children of either sex did prior to the enrichment program. Each of these studies points to children’s perceptions of engineering as male, and to the fact that that perception can be changed. A note of caution is introduced, however, by Robinson and Pérez-Quiñones (2014), whose study of 19 middle-school minority girls found that participation in a program on human-computer interaction changed the girls’ perception of the discipline, but had little effect on their interest in pursuing a career in computer science.

One common intervention designed to combat the gender typing of engineering and STEM disciplines overall is mentoring. Underlying this approach is the idea that girls are less likely to be attracted to engineering and related fields because they see relatively few females in these occupations and have little contact with women who could serve as models. Two studies we reviewed, however, raise questions about this approach. Draus et al. (2014) surveyed 695 women in informational technology, who reported that they did not see a great deal of usefulness in mentoring; 57 percent said that mentoring had little or no effect on their decision to go into IT. Bamberger (2014) reports on a study of Israeli ninth-grade girls whose school was visited by 12 highly educated female scientists and engineers from one of the country’s leading high-tech companies. Sixty girls participated in the visit, while a group of 30 did not. This intervention actually backfired, as the girls who participated in the visit had a more negative view of women in STEM and were less likely to express an interest in STEM careers afterward, while the control group showed no change. Bamberger speculates that the participants in the visit were frightened by the scientists and engineers, who used terms and concepts the girls found foreign and incomprehensible.

Valla and Ceci (2014), in a brief but provocative research note, point to another possible explanation for the fact that relatively few young women are attracted to careers in engineering and related fields. They raise questions about the ongoing focus on math ability in analyses of female underrepresentation in STEM, arguing that whether women have strong math ability tells only one part of what one needs to know to explain their occupational choice. The authors cite other research indicating that career choices are shaped by interests, not just aptitudes. Moreover, they note that women who score high on math ability also tend to score high on verbal ability, something that is less true for men. As a result, talented women have abilities in multiple areas, and may opt to enter fields in which verbal, rather than math, skills are central, whereas men who are good at math are more limited to what the authors describe as "narrow," math-centered activities. Valla and Ceci argue for a "breadth-based model" of women’s underrepresentation in STEM, suggesting, in effect, that women choose not to enter these fields, rather than being excluded from them. This complements their contention, to be discussed below, that women are not treated unequally in science but are choosing NOT to enter math-based STEM fields, largely because of their choice to devote time to family and child-rearing activities.

Valla and Ceci raise an interesting question about whether talented young women are "voting with their feet" in avoiding science. However, the authors take for granted the definition of certain career paths (including engineering) as "narrow." Their argument is that boys who are only good at math are drawn to engineering and computer science because these fields demand those skills exclusively. But, as the National Academy of Engineering’s Engineer of 2020 report and ABET’s 2000 reconfiguring of accreditation criteria for engineering programs advocate, engineering actually requires a variety of professional as well as technical skills.

Valian (2014) raises precisely this question in a commentary on previously published studies on the use of occupational interest inventories. Valian argues that "gender schemas" (stereotypes) are built into these scales and that sex differences in interests are changeable and sensitive to environmental cues (such as changes in which women are represented in fields formerly dominated by men). Women’s interest in math and science fields will increase if they have a feeling of belonging and an expectation of success, so that Valian concludes that "if we change the environment of math and science, we will change women’s interest in math and science" (229). Perhaps, then, if engineering and math became (or were perceived as) less "narrow," they might attract more of the more broadly talented women Valla and Ceci contend.

What happens in university?

Many young people make their decision about a college major well before they enter the post-secondary system. But, many of them also change their minds, often multiple times, while others complete high school without a clear sense as to what their intended major is. Thus, explanations for the underrepresentation of women in engineering need to look not just at what happens before college, but also what recruitment practices and messages are used to attract engineering students. Moreover, since at least some students leave engineering before completing their degrees, and because there has been considerable talk about the "leaky pipeline" for female engineering students, it is also important to consider how engineering programs endeavor to retain students, particularly female students.

