argument in an online debate2, and an almostidentical view was elaborated in a 2006 essayby Peter Lawrence entitled ‘Men, Women andGhosts in Science’3. Whereas Summers pref-aced his statements by saying he was trying to be provocative, Lawrence did not. WhereasSummers talked about “different availability of aptitude at the high end,” Lawrence talkedabout average aptitudes differing. Lawrenceargued that, even in a utopian world free of

bias, women would still be under-representedin science because they are innately differentfrom men.

Lawrence draws from the work of SimonBaron-Cohen4 in arguing that males are ‘onaverage’ biologically predisposed to systematize,

When I was 14 years old, I had anunusually talented maths teacher.One day after school, I excitedly

pointed him out to my mother. To my amaze-ment, she looked at him with shock and saidwith disgust: “You never told me that he wasblack”. I looked over at my teacher and, for thefirst time, realized that he was an African-American. I had somehow never noticed hisskin colour before, only his spectacular teach-ing ability. I would like to think that my par-ents’ sincere efforts to teach me prejudice wereunsuccessful. I don’t know why this lessontakes for some and not for others. But now thatI am 51, as a female-to-male transgenderedperson, I still wonder about it, particularlywhen I hear male gym teachers telling youngboys “not to be like girls” in that same deroga-tory tone.

Hypothesis testingLast year, Harvard University president LarrySummers suggested that differences in innateaptitude rather than discrimination were morelikely to be to blame for the failure of womento advance in scientific careers1. Harvard pro-fessor Steven Pinker then put forth a similar

to analyse and to be more forgetful of others,whereas females are ‘on average’ innatelydesigned to empathize, to communicate andto care for others. He further argues that menare innately better equipped to aggressivelycompete in the ‘vicious struggle to survive’ inscience. Similarly, Harvard professor HarveyMansfield states in his new book, Manliness5,that women don’t like to compete, are riskadverse, less abstract and too emotional.

I will refer to this view — that women arenot advancing because of innate inabilityrather than because of bias or other factors —as the Larry Summers Hypothesis. It is a viewthat seems to have resonated widely with male,but not female, scientists. Here, I will arguethat available scientific data do not providecredible support for the hypothesis but insteadsupport an alternative one: that women are notadvancing because of discrimination. Youmight call this the ‘Stephen Jay Gould Hypoth-esis’ (see left). I have no desire to make meninto villains (as Henry Kissinger once said,“Nobody will ever win the battle of the sexes;there’s just too much fraternizing with theenemy”). As to who the practitioners of thisbias are, I will be pointing my finger at women

Vol 442|13 July 2006

COMMENTARY

133

Does gender matter?The suggestion that women are not advancing in science because of innate inability is being takenseriously by some high-profile academics. Ben A. Barres explains what is wrong with the hypothesis.

“Few tragedies can be moreextensive than the stunting of life,

few injustices deeper than thedenial of an opportunity to strive or

even to hope, by a limit imposedfrom without, but falsely identified

as lying within.”— Stephen Jay Gould

M. G

OLD

WA

TER

/ALA

MY

13.7 Commentary NEW WF 11/7/06 11:53 AM Page 133

Nature Publishing Group ©2006

COMMENTARY NATURE|Vol 442|13 July 2006

134

as much as men. I am certain that all the pro-ponents of the Larry Summers Hypothesis arewell-meaning and fair-minded people, whoagree that treatment of individuals should bebased on merit rather than on race, gender orreligion stereotypes.

The sums don’t add upLike many women and minorities, however, I am suspicious when those who are at anadvantage proclaim that a disadvantagedgroup of people is innately less able. Histori-cally, claims that disadvantaged groups areinnately inferior have been based on junk sci-ence and intolerance6. Despite powerful socialfactors that discourage women from studyingmaths and science from a very young age7,there is little evidence that gender differencesin maths abilities exist, are innate or are evenrelevant to the lack of advancement of womenin science8. A study of nearly 20,000 mathsscores of children aged 4 to 18, for instance,found little difference between the genders(Fig. 1)9, and, despite all the social forces thathold women back from an early age, one-thirdof the winners of the elite Putnam Math Competition last year were women. Moreover,differences in maths-test results are not corre-lated with the gender divide between thosewho choose to leave science10. I will explainwhy I believe that the Larry Summers Hypoth-esis amounts to nothing more than blamingthe victim, why it is so harmful to women, andwhat can and should be done to help womenadvance in science.

