Here is an examination of the scientific flaws in the recent New York Times (NYT) Op-Ed: “Academic Science Isn’t Sexist.” The Op-Ed authors, psychologists Professor Wendy Williams and Professor Stephen Ceci, put forward various wide-sweeping statements about the effect of gender on academic careers of women scientists. The article outlines the fact that women make up a minority of junior faculty members, particularly in maths-intensive fields like engineering and computer science (25%-30%) and an even smaller proportion in senior positions (7%-15%).
Williams and Ceci argue that much of the empirical studies that established gender inequality in academia are outdated (mostly published prior to the year 2000). They argue that more recent data show that inequality has been diminished in academia. The researchers claim that women are promoted and remunerated at the same rate as men – except in economics. Williams and Ceci further argue that women’s numbers have been steadily growing in the life sciences and psychology. They note that the proportion of women in maths-intensive fields has also been growing, but not as much. Their analysis attempts to explain why this is the case.
The central argument presented in their NYT article is that women would fare well in maths-intensive subjects, “if they choose to enter these fields in the first place.” To put it another way, the problem as they see it, is that gender inequality is a myth, and that the discrepancies between men and women would be reduced if women chose to stay in STEM.
The Op-Ed is based on the co-authors’ study published in November in the journal, Psychological Science in the Public Interest. In their study, Ceci is first author and they are joined by two economists, Professor Donna Ginther and Professor Shulamit Kahn. The research team see that the sex variations within the fields of Science, Technology, Engineering and Mathematics (STEM) represent a “contradiction” and a “paradox.” The logic of their argument is that because there are more women in STEM fields today in comparison to the 1970s, and because there are different patterns of attrition amongst various disciplines, this is evidence that sexism in academia is a moot point. The crux of their argument is simple: if there are differences between men and women’s career trajectories in STEM, these arise from personal preferences, and not due to a culture of sexism.
The are several problems with the Op-Ed, which overly simplifies the body of literature the authors reviewed, but the analysis of study itself is highly flawed. The most glaring issues include the concepts used, such as the authors’ confusion of sex and gender and how these relate to inequality. Another set of problems arise from the authors’ methods. Put simply: the way they measure gender inequality does not match the data they have available, and their interpretation and conclusions of the data are therefore invalid. In science, a study can be seen to be valid when the phenomenon measured matches the instruments used. The concepts, data collection and analysis need to match the authors’ research questions. This is not the case with this study.
Let’s start with the key concept the authors measured: gender inequality, which is also discussed as “academic sexism.”
“Sex” Versus Gender Inequality
Sociology defines sex as the biological categories for men and women, which draw on ideas of physical differences. Many people in society perceive that sex differences are fixed and innate – that men and women are simply “wired differently,” and that as a result, women and men have different skills and interests. This idea is important in STEM because, as I’ll show, the notion that girls and boys are categorically different (for example, the idea that girls are simply not as good at maths) is a myth that perpetuates inequality.
Gender is a more fluid definition of how people understand their identities and their relationships to other people. Gender is more than what our bodies look like. Gender is a concept measuring how our ideas of sex are shaped by culture. Gender is the reason why experiences of being a woman in one society differs from women’s experiences and life chances in another culture. The concept of gender captures positions beyond male/female. It includes transgender, intersex, queer, and other gender positions. To match Ceci and colleagues’ data and analysis, I discuss differences between (cis) men and (cis) women, but I will later show why this narrow gender focus can also obscure diversity issues experienced by minorities in STEM.
Gender matters in STEM because science and technology fields have been dominated by men; specifically White, heterosexual, educated, middle-class men. This means that as more women have entered STEM fields, our presence threatens the status quo. The fight for gender equality continues. The fact that Ceci and colleagues’ article was even published in the first place is evidence of this ongoing struggle. Despite countless scientific evidence showing how women are disadvantaged with respect to hiring, pay, promotions, publications, awards and all other aspects of science – women and our allies are still having to debate even the most basic issues, such as whether gender bias really exists.