These issues were central to a number of the publications we reviewed this year. Holloway et al. (2014) considered the question of whether admissions criteria are contributing to women’s underrepresentation in engineering. They analyzed admissions data from the 2006-2010 cohorts at a Midwestern public university for applicants to the college of engineering. They found that there was gender bias in the admissions process and that it was built into the criteria and policy in use. The university responded by changing its admissions practices; this involved both reorganizing the admissions office and changing the criteria. Some of the changes included reducing emphasis on math scores on standardized test and increasing the focus on other cognitive factors such as verbal or written test scores, and on noncognitive factors such as leadership and academic motivation. Holloway et al. report that these changes were effective in eliminating gender bias in the admissions process at this university.

Several studies we reviewed discussed the question of whether engineering is, or could be, presented in ways that respond to female students’ interests. Lehr, Finger, and Christine (2014) conducted online surveys with first-year engineering students at a polytechnic state university in California and report the relatively familiar finding that efforts to recruit female students to engineering may be more successful if they convey different messages about what it means to be an engineer (e.g., make a difference) and/or if they occur in different settings (e.g., English classes). Klotz et al. (2014) report on a large survey conducted as part of the larger "Sustainability and Gender in Engineering Survey" of college students in introductory English classes. They found that, in general, sustainability was a theme that could attract both male and female students to engineering. However, there were differences between male and female engineering students—men were more likely to be interested in energy, while women were more interested in addressing issues of disease and poverty. Findings such as this add support to the argument that engineering may need to present itself differently to women than to men.

Cech’s (2014) findings from a study of more than 300 engineering students at four New England schools, however, raise questions about whether an emphasis on issues such as sustainability, health, or poverty is actually at the heart of engineering students’ sense of self. She reports that engineering students display a culture of "disengagement" and did not put public welfare consideration at the forefront of their professional identity. Moreover, she finds that students’ commitment to public welfare concerns decreased over time, as she interviewed students as first-year students, then again as seniors. Thus, it may be that students lose their interest in these issues fairly quickly, so that basing engineering recruiting materials on their centrality may not make sense. Or, it may be that students who are interested in these issues will find it difficult to thrive in a student culture dominated by a very different set of concerns, resulting in high attrition rates and levels of dissatisfaction.

Brawner, Orr, and Ohland (2014) draw attention to another aspect of recruitment: When does it occur? While most engineering graduates begin their college careers in engineering, a minority does not. The authors’ analysis of data from the Multiple Institution Database for Investigating Engineering Longitudinal Development shows that students who shift into engineering from other math and science disciplines are more likely to be female (the difference is not huge, but still significant). Thus, it may be that paying more attention to recruiting students from other majors will prove to be an effective way of increasing the numbers of women in the discipline, an idea that would be consistent with the argument that there is no single pipeline into the profession. Terenzini et al. (2014) note that engineering students in community colleges are unusually likely to be minority students, first-generation college students, and non-native English speakers. Looking outside the traditional high school recruitment may be a way to increase the diversity of engineering programs in other ways as well.

Once women declare an interest in engineering or STEM disciplines, do they stay? If they stay, what contributes to their retention? Gayles and Ampaw analyzed secondary data on 1,488 female STEM majors at four-year institutions in the U.S. between 1996 and 2001. They found that regular interaction with faculty was an important predictor of completion, particularly for women. So was being a full-time student. Surprisingly, however, good social experiences were not predictive of completion (although women students noted negative aspects of their social experiences); and, GPA was a weaker predictor of completion for women than for men, implying that some successful female students leave STEM disciplines despite high grades.

Several studies (Raelin et al. 2014; Flores et al. 2014) point to feelings of self-efficacy as a particularly important predictor of degree completion for women in engineering and STEM more generally, which echoes research we have reported in previous years. Raelin et al. add that contextual support is very important in encouraging women to stay in school and complete STEM degrees. Miller et al.’s study of 286 engineering students at two predominantly white and two HBCUs notes that female students mentioned seeking help from other female students as an important strategy for coping with the challenges of their minority status.

