False choice: Is the underrepresentation of women in science by choice or by discrimination?
This post is about Ceci and William’s PNAS article, Understanding current causes of women’s underrepresentation in science, which has spawned a particular kind of narrative -- one that has been around for a while, but which now bears the imprint of evidence. This narrative is captured in a recent headline from ScienceDaily: "Choices -- not discrimination -- determine success for women scientists, experts argue." The implication is that if only women would stop complaining about their feelings of “isolation, dissatisfaction and discrimination” (p. 3160), we could pay attention to important problems that are real and not imaginary.
Ceci & Williams 2011, mostly in their own words
Ceci & William’s goal is to find out whether there is currently sex discrimination in three important areas: (1) manuscript reviewing, (2) grant reviewing, and (3) interviewing/hiring. “Current” means within the last 20 years. They make a strong case that while such discrimination may have taken place in the past, there is no evidence of discrimination against women in current large, carefully analyzed studies of real world reviewing and hiring data. On p.3161, Ceci & Williams conclude that “past strategies to remediate women’s underrepresentation can be viewed as a success story; however, continuing to advocate strategies successful in the past to combat shortages of women in math-based fields today mistakes the current causes of women’s underrepresentation.”
Still, it is true that men come out ahead of women in these three areas even in the last 20 years. Ceci & Williams point out, however, that these differences go away if you control for institution, teaching load, funding, and research assistance. “A key issue", they say, "separable from sex discrimination, is why women occupy positions providing fewer resources and what can be done about this situation. Some of these choices are freely made; others are constrained and should be changed.” And (in the supplementary text) “When women PhD recipients choose not to apply for tenure-track posts, their refusal represents a choice, one that many of their male and many of their female colleagues do not make.”
Some grounds for skepticism
Here are just a few of many reasons to be skeptical of Ceci & Williams' claim that they have definitely debunked the existence of sex discrimination in grant/manuscript reviewing, and interviewing/hiring...
(1) Narrow definition of discrimination. Ceci & Williams explicitly state that their definition of sex discrimination is restricted only to cases in which there are gender disparities once you control for resources (institution, teaching load, funding, and research assistance) (p. S2). Maybe you agree with this definition, in which case, break out the champagne! No more discrimination! But it seems to me that accepting this definition requires accepting that it is all down to women’s “choices” (more on this below) whenever they end up with fewer resources. That’s a biiig assumption and very unlikely given the psychology literature.
(2) How to cure discrimination. Assuming Ceci & Williams’ analyses are sound and we agree with their narrow definition of sex discrimination, do we take a lack of sex discrimination as reason to halt all our efforts to prevent discrimination? Well that depends on what mental model you choose. Ceci & Williams seem to think of the problem of sex discrimination in a business framework. Like a debt we need to pay, a certain amount of effort is required to reach the desired sum. Once we’ve hit our target, we’re done. Therefore, once we’ve reached a point at which there is no gender imbalance, Mission Accomplished. By the same lights -- once we reach a certain level of parity in civic realms, why continue to waste resources enforcing anti-discrimination laws? Well, you may respond, perhaps sex discrimination is more like a chronic disease. If we support this analogy, then Ceci & Williams have shown that we’re out of the Intensive Care Unit -- but that doesn’t mean we can shirk the effort necessary for future disease prevention and maintenance.
(3) Conflating causes and forms of discrimination. See how Ceci & Williams move effortlessly from talking about the forms of discrimination to an inference about the causes of discrimination on p. 3158: “Although real barriers are still faced by women in science, especially mathematical sciences, our findings suggest that historic forms of discrimination cannot explain current underrepresentation, and that resources should be redirected toward current rather than historical causes of women’s underrepresentation in math-based careers.” Sorry, C&W, I have to dispute that. Just because the form of discrimination has changed over time, doesn’t mean the causes of discrimination have also changed. One hypothesis is that the cause of discrimination of women, where it exists, is a set of unfairly held stereotypes and misapplied generalizations that hold women are inferior in math and science. Just because these ideas are not acted on in the same way they were in the 1950s doesn’t mean these ideas are no longer endorsed -- or even if they are not explicitly endorsed, that they do not influence behavior.
