Academic Responsibility and Gender Bias
On January 14, the President of Harvard, Lawrence Summers, stopped in at a conference held by the National Bureau of Economic Research (NBER) to deliver a luncheon speech to about 40 experts and scholars gathered to discuss the under-representation of women and minorities in the Science and Engineering workforce in the U.S. A stated purpose of the meeting was to address the question: "What programs and policies can further the process of diversifying the science workforce?" Earlier in the day I had spoken about MIT's efforts to address this question.
Several months before this meeting, the media had reported that during Summers' leadership there has been a sharp decline in tenured faculty offers to women in Harvard's Faculty of Arts and Sciences (FAS) in all fields of the humanities and social sciences, as well as in science and engineering. Given the topic of the NBER meeting I expected to hear Summers' thoughts on this recent decline, or possibly innovative approaches to solving well-documented problems that drive women from science and engineering careers.
Instead, in what have become widely known comments, Summers offered three hypotheses to explain the absence of women from the top of many professions, including specifically academic science, math, and engineering. As discussed below, much of what he said flies in the face of decades of research. Furthermore, research shows that the unfounded attitudes Summers expressed can lower women's expectations in the classroom and the workplace, and negatively impact judgments about their performance and merit.
I walked out of Summers' talk in personal protest.
As a molecular biologist and geneticist for over 40 years who has spent 10 years studying the under-representation of women in the high end of the science professoriate, I knew that many of Summers' remarks were factually wrong, contradicted by decades of research, and deeply damaging to women. As a Harvard alumna (BA and PhD) I was personally offended. Later that day I happened to receive an e-mail from a Boston Globe reporter about an unrelated matter. At the bottom she queried: "PS: Did today's conference turn out to be anything interesting?" I told her about Summers' talk. She tracked down other NBER participants and gathered their responses to the speech. On January 17th the Globe published her story. The news sped round the world and has occupied front and editorial pages for more than eight weeks. On February 17, Summers published a transcript of his remarks (http://www.president.harvard.edu/speeches/2005/nber.html) with a cover letter (http://www.president.harvard.edu/speeches/2005/facletter.html) modifying and apologizing for the comments. While gracious, the letter did not explain why the comments were both wrong and damaging. It also failed to allay unwarranted public concerns about academic freedom.
Because of the media frenzy and public misunderstandings about this incident, the Editorial Sub-Committee asked me to write a piece for the MIT Faculty Newsletter. I take this opportunity to describe some of the background and research that may have caused Summers to apologize for his comments, and to explain why this incident is not about "political correctness" or "academic freedom."
Why I was a speaker at the NBER conference
I have been a professor of molecular biology and genetics at MIT for 31 years. For 17 years my lab studied genes involved in cancer. For the past 14 we have studied genes required for vertebrate development. I am a member of the National Academy of Sciences, Institute of Medicine, and American Academy of Arts and Sciences.
I entered science convinced that civil rights laws had eliminated gender discrimination from the work place, including from academia. However, I learned I was wrong. I gradually came to realize that women and men who made equally important scientific discoveries often were not valued equally by our system. In 1994, as described in an article in this newsletter in 1999 ( http://web.mit.edu/fnl/women/women.html), I discovered that other tenured women faculty in the School of Science at MIT had come to the same realization that I had. Then Dean of Science Bob Birgeneau (now the chancellor at Berkeley) established a committee to study the problem systematically. The committee, which I chaired, consisted of six tenured female faculty and three tenured male professors of science. We interviewed tenured and untenured women faculty in Science and gathered data on lab space, committee membership, salaries, etc. The resulting 150-page report presented vivid evidence of how women professors enter science believing that gender discrimination is a thing of the past, but, as they approach the age of the men in power, suffer marginalization and day-in- and day-out biases. These small inequities, as they accumulate, make doing science and attaining top positions much more difficult for women.
The committee report was detailed and personal, and hence needed to remain confidential, but in 1999, then chair of the faculty Professor Lotte Bailyn, asked our committee to provide a narrative report of what we had done and found, and how the Institute had responded. This summary, which was published in a special edition of the MIT Faculty Newsletter [Vol. XI No. 4, March 1999], came to be known as the MIT Report on Women in Science. A sentence that President Vest wrote in comment proved to be of particular importance: "I have always believed that contemporary gender discrimination within universities is part reality and part perception. True, but I now understand that reality is by far the greater part of the balance." The Report was covered widely in the press and elicited an outpouring of e-mail from women scientists in the U.S. and abroad noting that they too had experienced discrimination but had been unable to get their administrators even to acknowledge, much less to address, the problem.