If innate intellectual abilities are not toblame for women’s slow advance in sciencecareers, then what is? The foremost factor, Ibelieve, is the societal assumption that womenare innately less able than men. Many studies,summarized in Virginia Valian’s excellentbook Why So Slow?11, have demonstrated asubstantial degree of bias against women —more than is sufficient to block women’sadvancement in many professions. Here are a few examples of bias from my own life as a young woman. As an undergrad at the

Massachusetts Institute of Technology (MIT),I was the only person in a large class of nearlyall men to solve a hard maths problem, only tobe told by the professor that my boyfriendmust have solved it for me. I was not given anycredit. I am still disappointed about the presti-gious fellowship competition I later lost to amale contemporary when I was a PhD student,even though the Harvard dean who had readboth applications assured me that my applica-tion was much stronger (I had published sixhigh-impact papers whereas my male com-petitor had published only one). Shortly after Ichanged sex, a faculty member was heard tosay “Ben Barres gave a great seminar today, butthen his work is much better than his sister’s.”

Anecdotes, however, are not data, which iswhy gender-blinding studies are so important11.These studies reveal that in many selectionprocesses, the bar is unconsciously raised sohigh for women and minority candidates thatfew emerge as winners. Forinstance, one study found thatwomen applying for a researchgrant needed to be 2.5 timesmore productive than men in order to be consideredequally competent (Fig. 2)12.Even for women lucky enoughto obtain an academic job,gender biases can influence the relativeresources allocated to faculty, as Nancy Hopkinsdiscovered when she and a senior faculty com-mittee studied this problem at MIT. The datawere so convincing that MIT president CharlesVest publicly admitted that discrimination wasresponsible. For talented women, academia isall too often not a meritocracy.

In denialDespite these studies, very few men or womenare willing to admit that discrimination is a serious problem in science. How is that pos-sible? Valian suggests that we all have a strongdesire to believe that the world is fair11.

Remarkably, women are as likely as men todeny the existence of gender-based bias13.Accomplished women who manage to make itto the top may ‘pull up the ladder behind them’,perversely believing that if other women areless successful, then one’s own success seemseven greater. Another explanation is a phe-nomenon known as ‘denial of personal disad-vantage’, in which women compare theiradvancement with other women rather thanwith men11.

My own denial of the situation persisteduntil last year, when, at the age of 50, severalevents opened my eyes to the barriers thatwomen and minorities still face in academia. Inaddition to the Summers speech, the NationalInstitutes of Health (NIH) began the most pres-tigious competition they have ever run, thePioneer Award, but with a nomination processthat favoured male applicants14. To their credit,in response to concerns that 60 of 64 judges

and all 9 winners were men,the NIH has revamped theirPioneer Award selection pro-cess to make it fairer. I hopethat the Howard HughesMedical Institute (HHMI) willaddress similar problems withtheir investigator competi-tions. When it comes to bias, it

seems that the desire to believe in a meritocracyis so powerful that until a person has experi-enced sufficient career-harming bias them-selves they simply do not believe it exists.

My main purpose in writing this commen-tary is that I would like female students to feel that they will have equal opportunity intheir scientific careers. Until intolerance isaddressed, women will continue to advanceonly slowly. Of course, this feeling is alsodeeply personal to me (see ‘Personal experi-ences’). The comments of Summers, Mans-field, Pinker and Lawrence about women’slesser innate abilities are all wrongful and per-sonal attacks on my character and capabilities,

Sta

ndar

dize

d m

ath

scor

es

75

50

25

04 6 8 10 12 14 16 18

Age when tested

GirlsBoys

Figure 1 | Maths-test scores for ages 4 to 18. In theUnited States there is little to distinguish the maths-test scores of boys and girls throughout school.

“I am suspicious whenthose who are at anadvantage proclaim

that a disadvantagedgroup of people isinnately less able.”

3.0

2.8

2.6

2.4

2.2

2.00–19 20–39 40–59 60–99 >99

Scientific productivity

Com

pete

nce

scor

e

Men

Women

Figure 2 | Competence scores awarded after peerreview. Peer reviewers in Sweden award lowercompetence scores to female scientists than tosimilarly productive male scientists.

Few women, as well as men, are willing to admitthat there is discrimination in academia.

P. T

URN

LEY/

CO

RBIS

13.7 Commentary NEW WF 11/7/06 11:53 AM Page 134

Nature Publishing Group ©2006

NATURE|Vol 442|13 July 2006 COMMENTARY

135

as well as on my colleagues’ and students’ abil-ities and self esteem. I will certainly not sitaround silently and endure them.

Mansfield and others claim that women aremore emotional than men. There is absolutelyno science to support this contention. On thecontrary, it is men that commit the most vio-lent crimes in anger — for example, 25 timesmore murders than women. The only hysteriathat exceeded MIT professor Nancy Hopkins’(well-founded) outrage after Larry Summers’comments was the shockingly vicious newscoverage by male reporters and commentators.Hopkins also received hundreds of hateful andeven pornographic messages, nearly all frommen, that were all highly emotional.

