Stanford Class receiving instruction via video (1989).
What is the Academic ‘Pipeline’?
“The Leaky Pipeline” is one of the most ubiquitous and widely accepted explanations for low female retention in STEM academic spaces. The pipeline model examines the progression of people in STEM fields through the academic “pipeline”. The model assumes a “traditional” pathway to a tenured faculty position starting from high school to undergraduate study, to graduate school, to a postdoctoral position, and then through the tenure-stream faculty ranks (i.e., assistant, associate, full professor). According to the model, all groups “leak out”, but groups can leave at different rates and at different locations in the pipeline. The pipeline model was first articulated by Sue Berryman in 1983 (Berryman). Since then, her advice for increasing representation by reducing attrition or increasing the candidate pool has been commonly touted. Diversity initiatives since the inception of this model have almost exclusively focused on the recruitment and retention of underrepresented groups during the educational stages (i.e., high school, undergraduate, graduate school). Federal funding agencies provided grants for recruitment and retention (NSF, “Search NSF”) and diversity officers were hired to scrutinize admission. The notion was to increase numbers, and eventually, this would result in a diverse faculty.
The route (“pipeline”) of an academic in STEM is from undergrad, graduate school, postdoc, to faculty position and eventually tenure.
Nearly fifty years later, these promises ring hollow. The pipeline, in one sense, has become deliberately “clogged” between the doctorate and faculty stages. Doctorate production in some fields has reached 50% gender parity. However, these improvements are not translating into more diverse tenure-track faculty hires. There is a PhD bottleneck, which often gets shunted into short-term contract hires. The research university, for all its talk of diversity, has refused to increase tenure-track jobs. And, in essence, it has refused to support those in tenure-track jobs or those seeking tenure track jobs who do not fit into the mold a white, male heterosexual subculture. In another sense, the pipe has become “rusted” at the educational stages, in that this entrenchment of the white masculine space has resulted in lower retention and increased alienation of those who seek STEM degrees. Ultimately, the idea of increasing the “pool” that flows through the pipeline is naïve, and ignores serious institutional barriers for diversity, but is unsurprising. As Banerjee and Pawley argue, the pipeline model is popular because it situates STEM underrepresentation of women as a numbers problem, rather than a structural problem (Banerjee and Pawley).
Bachelor's degree Attainment
Even though nearly half of STEM bachelor’s degrees were awarded to women in 2016 (NSF 2019b), the majority of those degrees were awarded in the fields of Psychology, Biological Sciences, and Social Sciences (NSF 2019b).
Computer sciences, Engineering, and Physical Sciences are each individually below 21% of bachelor’s degrees awarded to women annually (NSF 2019b).
For African American women and Latina women, bachelor’s degrees awarded in these subfields are less than 4% (NSF 2019a).
Doctorate Obtainment
Even though doctorate attainment was at gender parity by the 2000s, there was deep sex and race inequities by field. By 2004, 51% of all earned doctorates in the U.S. were awarded to women (Hoffer et al. 2004). Specifically, women earned the majority of doctorates in such fields as education, the humanities, and the social sciences (p. 12).
There is persistent sex segregation among “prestige” STEM doctoral programs and fields, despite the fact that nearly half of doctorates in the United States are earned by women (Weeden, Thébaud, and Gelbgisera 2016).
From 1974 to 2004, doctorates for women in physical sciences increased from 8% to 27%, and 1% to 18% in engineering (Hoffer et al. 2004). But by 2021, these levels stagnated with women only earning 32.5% in physical sciences and 24.5% in engineering (NCSES 2021).
From the 1970s, doctorate obtainment for women of color in STEM have only marginally improved.
Of 42,980 S&E doctorates awarded in the U.S. in 2019, white women received 19%, Latina women 2%, Black women, 2%, Asian women, 3%, women temporary visa holders 12%, American Indian or Alaskan Native women 0.08% (NSF 2021).
Faculty Obtainment
By 2017, there was a low % of female faculty at most levels of the professoriate in the U.S.
In 2017, Black or African American women made up 7% of the U.S. population, American Indian or Alaskan Native 1%, Latina women 16%, and Asian women 3% (Census 2020). Only 1% of full professors in science and engineering fields were Black women, 1% Latina, and 0.1% American Indian or Alaskan Native.