Following publication of the MIT Report, and with support from the Ford Foundation, President Vest convened a meeting of presidents, provosts, deans, and leading female scientists and engineers from eight other leading research universities. They met in January 2001, returned to their campuses to study the problem, and met again in the spring of 2004.
Most universities analyzing the problem found results very similar to those at MIT, although five years after the MIT Report was made public, easily measurable inequities in compensations or resource allocations to women have become relatively uncommon.
MIT also established a Council on Faculty Diversity, which Provost Bob Brown and I co-chair. Its goal is to prevent and eliminate inequities that arise from bias, to increase the flexibility in academic careers to help both male and female faculty better manage family and work goals, to attract more minority students to PhD programs in science and engineering, and to ensure that search committees seek to identify outstanding women and minority candidates. Yet, despite real progress at MIT and many other universities, marginalization and unintended gender bias remain a serious problem.
A further consequence of the publication of the MIT Report was that I have received nearly 500 invitations to talk on this topic at universities, research institutes, and conferences. I have accepted 90. One was the NBER meeting.
A landmark report on women scientists in Europe and research in psychology have documented the impact of unintentional gender bias
Tenured women faculty at MIT were not alone in identifying gender bias as a barrier to women trying to advance in science careers. In 1997, a landmark study by Wenneras and Wold opened Europe's eyes to the magnitude of gender discrimination there (Nature 397, 341-343 (1997). Wenneras and Wold asked why a much smaller fraction of female than male applicants obtained postdoctoral fellowships in biomedical research from the Swedish Medical Research Council. They re-examined the merit of applicants using quantitative criteria of accomplishment. They examined applicants' publication records using impact factors of the journals the applicants had published in, and the number of publications and first author publications of each applicant. (They used legal means to obtain the data since the granting agency refused their request to provide it.) Wenneras and Wold's stunning finding was that a woman had to be 2.5 times more productive than a man to receive an equal score in scientific competence. This meant, for example, that she had to have published the equivalent of three more Nature or Science papers to be judged equal to a man.
What is so striking about the Swedish study is that many people assume that women often receive special advantages.
While sometimes true, data show that in highly significant ways, throughout their careers, the opposite is more often true. How could this be? Importantly, while these studies were in progress, another line of research was advancing which provided an intellectual basis for the surprising findings of reports such as those from MIT and Sweden. This research comes from the field of psychology. Much of the relevant work has been summarized in Virginia Valian's book, Why So Slow.
Over recent decades, psychologists have demonstrated that identical intellectual work performed by men and women is frequently not valued equally. For example, if one xeroxes a manuscript and puts a man's name on one copy and a woman's on the other and sends the two articles out for review, the identical work receives a higher score or more positive comments if reviewers think it was authored by a man. Strikingly, it does not matter whether the reviewers are men or women! Most of us - women and men alike – tend to under-value work if we think it was performed by a woman. It does not take research or imagination to understand the devastating impact on professional women, of having their work judged inferior when in truth it is at least equal in merit . The inability of most people to believe they are capable of such unconscious bias and unfair judgment is perhaps the greatest obstacle to equality in the workplace.
But there is some good news: Research shows that unconscious bias can be alleviated by societal changes and by recognition of the problem. Judgments of women's work have of late become more equitable, presumably due to society's acceptance of women's competence in the workplace. In addition, work by Harvard psychologist Professor Mahzarin Banaji and others has shown that making people aware of their unconscious bias can reduce its magnitude.
Recently, NSF established a program called ADVANCE to address bias and other well-documented barriers to women's careers in science and engineering. The program has awarded $50 million in grants over the past three years to support efforts at 19 universities aimed at institutional changes that might level the playing field for women, and also for minorities, who face some of the same barriers as women in entering science and engineering fields. Women (including minority women) are 50% of the U.S. workforce and minority males are an additional 15% of the population. In failing to draw on this pool, we squander two-thirds of the available science talent in America. Encouraging women and minorities to enter science and engineering is deemed critical by the NSF to future U.S. competitiveness.