(3) False choice between discrimination and choice. By asserting that discrimination and choice are distinct, Ceci & Williams pave the way to the conclusion that women’s underrepresentation is something that they do to themselves -- whether freely or as a way to satisfy women-specific societal and biological imperatives. Sorry, C &W, I have to dispute this too. Differences in choices made between two groups may reflect that they don’t actually get to make the same choice. Perhaps one of these groups is getting a raw deal. Even within Ceci & Williams’ own paper, we find examples of bias against women in science and academic contexts found in scientific studies conducted since 2000 and thus “current”. Here’s one on p. 3160: “For stereotypically male tasks, if there is ambiguity about the quality of the women’s contribution to a joint task, it is downplayed.” Well, gee. As Alison Gopnik writes in Slate,
“Those studies show that women are subject to bias from the very start of their careers. Is it any wonder that many of them, keenly aware that their efforts are being downgraded compared to those of men, would withdraw from a competition that is systematically unfair?”
Ceci and Williams argue that using the language of discrimination is “not only inaccurate, it also leads to interventions that are unlikely to remedy the underrepresentation.” The language of choice may do the same thing.
And now for something completely different
In the above, I tried to provide some reasoned arguments to dispute the idea that Ceci & Williams’ analysis proves that discrimination is dead. And I’d like to say that these were the main reasons why I disagree with their conclusions. Deep down, however, I have to admit that some of my motivation comes from flagrantly anecdotal person experiences and I feel it would be disingenuous to conclude without touching upon them.
Scientists and intellectuals have been my heroes since childhood. I don’t know how it started or why, but that’s how it was for me. In elementary school, I had a Marie Curie phase. I loved the idea that someone could be so curious about something that they would follow their passion no matter where it led, even if people told them they couldn’t. I believed that Marie Curie inoculated me against future foes. If someone important told me one day that I couldn’t do something because of some stupid stereotype, I wouldn’t let them dissuade me. Of course, this never happened. No one told me I couldn’t do something because I was a woman, probably because I started my training 100+ years after Curie did. The idea I had prepared to fight was something I never encountered. Phew.
What I ended up really struggling with, what I wasn’t prepared for, were the smaller choices that together make up the big choice of whether or not you want to be a research scientist. As an apprentice researcher, you often depend on the advice of those senior to you about what projects are appropriate, what subfields and techniques you are well-suited for, and so on. Based on what happened with those decisions, you would have different outcomes and consequently, different choices later on. Like a Choose Your Own Adventure book for PhDs, early scientific training is a massive decision tree of contingencies. No one can know all the consequences of such decisions, but you make your best guess with the guidance of your mentors.
So, as in this Slate article, you may find yourself a young scientist in a new lab tangling with some new project or technique, and urged to decide whether something is your “true interest” or your “real talent.”
Perhaps you are struggling to master new skills in order to answer a question you judge to be important and interesting -- aka, the story of the last five years of my life. Perhaps you are starting to doubt yourself, or a mentor suggests, "why don't you do what you're really good at?"
Now, is this person guiding you to a different path because they know you well and have assessed your abilities and potential with unbiased accuracy? Will your answer to this question depend on whether the person guiding you is or isn’t a women, and if you are or aren’t a women, and if the path is or isn’t leading in the direction of a traditionally male-dominated field? As Vendatam writes in this article, it could be that the reasons we judge ourselves and others to be well-suited for a particular goal may be “only partially—and perhaps tangentially—tied to our interests, determination, and talent.”
Imagine that in the future, we’ve done all the right studies and we’ve found out that Ceci & Williams (and Tierney and Hoff) are right. Gopnik and Vendatam and everyone else including me are wrong. Awesome! I would love to know that all my worries are unfounded and that I can devote that precious thoughtspace to other topics. I would love to know that I can believe the best about the human community I am a part of, and that everyone who comes through a science PhD program ends up doing exactly what they most want to do in life.
I just can’t help but wonder if that might not be true...
Further reading
Valian, V. (1999). Why so slow?: The advancement of women. MIT press.