Imran, Nasor, and Hayati (2014) describe an interesting case study showing how admissions criteria and curricular design can influence the retention of engineering students. They studied two groups of undergraduate electrical engineering students, one composed of students admitted before changes to admissions and curricular requirements, one composed of students admitted after those changes were in place. The changes included reducing the total number of hours required, curricular redesign to ensure a smoother transition from junior to senior years, and enhancements to the program designed to encourage interest and engagement. The authors report that the changes significantly increased retention for students in general, and for female students specifically; the percentage of women in the second study group was also larger than in the first group, perhaps because of changed admissions rules, perhaps because of more effective retention.

The workplace: a gender-biased academy?

The argument that academic departments are not welcoming to female scientists and engineers and that women encounter a hostile, chilly, even sexist climate when they enter academic science has been central to the literature on women in STEM disciplines. Part of the rationale for the existence of programs like the National Science Foundation’s ADVANCE is the perceived need to take positive steps to counteract the effects of both conscious and unconscious gender bias in the academy. Not everyone agrees that this is still a problem, however, as this year’s literature review made clear. In a major monograph-length article, Ceci et al. (2014) summarize their own and others’ research challenging the view that a major cause of the underrepresentation of women in academic science is gender discrimination in these disciplines. The monograph reviews the research underlying the op-ed they published this year in The New York Times (Williams and Ceci 2014). They contend that the most recent evidence shows that women who enter academic science disciplines fare at least as well as their male counterparts. In math-intensive fields where women are underrepresented, they find that female candidates are at least as likely to be invited to interview for tenure-track positions. Similarly, they find that manuscript and grant funding are now gender neutral, with female authors and principal investigators experiencing similar acceptance and funding rates as their male counterparts.

In their scholarly monograph, Ceci et al. argue that, instead of focusing on discrimination against women who enter academic science, researchers seeking to explain the underrepresentation of women in science should focus on precollege factors and the likelihood that women will major in math-intensive STEM disciplines. Their review of the current research does not support the view that early sex differences in spatial and mathematical reasoning are biological; indeed, they find evidence that sex differences in math ability, even at the high end (the right tail) from which most STEM academics are drawn, can and have changed over time and vary across cultures. They also note that gender differences in attitudes and expectations about math-intensive careers appear very early in children’s development and lead to a lower percentage of women choosing to major in math-intensive STEM disciplines. The clear implication of their argument is that interventions designed to increase female participation in academic science should target young children, not adults, and focus on influencing girls’ attitudes to careers and STEM majors that involve advanced mathematics.

Ceci and Williams’ arguments have been sharply criticized by others. Rebuttals, particularly "STEM Women, Sexism in Academic Science: Analysis of The New York Times Op-Ed" (2014), contend that Ceci and Williams make selective use of data, don’t acknowledge evidence of sexism in science, and focus on individual-level data, rather than the structural characteristics of academic departments and institutions. Particular criticism is leveled at their analysis of the evidence regarding the allegedly equal treatment received by female academics in publication and tenure and promotion decisions. Ceci and Williams acknowledge that female assistant professors are less "productive" scholars, but see this as evidence not of sexism, but of choices women make to prioritize family and child-rearing over academic work. From their point of view, if women publish less and are, therefore, less likely to be granted tenure and promotion, it is not because they are women but because they are less successful within the gender-blind competition that is scientific work.

Critics note, however, that this normalizes a highly gendered reality in which women (and not men) are expected to and expect to have primary responsibility for child-rearing and in which definitions of scientific success assume an "ideal worker" who can devote all or most of his (or her) time to scientific work. From the critics’ point of view, then, science is organized in a way that is gendered and that favors the experiences and realities of men over those of most women (unless those women are willing to "act like men"). Ceci and Williams are charged with ignoring this aspect of the social organization of gender in STEM.