By 2005, only 14.8% of faculty members in top S&E research departments in the U.S. were female, with only 8.5% in engineering, 8.4% in physical sciences, 33.4% in psychology, and 20.2% in biological sciences (Beutel and Nelson 2005).
By 2005, only 1.8% of faculty members in top S&E research departments were Black, 2.3% Latino, and 11.7% Asian (Beutel and Nelson 2005).
Percentage of Women in faculty positions in 2017 in S&E fields by race; adapted from NCES and NSF data. Professor = tenured; Associate and Assistant professors are tenure-track.
The Curious Case of Computer Science
1994 Students in a Computer Lab at Spokane Falls Community College (Photo by Bob Rowan/Corbis via Getty Images)
Notably, the percentage of bachelor’s degrees awarded to women in the computer sciences has declined over the past decades. After an initial surge in female enrollment in computer science majors in the 1980s, there was a subsequent plateau and decline in female enrollment (Rossiter 2012). In 1997, 27.2% of bachelor’s degrees in computer science were obtained by women which fell to 18.7% in 2016. Women’s shares of doctorates in computer science fell from 21.7% in 2006 to 20.1% in 2015. (NSF 2019b). This was likely due to the lack of solidarity and support many women experienced in computer science majors. In fact, girls and boys demonstrate similar interest in computer science during childhood, but can often be discouraged by masculine stereotypes and negative discussions of women and tech.
Number of bachelor’s degrees awarded to women in computer science in the United States, screenshot from (Rossiter, Margaret W. 2012 p. 61). Bachelor degrees rose rapidly from the late 1970s, peaked around 1986, and then rapidly fell.
A Carnegie Mellon study interviewed students from 1995-1999 confirmed this male-oriented culture in computer science programs (Hill, Corbett, and Rose 2010). Unsurprisingly, women’s interest and enthusiasm for the major often waned after a few semesters in the department (Hill, Corbett, and Rose 2010). At an undergraduate level, women of color are less likely to report a sense of belonging in their STEM major compared to white men and women and men of color (Rainey et al. 2018). They may be further discouraged once they enter the workforce, where they are less likely to be hired or promoted despite having equal qualifications to their male counterparts.
The Myth of “Too Many” PhDs
While there was increasing support for diversity in STEM, the nature of academia started to change. Corporate tactics began to take hold, which worked to weaken faculty governance, conserve labor costs, and increase revenue. Since the 1980s, the percentage of tenure-stream faculty appointments declined (Baldwin and Blackburn 1981). There was also an increased tendency towards lateral hiring, in the chase for “prestige”. By the early 1990s, prestigious institutions like Harvard, obtained about 71% of faculty from lateral appointments, not freshly-minted PhDs (Cross 1994). Between 2008-2012, full-time tenure track faculty only increased 1% in the United States while full-time contingent faculty increased 11% and part-time contingent faculty increased 18% (AAUP 2018). This data undermines the common administrative “excuse” that there was a scarcity of qualified candidates of color and that there were “bidding” wars over hiring them (Tuitt, Sagaria, and Turner 2007). Instead, it reveals that administrators have deliberately “narrowed” the pipeline at the faculty stage.
Case in point, STEM fields where awarded doctorates are at 50% gender parity, still low percentages of female faculty. By 2004, women made up about 50% of life sciences doctorates (NSF 2021) but were only 26% of full professors in life sciences by 2017 (NSF 2019c). If it truly was a “pipeline” issue in faculty hiring, then the % of female faculty in Life Sciences should have been close to 50% by 2017. The persistent gap between female earned doctorates and full professorships cannot be attributed solely to individual choice to leave academia since the “numbers” at the beginning of the pipeline are higher. For example, the field of mathematics has had significantly higher attrition rates for female faculty (Kaminski and Geisler 2012).
Earned doctorates by women in S&E fields in 2004 v. Full Professorships in 2017. Even in a field like Life Sciences where female doctorate attainment was 50% in 2004, full professorships were only held by 26% of women in 2017.