President Summers' NBER remarks
It was in the above context that Summers made his remarks to the NBER audience. He was listed on the program as "Lawrence Summers, President, Harvard University." Summers' remarks were of particular importance because the views and leadership of a President are critical to making significant progress in this endeavor. Furthermore, women faculty at Harvard's FAS had expressed concern last year that Summers may not understand these issues. I had already discussed the poor record in hiring female molecular biologists in the FAS with Summers in the fall of 2004.
Summers told the NBER audience that he would offer three hypotheses to explain the under-representation of women in tenured positions on the faculties of leading universities, particularly in the fields of science, math, and engineering. As set forth in the transcript of his remarks, the three, in order of importance, were: 1. Women's family responsibilities and unwillingness to work the 80 hour week it takes to get to the top;
Below I explain what is factually wrong and what is right with Summers' three hypotheses, according to current research. To summarize, research has identified two important, and even related, barriers to women striving to reach the top in science and engineering: greater family responsibilities, as Summers put forward in hypothesis #1, and socialization and gender bias, which Summers incorrectly dismissed. In contrast, and despite decades of research, intrinsic differences in aptitude that could explain the small numbers of women professors of science and engineering have not been found.
Hypothesis 1: As Summers proposed, the greater family responsibilities that have traditionally fallen to women vs. men have been a significant factor in causing women to opt out of high-powered careers in science and engineering. However, some women have always been able to "have or do it all." Furthermore, significant changes in recent decades have improved the situation for many more women. New expectations, goals, economic realities, reproductive technologies, and the length and variability of careers have changed the picture. Today young professional women and men demand more equal involvement in both family and work, and universities compete to accommodate them. Still, I personally believe that there is not yet enough change in universities or in society to create equality for men and women in this regard, so I believe that Summers' hypothesis #1 has merit. However, I found his implication that women are not willing to work hard enough to get to the top both insulting and wrong.
Hypothesis 2: "Intrinsic aptitude" differences means genetic differences. Not one shred of scientific evidence supports the assertion that women as a group are genetically inferior to men in the math, science, or engineering ability required to attain the top of the profession. Nor is this from a lack of research effort to find such differences. Of course there are myriad biological differences between men and women, including in brain and other anatomy, hormones, etc. However, none of these has been linked to any intellectual ability which itself has a causal link to achievement at the highest levels in any intellectual pursuit.
In sharp contrast, decades of research demonstrate that socialization and bias are powerful factors in determining career choice and academic success, including differences between men and women.
To explain the decades of research on this topic in biology and psychology is beyond the scope of this article. However, several key findings that argue against significant genetic differences in intellectual ability of men and women, as well as flaws in Summers' argument and facts are below. (See also Nature Neuroscience 8 , 253 (2005) for an excellent brief summary.)
1. Women's demonstrated ability to perform at a very high level in science and math has changed too rapidly to be explicable by genetic changes. For example, the figures show the percent of women undergraduates and graduate students at MIT each year for the past century. The percent of women PhDs in science and engineering nationally has increased similarly.
2. Summers asserted that research universities only hire exceptional people, and used as the measure of being exceptional, math SAT scores. He said that there are more men out on the tail of the bell curve of math ability than women, that this may reflect differences in intrinsic aptitude between men and women, and thus may explain the under-representation of women on science, math, and engineering faculties of elite universities. However, the tests he referred to are SAT tests taken by teenagers. By this age it is not possible to differentiate innate ability from the impact of society and its expectations on students' choices or performance. In contrast to tests of teenagers, tests of very young children have failed to detect significant differences between the cognitive abilities of males and females. Minor differences have been found, for example, in spatial rotation (favoring boys) and verbal skills (favoring girls) but even in these cases, there are no correlations of these skills to later career choices or success, and the differences detected are tiny compared to the impact of social factors.
3. Women near the high end of the SAT bell curve take the same math classes as men in college and do equally well in them. This was not true several decades ago, but has come about as women began to receive similar encouragement in school. Such a rapid change in women's math ability again argues against genetic changes and in favor of societal expectations and how boys and girls are educated. These changes also make long-term studies of high SAT scorers, which have sometimes been used to try to bolster Summers arguments, problematic.