Taboo or untrue?There is no scientific support, either, for thecontention that women are innately less com-petitive (although I believe powerful curiosityand the drive to create sustain most scientistsfar more than the love of competition). How-ever, many girls are discouraged from sportsfor fear of being labelled tomboys. A 2002study did find a gender gap in competitivenessin financial tournaments, but the authors sug-gested that this was due to differences in selfconfidence rather than ability15. Indeed, againand again, self confidence has been pointed toas a factor influencing why women ‘choose’ toleave science and engineering programmes.When women are repeatedly told they are less good, their self confidence falls and their ambitions dim16. This is why Valian has

concluded that simply raising expectations forwomen in science may be the single mostimportant factor in helping them make it tothe top11.

Steven Pinker has responded to critics ofthe Larry Summers Hypothesis by suggestingthat they are angry because they feel the ideathat women are innately inferior is so danger-ous that it is sinful even to think about it17.Harvard Law School professor Alan Der-showitz sympathizes so strongly with thisview that he plans to teach a course next yearcalled ‘Taboo’. At Harvard we must have veri-tas; all ideas are fair game. I completely agree.I welcome any future studies that will providea better understanding of why women andminorities are not advancing at the expectedrate in science and so many other professions.

But it is not the idea alone that has sparkedanger. Disadvantaged people are wonderingwhy privileged people are brushing the truthunder the carpet. If a famous scientist or apresident of a prestigious university is going topronounce in public that women are likely tobe innately inferior, would it be too much toask that they be aware of the relevant data? It would seem that just as the bar goes way upfor women applicants in academic selectionprocesses, it goes way down when men areevaluating the evidence for why women arenot advancing in science. That is why womenare angry. It is incumbent upon those pro-claiming gender differences in abilities to rig-orously address whether suspected differencesare real before suggesting that a whole group of

people is innately wired to fail.What happens at Harvard and other univer-

sities serves as a model for many other institu-tions, so it would be good to get it right. Toanyone who is upset at the thought that freespeech is not fully protected on universitycampuses, I would like to ask, as did third-yearHarvard Law student Tammy Pettinato: whatis the difference between a faculty membercalling their African-American students lazyand one pronouncing that women are innatelyinferior? Some have suggested that those who are angry at Larry Summers’ commentsshould simply fight words with more words(hence this essay). In my view, when facultytell their students that they are innately inferiorbased on race, religion, gender or sexual ori-entation, they are crossing a line that shouldnot be crossed — the line that divides freespeech from verbal violence — and it shouldnot be tolerated at Harvard or anywhere else.In a culture where women’s abilities are notrespected, women cannot effectively learn,advance, lead or participate in society in a fulfilling way.

Take actionAlthough I have argued that the Larry Sum-mers Hypothesis is incorrect and harmful, theacademic community is one of the most toler-ant around. But, as tolerant as academics are,we are still human beings influenced by ourculture. Comments by Summers and othershave made it clear that discrimination remainsan under-recognized problem that is far fromsolved. The progress of science increasinglydepends on the global community, but only10% of the world’s population is male and cau-casian. To paraphrase Martin Luther King, afirst-class scientific enterprise cannot be builtupon a foundation of second-class citizens. Ifwomen and minorities are to achieve their fullpotential, all of us need to be far more proac-tive. So what can be done?

First, enhance leadership diversity in acade-mic and scientific institutions. Diversity pro-vides a substantially broader point of view, withmore sensitivity and respect for different per-spectives, which is invaluable to any organiza-tion. More female leadership is vital in lessening

As a transgendered person, noone understands more deeplythan I do that there are innatedifferences between men andwomen. I suspect that mytransgendered identity wascaused by fetal exposure to highdoses of a testosterone-like drug.But there is no evidence thatsexually dimorphic brain wiring is at all relevant to the abilitiesneeded to be successful in achosen academic career. I

underwent intensive cognitivetesting before and after startingtestosterone treatment about 10years ago. This showed that myspatial abilities have increased as a consequence of takingtestosterone. Alas, it has been tono avail; I still get lost all the timewhen driving (although I am no longer willing to ask fordirections). There was one innatedifference that I was surprised tolearn is apparently under direct

control of testosterone in adults —the ability to cry easily, which Ilargely lost upon starting hormonetreatment. Likewise, male-to-female transgendered individualsgain the ability to cry more readily.By far, the main difference that Ihave noticed is that people whodon’t know I am transgenderedtreat me with much more respect:I can even complete a wholesentence without beinginterrupted by a man.