Therefore, postdoc positions became a frequent stopping point for many in the academic “pipeline”. Once considered prestigious and rare, postdoctoral fellowships became standard and ubiquitous with no guarantee of gaining a tenured-track position (Rossiter 2012). By the 2000s, these positions required mobility, long hours, and minimal benefits with an uncertain job outcome and ultimately did not become an effective “entering wedge” for a career in academia. Rossiter argues that postdocs became a new way to exploit young scientists who faced uncertain long-term career prospects (p. 118). Nearly half of female postdocs were international students on visas, which also made them more vulnerable to exploitation (pp. 118-119). However, a double standard emerged that persisted in the 1990s, with more men getting tenure-track positions without postdocs than women (p. 140). Furthermore, women with PhDs (57.5%) were less likely to have academic employment commitments at the end of their PhDs than men (48%) and were less likely to be able to obtain postdoctoral appointments (23%) than men (30%) (p. 29). Even women’s colleges had few tenured women faculty into the 1990s (Rossiter 2012; Roush 1996). Unsurprisingly, postdocs were often a point where women left academia.
Common recommendations for this “postdoc crisis” are to constrain PhD. production and provide better career guidance for non-academic jobs (Micoli and Wendell 2018). This is particularly ironic since there is such inequity at later stages in the “pipeline”. Since women and URMs are far less likely to get past the “postdoc” stage than their white, male counterparts, critics should take a closer look at who is “leaking” out before making such broad recommendations. Furthermore, the solution of “constraining” PhD. production is endemic of neoliberal discourse and undermines the purpose of higher education: to educate. Denying people higher educational opportunities is the wrong tactic to take.
In a sense, there appears to be little regard for who is in the pipeline and what their experiences are. It is only the worker’s economic value to the university that matters. And if PhD production were constrained, new tenure-track faculty jobs would not be created, as evidenced by the college enrollment surges of the 1990s and lack of tenure-track job creation. Ultimately, the university will find another source of cheap labor.
The pipeline is a myth because:
Lie: Women and women of color are not interested or don’t have aptitude for STEM
Truth: Women, especially women of color, are unsupported and unwelcomed into many S&E majors including computer science and engineering. Federal funding and diversity programs focus too strongly on recruitment, and not enough on supporting current students or reforming culture. Gifted and Talented programs have also historically catered to white boys and overly relied on flawed metrics like IQ to determine STEM aptitude and promoted the lie that boys have more talent for math then girls.
Lie: The pool of female STEM doctorates is too small.
Truth: Some fields like Life Sciences have had 50% of women earning doctorates since the early 2000s and only a 1/4 of full professors are women. Universities prefer to hire cheap labor (grad students, postdocs, adjuncts) to teach and conduct research.
Lie: There are “bidding” wars for female faculty, especially women of color.
Truth: Tenure-track faculty job creation has stagnated since 2008. Postdoc and contingent faculty position creation has increased rapidly during that time. Lateral hiring for faculty positions is increasingly common.
Ultimately, the STEM ‘pipeline’ is the best myth out there: it lays the burden on the individual not the structure.
Micoli, Keith, and Steve Wendell. 2018. “History and Evolution of the Postdoctoral Scholar in the United States.” In The Postdoc Landscape: The Invisible Scholars, eds. Audrey J. Jaeger and Alessandra J. Dinin. Academic Press, 1–13.
Rossiter, Margaret W. 1995. “Faculty and Major Universities: The Antinepotism Rules and the Grateful Few.” In Women Scientists in America: Before Affirmative Action 1940-1972, Baltimore and London: The Johns Hopkins University Press, 122–48.
Rossiter, Margaret W. 2012. Women Scientists in America: Forging a New World since 1972. Baltimore: The Johns Hopkins University Press.
I love how you laid this out so clearly and anticipate sharing this with folks in the workshop I do. Another issue I've seen: academic culture remains alienating to those who have been historically excluded/marginalized. Often, they cannot stay in academia and retain their values or move from what originally motivated their research. So, people self-select out to pursue other careers that better align with their values and motivations. I often tell administrators that we are losing our BEST people because of the stagnancy of university life.
Great article! Your articulation of the leaky pipeline of academia and its connection to the underfunding of universities really resonates.
When I was working in tech as a ghost writer for executives, a woman exec told me that the ad industry also played a big role in the reduction of women entering computer science majors in the late 80s and onward. The story went that while trying to get more computers into the home and push computer use, advertisers also began pushing the idea that computers were for white men and “hobbyists”.
There certainly seem to be gender roles in these ads from the 80s where women are typically typists or using the personal computer for family maintenance (grocery lists, etc). Out of curiosity, have you seen any studies connecting gender roles in advertising to that dip in enrolment post-80s?