Several studies we reviewed emphasized the gendered reality of academic engineering and science and challenged the view, propounded by Ceci and Williams, that academic science is now gender neutral. Beddoes and Pawley (2014), for example, interviewed 19 STEM faculty (four of whom were men) at a Midwestern public research university between 2009 and 2011. The participants in this study all note that there are significant conflicts between being a female faculty member and having children and a family. Within the university, views similar to Ceci and Williams’ prevail: These conflicts, while acknowledged, are generally regarded as a matter of individual choice, not of inherent, structural conflicts between prevailing models of family life and work. The respondents identified this as a primary reason for the underrepresentation of women in academic STEM departments and expressed the view that jobs in industry were actually more accommodating.

Cherkowski and Bosetti (2014) in a study of female faculty in all academic disciplines reported that academic women feel pressure to "do it all," to observe academic publication norms but also to be excellent mothers, a kind of pressure not experienced by their male counterparts. Finally, Cech and Blair-Loy (2014) surveyed 506 STEM faculty members at a research-intensive university regarding their perceptions of the consequences of taking advantage of the university’s "family-friendly" policies. They found that faculty believed that using the work/life policy would have negative consequences for their careers because of the perception that parents have lower worker commitment. Like Beddoes and Pawley’s respondents, these academics were more likely to want to leave for industry, where family-friendly accommodations were perceived as more widely accepted. Cech and Blair-Loy found no difference between mothers and fathers in their attitudes toward family-friendly policies. Because it is women, far more than men, however, who take the primary role in child care and other care work, the effects of these perceptions are experienced primarily by female faculty.

Other research we reviewed points to evidence of continued biased or discriminatory practices in academic science, questioning Ceci and Williams’ view that academic STEM departments have become gender neutral. For example, Milkman, Akinola, and Chugh (2014) report on an interesting experimental study examining whether bias persists in graduate education. They conducted an "audit study" involving more than 6,500 professors at top U.S. research universities; faculty were contacted via e-mail by prospective graduate students whose names signaled their gender and race/ethnicity (male/female, Caucasian/African-American/Hispanic/Chinese/Indian). The e-mails asked faculty to meet with prospective graduate students about possible research opportunities prior to applying to the doctoral programs in which they taught. The authors hypothesized that faculty would be less responsive to female and minority applicants, and that the pattern of discrimination would be stronger in fields in which women and minorities were underrepresented because faculty would be less accustomed to female and minority colleagues; and in which pay was higher because these fields tend to be dominated by white males.

The results largely confirmed Milkman, Akinola, and Chugh’s hypotheses, although there were interesting exceptions. Female and minority applicants did receive fewer responses, with Indian and Chinese students encountering particularly pronounced discrimination. For the most part, the pattern held across disciplinary areas, although there was no discrimination found in the health sciences and, in the fine arts, women and minorities were actually favored. Somewhat surprisingly, and contrary to what many other analyses would have predicted, the representation of women and minorities in the various disciplines was not a predictor of discrimination; the treatment of women and minorities was not better in disciplines with higher female and minority representation. There also was little evidence in this study that female students benefited from contacting female faculty, or that minority students benefited from contacting faculty in the same minority group, with the exception of the Chinese students. The study did find that discrimination was greater in disciplines in which faculty earn higher salaries. Faculty in public universities were less discriminatory than those in private universities; however, the ranking of the university was not a predictor of the level of faculty discrimination against female and minority candidates.

While this study is not exclusively focused on engineering or even STEM programs, it does provide evidence that engineering/STEM faculty continue to be affected by implicit biases that lead them to favor white male applicants to graduate school. As the authors of the study point out, the kind of discouragement a student receives from a nonresponse to the kind of request simulated in this study may lead them to not apply to a particular graduate program or to give up on their plans for graduate study altogether. This study revealed the importance of combating this kind of bias if the numbers of female faculty in STEM disciplines are to increase. Moreover, if the evidence found by this study proves to be correct, eliminating the effects of implicit bias is not simply a matter of increasing the numbers of female and minority faculty in academic departments, since the demographic composition of departments did not appear to affect the level of discrimination prospective students encountered. Instead, it would appear to be a matter of directly confronting the implicit biases of faculty of all kinds (male/female, majority/minority), a major component of many NSF-ADVANCE-funded projects over the years.