Margolis, J., & Fisher, A. (2002). Unlocking the clubhouse: Women in computing. MIT press.
science-professor.blogspot.com
Helen De Cruz 27 March 2011 (16:01)
Thank you for your thoughtful post, and for bringing the Ceci & Williams debate on the Cognition & Culture radar. What strikes me is the following sentences "When women PhD recipients choose not to apply for tenure-track posts, their refusal represents a choice, one that many of their male and many of their female colleagues do not make". Now, I and most of my female colleagues in philosophy in Belgium are not on tenure track positions, but this is not because we do not "choose to apply" for such posts. I have seen several of my female colleagues leave the profession because lack of support, mentorship, and opportunities, which are readily offered to their male colleagues. I have seen at least two women not being selected for the few lecturing positions that are open because "she would probably not be able to combine the heavy duties of a professor with having to care for her children".
Nicolas Baumard 27 March 2011 (17:29)
You think that there is no prejudice any more? Compare how Science Daily and the Guardian cover the story! Science Daily: Choices -- Not Discrimination -- Determine Success for Women Scientists, Experts Argue The Guardian: Sexual discrimination against women in science may be institutional In other words, Science Daily suggests that the fight for equality is over--inequality in science is just a matter of choice--while The Guardian emphasizes that the discrimination does not consist in prejudice, but in the building of society--female are constrained in their choices, having to give up some opportunities because society does not offer arrangements for individuals in charge of a family. There may be no prejudice that women are not as smart as men, but there is still the prejudice that societies offer equal opportunities to men and women.
Nicolas Baumard 27 March 2011 (18:03)
I agree with Davie that 'choices' may not be the right term if these 'choices' are actually 'forced choices'. However, Ceci and Williams' article may be a good move in the debate. If what mainly prevents women from getting good positions is their family obligations, then we are loosing lots of time in focusing only on prejudices and not discussing institutional changes (child care, part time work in competitive jobs, etc.). In other words, the debate may not be 'is prejudice over?', but 'what is the most effective way, today, to help women in science: fighting against prejudices or providing more resources to helping women take care of their family?'
Davie Yoon 29 March 2011 (21:51)
A report on MIT's efforts to de-institutionalize gender inequity. It sounds like they have done good job of making sure women no longer make less money and have access to fewer resources and are pressured to do more teaching. However, it sounds like there is plenty more work to do: women seem overly tapped for service commitments and a new perception as being held to lower standards. http://www.nytimes.com/2011/03/21/us/21mit.html My favorite quote [quote]“To women in my generation, these residual issues can sound small because we see so much progress,” said Nancy H. Hopkins, a molecular biologist who instigated the first report. “But they’re not small; they still create an unequal playing field for women — not just at universities, and certainly not just at M.I.T. And they’re harder to change because they are a reflection of where women stand in society.” [/quote]
Davie Yoon 29 March 2011 (21:52)
Many thanks Helen and Nicolas for your comments. Helen, it sounds like conditions are ripe for a study of sex discrimination in Belgium! Yikes. I'm sorry to hear the difficulties you and your colleagues face and even more grateful that the situation in the US has improved as much as it has. Your comments make me realize that as much as I would like progress to continue, the progress already made is nothing to sneeze at! Nicolas, your comments have made me think much more about institutionalized differences and discrimination and whether they are one and the same. Take the high school system I went to. I was a student at St. Lucy's Priory High School, an all-girls school -- boys went to Damien High School. I believe that both schools were run by the same Catholic governance organization (although St. Lucy's had nuns at its head and Damien had priests). As a student at St. Lucy's from 1996-2000, I was really frustrated that Damien had a whole range of science and math classes St. Lucy's didn't: environmental science, computer science, and more advanced biology, chemistry, and calculus classes. Why did the boys get to learn and do stuff we girls couldn't do? The boys had more opportunities to study math and science, plain and simple. This was an institutionalized gender difference in math and science training and gave girls a poorer foundation for futures in these fields. But was it discrimination? I remember the science teachers at St Lucy's being very supportive of their students -- I never got the sense that they thought we girls were less capable. When I tried to make the biology course at St. Lucy's more like Damien's with real experiments rather than just textbook exercises -- I was met with excitement and encouragement. If someone offered the school the better resources for science and math teaching -- I would expect them to accept happily rather than disdain the offer as inappropriate for girls. So in the end, I think that Ceci & Williams are right to shine a spotlight on institutionalized sources of gender inequity. I do wish that they hadn't gone out of their way to highlight institutionalized inequities are NOT discrimination, but I see their point that it is a different kind of source of female underrepresentation, requiring a solution at an institutional rather than individual level. I would NOT agree that these institutionalized differences are by definition free from the kinds of ideas we would call discrimination. Who institutionalized the rules that put women at a disadvantage? Who resists their change and why? An institution is not an individual, but individuals run it.