The greatest problem with Ceci and colleagues’ study is that they present their analysis as if STEM professionals work in a vacuum; as if their work productivity is separate from the rest of society. STEM professionals are people, and social dynamics are always a factor when people interact with other people. Culture, race, sexuality, socio-economics and other social experiences that govern the personal lives of STEM workers also influence interactions and processes at work. Are women making choices about their lives that are different from men? Yes. The thing is that, as I will show, these individual choices are influenced and constrained by institutional factors.
An analysis of sexism in academia needs to seriously address gender as a social system, not simply document superficial differences between men and men. Ceci and colleagues are simply looking at the outcomes of women’s STEM careers in comparison to men’s, without adequately measuring how these outcomes arise, and how they’re connected to broader socio-economic patterns in society.
When Data and Methods Don’t Match Conclusions
Beyond the paper’s conceptual problems, the analysis is riddled with methodological confusion. The authors claim to present a life-course analysis. This is an established methodology within the social sciences, that takes into consideration how individuals experience identity, health, family or other social phenomena at different stages. Typically, this means collecting data about childhood, adolescence, adulthood and late age, and identifying how significant life events or transitions (marriage or children for example) affect individuals’ decision-making. A life course approach will take these individual biographies and connect them to broader socio-economic patterns.
A life course approach usually involves following a cohort over a long period (collecting longitudinal data over a number of years). Otherwise, it can also involve the analysis of qualitative data, such as people’s oral histories (gathered via interviews or other observations). The problem is that Ceci and colleagues have not been faithful to the life course approach and their analysis is profoundly flawed as a result. They look at different stages in women’s education and careers as if they are independent, rather than interconnected. For example, the fact that studies show that senior male researchers prefer not to take on women as summer interns or as students is treated separately, and dismissed by Ceci’s team as a factor in the hiring patterns of women. There are incremental inequalities at various stages of academia, which make it even harder for women to be successful.
There are many other methodological problems throughout the study. The authors argue that studies supporting the effect of sexism in academia reach this conclusion using a limited dataset. The authors criticise research for having smaller sample sizes, or for being outdated. Yet in various places through their analysis, they draw on studies from the 1980s and 1990s to establish their own argument that academic sexism has been eradicated, particularly by showing how trends in recent years are not as bad as previous decades. Similarly, while the authors make the point that international and cross-disciplinary studies are important, at other points, where such studies disprove their point, they rely on American data from one field. Their focus is academia, yet on a couple of instances they use data on non-academic women to boost their argument. This selective treatment of the data is problematic to say the least.
Finally, in a study that goes to great pains to argue that publishing is a prime measure of academic egalitarianism and “productivity,” this paper is in itself a curious example of the established literature. Ceci and colleagues show that men publish more than women, especially in senior ranks, but they argue this is not due to sexism (they say that women can’t keep up with the “fast paced” nature of work). Earlier, I noted that the authorship of the NYT article is different from the study. In the Op-Ed, Williams, a White woman, is first author, and Ceci, a White man, is listed second. In their study Ceci is first author, their two women colleagues are second and third authors, and Williams is last (in psychology, authorship is different than the natural sciences, where each author is listed according to their academic contribution, based on a points scale). It is does not escape attention that this study on sexism has a White male Professor as first author, spearheading the case against sexism with a research team of tenured Professors. (I note two of the authors, Ginther and Kahn, have published on women of colour in STEM similarly focusing on individual choices, and without reference to institutional factors such as racism or sexism.)
The present study barely touches on culture, and fails to account for the intersections of sexism with other forms of academic inequality like class, race, sexuality and disability. This is the theory of intersectionality: that experiences of inequality are inextricably linked. Any study serving the interests of White and male privilege in academia should be treated with extreme caution.
Even though they do not have the evidence to support their conclusions, they nevertheless argue that women “opt out” of academic competition. They see that the sex discrepancies in STEM arise by women choosing to leave STEM. They argue this happens partly because women cannot maintain the same level of “productivity” of their male peers. They also see that women leave academia to have children, and by so doing, they accept the “child salary penalty.” Both of these arguments are sexist at their core. They presume that the unfairness of the academic system, which penalises women in ways that men are not, should not be questioned.