Joshi (2014) conducted online surveys with more than 400 participants in interdisciplinary research centers at a U.S. public university. The goal of the study was to see whether researchers judged men’s and women’s contributions differently. Results showed that recognition and utilization of the expertise of male and female scientists was contingent on the gender and gender identification of the actors assessing that expertise, the team’s gender composition, and the degree of female faculty representation in the discipline in which the teams were embedded. Thus, men who strongly identified as male tended to evaluate women’s contributions more negatively. Teams with more women were more likely to utilize the expertise of highly educated women. And in fields with more female representation, teams that had many women were found to be more productive.

Mason et al. (2014) found that women receive proportionally fewer awards than male professors. Based on historical institutional data from a large technical university, they found that between 1964 and 2012, 19.8 percent of university awards were given to women, and between 2007 and 2012, women received 12.1 percent of awards, but were approximately 30.5 percent of the faculty population. Challenges to creating less-biased award systems are identified, including unclear criteria, the use of student evaluations, and the use of letters of recommendation, all of which have been shown to perpetuate gender biases detrimental to women. The authors recommend methods for combating these clear indicators of remaining bias in academic science, including more explicit evaluation criteria, holding decision makers accountable, and legitimizing female leaders in order to establish a basis for positive ratings.

Finally, the argument that the workings of academic science have become increasingly fair, rational, and gender blind is called into question by research pointing to the ambiguity of tenure and promotion processes (Jones et al. 2014; Beddoes, Schimpf, and Pawley 2014). Intensive interviews with male and female faculty at a large public university in the Midwest revealed that faculty members described the tenure and promotion process as opaque, confusing, subjective, arbitrary, and blurry. Formal university and departmental policies, which spelled out ostensibly rational, measurable, and gender-neutral tenure and promotion criteria and processes, were seen as playing little or no role in defining the criteria for actual tenure and promotion decisions. Although the authors of these studies do present direct evidence of the link between this "foggy climate" for tenure and promotion and gender inequality in STEM departments, they raise the question of whether gender bias may be at play if tenure and promotion decisions in academic science and engineering are governed by the subjective and ambiguous procedures they describe.

The question of whether academic science is still affected by sexist attitudes and practices is obviously critical to determining whether programs such as NSF-ADVANCE remain necessary. Clearly, those programs continue; our literature review turned up several descriptive accounts of ADVANCE programs on university campuses—e.g., Carpenter (2014) on Louisiana Tech and Wadia-Fascetti et al. (2014) on Northeastern. McClelland and Holland provide a particularly interesting examination of the role of departmental leadership in increasing gender diversity in STEM departments. They interviewed a group of 31 STEM chairs and deans, most of whom were male, and 11 of whom were in engineering. They noted that these leaders varied in terms of how much they saw it as their responsibility to effect change; those who were low on responsibility tended to play down the need for change and/or to see it as something that female faculty needed to be responsible for. More generally, McClelland and Holland reported that leaders of both kinds tended to ask men to make relatively small changes—be more sensitive, learn more about your female colleagues—while women were asked to make significant life or attitudinal changes to foster gender diversity.

Interestingly, we also reviewed a "feminist reflection" on NSF-ADVANCE by Morimoto and Zajicek (2014) that took a very different view from the one propounded by Ceci and Williams. The authors accept that academic sexism persists, but wonder whether NSF-ADVANCE programs are up to the task of effectively combating it. They ask whether it is reasonable to expect that a program developed within a major scientific institution can be expected to be the agent through which that institution itself is transformed. Their verdict is somewhat mixed. They criticize many ADVANCE programs for focusing on "individual-level" solutions that don’t attempt to transform existing structures (e.g., mentoring or networking programs that seek to help individual women scientists do better within unchanged organizations). They also note that ADVANCE-funded efforts at organizational transformation can also, inadvertently, reproduce the status quo. For example, programs that seek to raise consciousness and change attitudes in academic departments often make women the consciousness raisers. Similarly, programs to encourage the development of family-friendly policies risk perpetuating traditional gender structures unless both men and women take advantage of those policies. Nevertheless, Morimoto and Zajicek hold out the hope that ADVANCE can be a truly transformative program. They add that it might be possible for programs such as ADVANCE to encourage openness to alternative models of scientific work and scientific success, models that do not construct the scientist as an "ideal worker" with unlimited time and a support structure at home and instead reward activities other than individual grant getting and publication, such as successful mentoring and fostering collaboration.