4. Some women have extreme math ability as traditionally measured. For example, there are female winners of the Putnam award now, although for many years there were none. This is an award made to the top five college students in math in the U.S. each year. Strikingly, this year 4 of the top 15 students in the competition were women. This award is so elite that even on the most elite math faculties only about 1 in 20 to 1 in 40 professors was a Putnam winner. Furthermore, while Putnam winners as a group are enormously successful academically, not all (or even most?) become academic math superstars. Nonetheless, given that some women can perform at this level, and thus must always have been capable of it, why has there never been a tenured woman professor in math at Harvard in its 369-year history (and only three ever at MIT)? I am told that three women have been offered tenured positions in math at Harvard in recent years. I spoke to one who told me she would not accept the job at Harvard or "in the math departments of any of those other elite east coast schools either" because of the gender discrimination. She said that in these departments faculty think that only men can be great mathematicians and she was not willing to work in such a biased environment. While anecdotal, such testimonies from women with this ability are important, and they are common among this elite group.
But could there be genetic differences between men and women which we might discover in the future and which might affect the ability or inclination of men vs. women to pursue certain fields? Yes. After socialization and bias are corrected for, perhaps girls will prove to be better at math than boys. But we can't use these unfounded possibilities to explain the lop-sided under-representation of women (or minorities) on current science, math, and engineering faculties. Worse, it is actively irresponsible to do so, because evidence from research in psychology demonstrates that such attitudes negatively impact performance and judgments of merit of those who belong to the group that is seen as genetically inferior (see, for example the work of Claude Steele). It has proven difficult to dissociate exceptional individuals from their group stereotype.
Although there are obvious biological differences between the sexes, including anatomical brain differences, making the leap from gross anatomical findings to explanations of subtle cognitive functions, let alone to complex career choices and outcomes is impossible (except perhaps where disease or damage to the brain is involved.) Such simplistic, unfounded leaps are junk science. Genetics has been one of the most powerful tools in modern biology. But so far, genetic analysis of even simple behaviors in animals has been an almost complete failure. It will be extremely hard, if even possible, to identify math ability genes, if they exist. Currently, to speak seriously of using genetics to justify negative societal outcomes in intellectual pursuits for groups of people, amounts to academic folly.
Hypothesis 3: As discussed above, socialization and bias, including, importantly, unintentional bias, are powerful factors that shape and can limit the careers of men and women. While we have seen extraordinary changes in the status of women and under-represented minorities in the past several decades, research unequivocally documents that full equality of opportunity has not yet been attained for these and other groups at the highest levels of our society.
This incident is not about academic freedom
While initial reports and commentary focused on Summers' remarks, a second wave of press tried to make the critics of Summers' speech the focus of this story. These individuals were accused of suppressing academic freedom. But this affair is not about academic freedom at all. It's about the role and responsibilities of a University President, and secondarily about responsibility in research.
A university needs a president and it needs one 24/7. Summers speaks as the voice of Harvard every time he utters a public word. So, is it acceptable behavior for the President of Harvard to make undocumented, incorrect comments outside his field when such comments damage and demean half his constituency? Of course he has the freedom to do so. The question is whether he should .
The second purported issue of academic freedom here is whether experts must stand up and argue with the President in a public forum in order to educate him about a field in which he does not work. If they fail to do so, or disagree afterwards, have they infringed on the President's academic freedom? The answer is no, precisely because he is President. Were Summers merely a professor who wanted to advance the genetic inferiority of women to do math and science at elite universities, the processes that accompany academic freedom and enable the academy ultimately to discover the truth, would kick in. The speaker would be required to present data, others would present contradicting data, replication of experiments would ensue: ideas would be tested and those found wanting would ultimately be weeded out. But the President is not a researcher in the field, and the processes associated with academic freedom of faculty to do research do no t pertain when he speaks. Summers is the boss. His words are the pronouncements and opinions of Harvard University. At no time during his talk did I doubt that these were the beliefs of the President of Harvard. Given my knowledge of these fields, I felt a moral obligation to protest and a desire to leave. I asked the women sitting to either side of me if they thought we should all walk out, but they were the speakers following Summers and so, although one said she would like to, they felt they could not leave.
My critics in the press
I have been criticized in the national press for my role in bringing Summers' comments to light. My critics seem to have a diversity of motives. The identity of some of the critics could have been predicted; for example, the ultra-conservative Independent Women's Forum and associated voices such as those of Judith Kleinfeld, Phyliss Shafly, even Cathy Young. Some of these women have long attempted to discredit the MIT Report, apparently for political or ideological reasons, and some have used deliberately false information to do so. So their reactions were not terribly surprising.