Where other researchers have found systematic gender bias, Ceci and colleagues see individual choices and disciplinary differences. Due to the variety of outcomes amongst women academics, such as differences between maths-intensive and the life and social sciences; differences amongst women working in top tier and less elite universities; and differences between women with and without children – the researchers argue that these disciplinary patterns nullify the question of gender inequality. This is not the case.
Ceci and colleagues see that one of the biggest reasons women “opt out” of academia is due to their family choices. Having children is not really the issue, as research shows that men’s academic career is actually bolstered when they have children. The data are telling us something different than what Ceci and colleagues would have us believe.
In summary, the data show that women in STEM fields on average are significantly less productive than men at the assistant-professor rank. Economists, including those coauthoring this article, believe that this is prima facie evidence that shifts at least part of the responsibility for women’s limited academic success away from employing departments—unless one wants to argue, as some do, that more should be done to help mothers with young children. Unfortunately, there are relatively few studies that provide compelling explanations for these productivity differences. [My emphasis]
That’s a lovely narrative about lack of evidence. It is unfortunate, for Ceci and colleagues at least, that this economist appraisal does not match the wealth of other social science research that has been conducted over the past 60 years. This body of evidence shows that, while domestic labour arrangements have been changing, women still do more domestic work than men and that this affects gender inequality. This is known as the double burden – it means that even as women work longer paid hours, their caring responsibilities do not diminish. To be sure, both men and women are forced to make trade-offs in their work and family lives, but the costs are higher for women’s paid career in many nations, especially those that do not support women adequately. The studies cited by Ceci and colleagues show that women are overburdened with childcare responsibilities, teaching, pastoral care of students, administration and other duties, while their male colleagues spend less time on family responsibilities and other teaching and admin, and more time doing research that better helps their career. How is this not evidence of a sexist culture?
Ceci and colleagues present their analysis of academia as discrete stages, rather than a fluid and dynamic process that is impacted by external factors. They presume that academia functions as a more or less even playing field, where everyone has an equal chance of success so long as they are willing to be “productive.” This is not the case. Women are not simply academics; women’s experiences of gender in STEM are mediated by their culture, language background, family and social networks, their experience of race, their sexuality, their disability or illnesses and other social identities. I will be elaborating on these points in an upcoming post (and will link back later), but here I sketch some broad issues:
- Women academics of colour face the presumption of incompetence. This phrase captures experiences of racism, such as being routinely asked to legitimise academic qualifications by students and colleagues
- Lesbian, bisexual, queer, transgender and intersex women in academia are denied a full sense of gender inclusion, with tacit and overt homophobia making it difficult or unsafe for women to “come out” at work
- Academics and students living with a disability and chronic illness navigate an uncaring system that reinforces the idea that disabled people can’t keep up with academic ambitions.
Gender exclusion is not discreet: for minority women, sexism is enmeshed with racism, homophobia, disability and other forms of discrimination. Trying to separate these issues from a study on sexism is profoundly counterproductive.
Why All of This Matters
Popular media articles such as that written by Williams and Ceci are incredibly damaging to gender equality efforts in general, and they undermine the progress of gender diversity in STEM more specifically. As this Op-Ed is published in a reputable, widely-read news source, The New York Times, the article contributes to the public’s misunderstanding of academia and scientific practices.
The article feeds into the pre-existing idea that gender inequality is a myth. Inequality is an issue that tangibly affects women scientists all over the world, as the United Nations has shown. At STEM Women, we’ve shown that the transition from education to a STEM career is riddled with institutional bias. As the life course methodology shows, experiences at each life stage feeds into later experiences. Despite evidence to the contrary, the authors argue the problems women experience in STEM are not driven by sexism. Instead, the researchers argue the issue is that women are not “leaning in” enough, evoking Sheryl Sandberg’s argument that women would be more successful if they only asked for more help and acted more confidently. This view is ignorant to the multiple institutional barriers that women and minorities face. An analysis of sexism in academia that does not adequately measure institutional factors is, in fact, not an analysis of sexism. It is just a replication of gender bias.
The companion piece to this post is now published on Diversity Postdoc. The article discusses why social scientists need to be especially careful when publishing about gender and science in the popular press.
Original photos by Argonne National Laboratory, CC 2.0. Image 1. Image 2. Image 3. All adapted by STEM Women.