The workplace: gender bias in industry?

As has been the case for several years, relatively little research published this year focused on employed professional engineers outside the academy, and several of the studies we did find were about countries other than the United States. We reviewed, for example, articles describing the disadvantages faced by female engineers in Portugal and their efforts to cope by trying to make their femininity "invisible" (Saavedra et al. 2014). Tacsir, Grazzi, and Castillo (2014) review the literature on women engineers in Latin America and report a familiar pattern of disadvantage and slow progress. There is a fairly obvious need to find ways to encourage more researchers to examine engineering employment and the experiences of female engineering graduates, especially since research published in previous years shows that many female engineering graduates do not go on to engineering jobs and that at least some women who do, eventually leave. There is an extensive literature on the experiences of women in professional and managerial occupations and the conflicts they encounter, but without focused studies on female engineers, we are forced to assume, without real evidence, that they are simply the same as women in other occupational roles.

We did review a few studies of working female engineers, most of which focused on the issue of bias and employer attitudes. Braun and Turner (2014) conducted a survey and a small, interview-based study with managers in STEM settings to examine their intention to engage in "women-friendly behaviors." Importantly, they found that what managers believed others expected them to do was not an important predictor of their willingness to engage in these behaviors, although this was the case for a subgroup of female managers. The managers were aware of the benefits attributed by advocates of promoting women, but also freely expressed reservations about female managers. Braun and Turner note that they were surprised by the openness with which managers talked about these reservations and acknowledge the existence of stereotypically negative views of women in these fields. The authors are encouraged, however, by the fact that managers’ beliefs and past histories are important predictors of their willingness to behave in woman-friendly ways—those who did so previously continue to do so; those with more positive attitudes to women are more likely to do so. They also note that supportive organizational settings can help encourage managers to be woman friendly. The logical inference is that managers’ practices can be made more woman friendly through efforts to promote attitudinal change, through hiring managers with positive attitudes toward women, and by creating the appropriate organizational climate.

Reuben, Sapienza, and Zingales (2014) point to a more discouraging reality—the persistence of stereotypical beliefs about women’s ability in scientific fields. They conducted an experimental study of 191 undergraduate students to examine their gender preference when it came to performing a math-based task. They found that both men and women prefer to hire men for such tasks. The discrimination is reduced when the candidate self-reports performance data, but the authors note that men are more likely to brag about their achievements, so that limits the degree to which bias is reduced. The authors see this experimental evidence as indicating that actual managers may be affected by the same stereotypical attitudes, which may, in turn, hinder women’s careers in STEM disciplines.

In August, Nadya Fouad reported on the ongoing research project at the University of Wisconsin-Milwaukee focused on the retention of women engineers. Invited to address the American Psychological Association’s annual meeting, her presentation, titled "Leaning in, But Getting Pushed Back (and Out)," received wide media coverage. It was picked up by national newspapers, radio, and online media, perhaps due to the tie-in with Sheryl Sandberg’s book, Lean In, and concurrent media attention to the dearth of women in "tech." Fouad’s research is based on a national survey of women engineers and has been discussed in previous literature reviews. Fouad’s team concludes that workplace climate and lack of advancement opportunities drive women out of the profession. They are currently collecting additional data from both women and men engineers for the next stage of their study.