Harder to understand, at first, are certain Harvard faculty critics, particularly our former colleague Steven Pinker, who has portrayed me in the press as being opposed to academic debate and inquiry (The New Republic , February 14, pg 15, 2005) and see http://www.thecrimson.com/article.aspx?ref=505366. Why would Steve imply such an obvious untruth? Some Harvard faculty told me that Pinker and his popular-press book The Blank Slate were the source for Summers' NBER comments. Having now read the poorly reasoned and unsupported section on gender in Pinker's book, this seems likely. If so, Pinker's defense of Summers makes sense. In fact I think he owes Summers an apology. But he owes me one as well. Ironically, Harvard psychologist Professor Elizabeth Spelke and I were scheduled to debate with Pinker on the Charlie Rose show about research in biology, genetics, and psychology that debunks Pinker's views, but Pinker backed out. The show was cancelled because, they told me, they could not find a psychologist to take Pinker's viewpoint who was willing to appear.
But my most visible and also semi-mainstream critic was columnist George Will, who labeled me a "hysteric," and did so before the text of Summers' remarks had even been released (http://www.washingtonpost.com/wp-dyn/articles/A40073-2005Jan26.html). It was a classic male response in many ways, since women who complain of gender discrimination have long been told they are "difficult women," even crazy. Will labeling me the old-fashioned female hysteric is how some men have tried to gag women forever. I think the explanation for his remarks is summed up well in a letter written by a female reader to Will and cc'd to me:
"Do you have some kind of personal grudge against Ms Hopkins, Mr. Will? Or is it just women and their progressiveness in general?. So Ms Hopkins got upset. So what? Many women get emotional. Know why? .When people don't listen its frustrating. By the way, would you have written your piece had you been speaking of a Jew getting upset about the Holocaust, or an African American reeling over some injustice? I doubt it. But we ALL know that women are fair game, right??? THINK, Mr. Will. THINK. You (and many other controlling, power-hungry males) would find any way, any excuse to keep women down and from getting ahead - even when it's in the best interest of humankind and the world."
Hundreds of women who learned of the NBER incident from press reports wrote me to say, "Thank you, thank you for your courage," while only about two dozen women criticized my actions or words.
In contrast, I received fewer than 30 e-mails from men supporting my actions and a couple of hundred that echo Will: some say they can't take me seriously because I'm a weak whining woman; some say they can't take me seriously because I didn't stay to argue with Summers - as if centuries of women had not argued with such men for long enough. These men do not understand that leaving was my reply, or why it was important to tell the exact truth when the Globe reporter asked me how this speech had made me feel: how else could I guarantee that those who do not understand gender bias would comprehend the impact it has on those who usually have to bear it in silence? But if everyone understood gender bias, this whole affair would never have happened. George Will shows us how far women still have to go to achieve equality - and he seems worried enough to write about it. Like Summers' NBER comments, he's the problem.
So the Summers flap is not about some obvious hot-button items: suppression of academic freedom, quotas, or lowering academic standards by admitting women to the top who don't belong there (which should never be permissible at the faculty level in universities). So what is it about? It's certainly about trying to make people understand that unintentional gender bias is real, that it keeps women from the top, and that it can be remedied. But still, given that so many people could care less about gender bias, why has this story held the interest of the press and public for over two months?
I think the fascination with this story is that we may be witnessing a skirmish in the final battle - a battle to get to the top, the last step in a process that has gone on for millennia. As Summers so accurately noted, women aren't at the top of any powerful profession in the U.S. It's not just the science or engineering faculties of elite universities where they are under-represented, but at the top of business, politics, law, the media - where the real power resides. And no one could believe that SAT scores explain why there has never been a female president of the United States!
Men hold at least 95% of the institutional power in America, and it's not easy to give that up. What many of us are waiting to see in this symbolic struggle is whether women are finally going to achieve equality or not. And I think it's not only men who fear such an outcome. I think many women do, too. If men and women shared power equally the world would be a different place. I for one think it would be better than anything we can imagine. But none of us knows for sure what it would look like. Given this uncertainty, it's easier to see why this story has tapped into so much fear, and provoked such anger and hostility - and among some of us, so much hope.