Davie Yoon 1 April 2011 (08:39)
I recently came across this commentary in the March 2011 issue of Perspectives in Psychological Science. It contains some pretty exciting data and I thank Kessel and Nelson for sharing these results -- especially as a counterpiece for Valla and Ceci's discussion about the relationship between sex-based differences in prenatal hormones, and brain and behavioral differences relevant for quantitative STEM fields. Over the past few decades, Kessel and Nelson show gains for women across STEM fields. Over the same period (see C&W), we have seen changes in society's views about women and programs to attenuate discrimination. Presumably prenatal hormone exposure has not changed during this time. It therefore seems reasonable to conclude that addressing structural and social impediments for women leads to increasing numbers of women entering STEM fields -- an increase we have no reason to think would be ultimately limited by the fact of hormonal differences between the sexes.
Christophe Heintz 1 April 2011 (22:51)
A cognitive study on the topic is done by the project implicit at Harvard. You can take a test about your implicit prejudices at: https://implicit.harvard.edu/implicit/Study?tid=-1
Rosalie Esther Ivády 26 August 2011 (16:58)
Thank you Davie for highlighting this article and all the invaluable comments on it, it is indeed an engaging debate, particularly with similar arguments lately appearing related to race (http://www.sciencemag.org/content/333/6045/1015).
I must admit I have the same Cheshire Cat feeling that I experienced when writing a paper on this topic on possible gender discrimination in PhD Schools - yet let me quote the authors as they highlight the problem much better than I ever could:
"If the underrepresentation of women in math-intensive fields is
not due to biased journal or grant reviews, perhaps it results from biased interviewing and hiring decisions? A study of mocksearch- committee recommendations for hiring of psychology professors (15) is often invoked for suspecting it does." (C&W p. 3160.) As the authors themselves note, it seems a rather curious practice to categorize psychology under the umbrella term STEM or alternatively try to explain a domain-specific problem with research done in another domain (where moreover females actually outnumber males, thus even if there was demonstrable discrimination, it would mean that ultimately it does not matter).
The Cheshire Cat face of underrepresentation of women in many fields of Academia seems to have vanished - with the obstinate grin of STEM related fields remaining behind. The reason for this stubborn STEM grin is still not clear to me - arguments appealing to prejudices and stereotypes do not seem very convincing, as it would need to be demonstrated how and why STEM differs from other fields surrounding the grin - it is likely that the first women to enter biology PhDs and MDs also experienced "isolation, dissatisfaction, and discrimination" (ibid.). So whence the Cheshire Cat STEM grin effect in Academia?
The argument is certainly appealing - my own personal experience would certainly support a stereotype explanation: I abandoned becoming a physicist, as I was told women were incapable of pursuing physics at an advanced level (just for the record: it's not a complaint, I'm perfectly happy with my Psychology MA). Yet this is anecdotal and furthermore I have heard the same argument physicist replaced with MD - so there must be a cog missing somewhere in the explanatory machine.
As a last note on stereotypes - thank you Christophe for drawing attention to the cognitive experiments revealing possible unconscious biases, yet I must say I have the same "appealing but unreliable" feeling about the IAT test. Although I see many improvements, still over and above the criticisms it received because of its proposed manipulability and lack of validity, my personal experience was that whenever I tried to use it - I think once it was actually a gender*science/humanities experiment - it just wouldn't work. I'm not arguing against the existence of non-conscious biasing effects or implicit prejudices, I'm merely trying to suggest to take IAT test conclusions with a grain of salt.