There was one important contribution to the literature on women in engineering workplaces this year: Bilimoria and Lord’s (2014) edited volume on Women in STEM Careers. The collection includes research from the U.S., Europe, and Australia, examining women’s work experiences, advancement, and leadership roles in STEM fields, both in industry and in the academy. The most important contribution of the volume is to ask readers to pose a different question. Typically, we ask why women leave engineering – the metaphor of the leaky pipeline is clear evidence of this focus. Bilimoria and Lord suggest that we ask, instead, why women stay. The essays collected in the volume examine a variety of organization initiatives that have been shown to make a positive difference for women’s careers in STEM. In addition, the final section of the volume emphasizes the need to reframe organizational discourse and practice to create the conditions under which women will stay. Among others, the essays in this section discuss the gendered character of definitions of success in STEM (which assume the ability to devote all of one’s time to one’s work), the ways in which care work and family responsibilities, while important to both male and female professionals, are not seen as an appropriate topic for discussion, and the need to make gender a more visible part of STEM curricula. Asking why women stay encourages us to examine the characteristics of the organization and the institution, not just the individual characteristics of the women who leave, thereby responding to the criticism that, sometimes, efforts to promote gender equity focus too much on "changing the women."

Conclusion—closing the gap?

Readers of the literature on women in engineering and STEM are generally accustomed to the notion that these fields are a particularly extreme case; that they are quite unlikely to attract women and, while there has been progress, that progress has been slow, limited, and may have stalled. Economist Claudia Goldin (2014), however, constructs an argument that may give us reason to think that things may not be quite so bad, after all. In an important article addressing the residual pay gap between men and women in the United States, Goldin offers another lens through which to examine the status of women in engineering and STEM disciplines in general.

Goldin’s concern is with the overall pay gap between men and women in the United States. She notes that average female pay has not yet caught up to average male pay, but the gap has narrowed significantly over recent decades. Goldin asks what would eliminate the last of the pay gap. She argues that the remaining pay gap is not the result of differences in human capital (e.g., education, skill) between men and women, nor can it be attributed to occupational differences (i.e., the concentration of women in different occupations than men). Instead, she argues that the remaining pay gap is primarily the result of the fact that, in certain occupations, value is placed on working long hours and on continuous employment. Women in these fields are more likely to seek to restrict their hours to accommodate family roles, or may interrupt their employment for childbirth and child-rearing, therefore suffering economic penalties and falling behind their male counterparts. Interestingly, Goldin finds that technology and science jobs are more flexible, and impose fewer economic penalties of this type on female employees.

Although Goldin’s article doesn’t specifically address the case of engineering, it delivers a good news/bad news report on STEM fields and, by extension, on engineering. In some occupational fields, reformers need to worry about both a pay gap and gender balance. Goldin’s argument implies that the problem in STEM isn’t a pay gap; it is primarily access—women who enter STEM fields do reasonably well economically, compared with men. They are better off relative to their male counterparts than are women in business. For various reasons, however, few women find their way into the field, so only small numbers of women benefit from this relative equity.

Goldin’s analysis implies that finding ways to increase the numbers of women entering the field, rather than combating pay inequity in engineering, should be the focus of equity-oriented policy. It is also interesting to note that Goldin’s analysis suggests that, at least in economic terms, the conditions for women in engineering appear to be relatively good. Given that, it would be important to look again at arguments suggesting that women leave engineering because they find it unwelcoming. Are women leaving engineering in unusually high numbers, or are we simply ignoring the reality that women are leaving other occupations, where conditions are less favorable, in even larger numbers? Subsequent research on women working in industry, particularly addressing the perspectives raised by Goldin, is needed to adequately shed light on these questions.

Peter Meiksins is vice provost for aca­demic programs and professor of sociology at Cleveland State University. Kacey Beddoes is a research associate in the School of Civil and Construction Engineering at Oregon State University. Maureen McCusker is a Ph.D. graduate student in the Industrial and Organizational Psychology program at Virginia Tech. Elsa Camargo is a doctoral student in the Educational Leadership and Policy Studies program at Virginia Tech. Katie Boyd is a doctoral candidate in Industrial and Organizational Psychology at Virginia Tech. Peggy Layne, P.E., F.SWE, joined Virginia Tech in 2003 as director of the AdvanceVT program and is currently assistant provost for faculty development in the office of the senior vice president and provost. This article originally appeared on SWE Magazine, SWE is a CFE Media content partner.