PERSONAL EXPERIENCES

Stephen Pinker, Larry Summers and Peter Lawrence (left to right) all argue that innate differences are at least partly to blame for the failure of women to advance in science.

H. B

ORD

EN; R

. FRI

EDM

AN

/CO

RBIS

; MRC

CA

MBR

IDG

E

13.7 Commentary NEW WF 11/7/06 11:53 AM Page 135

Nature Publishing Group ©2006

COMMENTARY NATURE|Vol 442|13 July 2006

136

the hostile working environment that young women scientists oftenencounter. In addition to women andunder-represented minority groups, wemust not forget Asians and lesbian, gay,bisexual and transgendered folks. Thereare enough outstanding scientific lead-ers in these racial and gender groupsthat anyone with a will to achieve adiverse leadership in their organizationcould easily attain it.

Speak outSecond, the importance of diverse fac-ulty role models cannot be overstated.There is much talk about equal oppor-tunity, but, in practice, serious atten-tion still needs to be directed at how torun fair job searches. Open searchesoften seem to be bypassed entirely fortop leadership positions, just when it matters most — search committeesshould not always be chaired by menand the committee itself should behighly diverse14,18. Implementation ofspecial hiring strategies and strong deans willing to push department chairs to recruittop women scientists are especially effective. Itis crucial in the promotion process that meritbe decided by the quality, not quantity, ofpapers published.

Women faculty, in particular, need helpfrom their institutions in balancing career andfamily responsibilities. In an increasingly com-petitive environment, women with childrenmust be able to compete for funding andthrive. Why can’t young faculty have theoption of using their tuition benefits, in whichsome universities pay part of the collegetuition fees for the children of faculty, for day care instead? Tuition benefits will be of nohelp if female scientists don’t make tenure.And institutions thathave the financialcapability, such asHHMI, could help bymaking more career-transition fellowshipsavailable for talentedwomen scientists.

Third, there should be less silence in the faceof discrimination. Academic leadership has aparticular responsibility to speak out, but weall share this responsibility. It takes minimaleffort to send a brief message to the relevantauthority when you note a lack of diversity inan organization or an act of discrimination. Idon’t know why more women don’t speak outabout sexism at their institutions, but I doknow that they are often reluctant, even whenthey have the security of a tenured facultyposition. Nancy Hopkins is an admirable rolemodel, and it is time that others share the burden. It doesn’t only have to be women thatsupport women. I was deeply touched by theeloquent words of Greg Petsko19 followingSummers’ comments. And it has been 30 years

since I was a medical student, but I still recallwith gratitude the young male student whoimmediately complained to a professor whohad shown a slide of a nude pin-up in hisanatomy lecture.

Fourth, enhance fairness in competitiveselection processes. Because of evaluation bias,women and minorities are at a profound dis-advantage in such competitive selection unlessthe processes are properly designed11,12,14,18. As the revamped NIH Pioneer Award demon-strates, a few small changes can make a significant difference in outcome. By simplychanging the procedure so that anyone canself-nominate and by ensuring a highly diverseselection committee, the number of womenand minority winners went up to more than

50% from zero. Thislesson can and shouldnow be applied toother similar processesfor scientific awards,grants and faculty posi-tions. Alas, too manyselection committees

still show a striking lack of diversity — withtypically greater than 90% white males. Whenselection processes are run fairly, reverse discrimination is not needed to attain a fairoutcome.

Confidence boosterFinally, we can teach young scientists how tosurvive in a prejudiced world. Self-confidenceis crucial in advancing and enjoying a researchcareer. From an early age, girls receive mes-sages that they are not good enough to do sci-ence subjects or will be less liked if they aregood at them. The messages come from manysources, including parents, friends, fellow stu-dents and, alas, teachers. When teachers havelower expectations of them, students do less

well. But we are all at fault for send-ing these messages and for remain-ing silent when we encounter them.Teachers need to provide muchmore encouragement to young people, regardless of sex, at all stages of training.Occasional wordsof encouragement can have enor-mous effects.

All students, male and female,would benefit from training in how tobe more skillful presenters, to exert apresence at meetings by asking ques-tions, to make connections with fac-ulty members who may help them toobtain grants and a job, and to havethe leadership skills necessary to survive and advance in academia.Because women and minorities tendto be less confident in these areas,their mentors in particular need toencourage them to be more proactive.I vividly recall my PhD supervisorcoming with me to the talks offamous scientists and forcing me to

introduce myself and to ask them questions.There is a great deal of hallway mentoring thatgoes on for young men that I am not sure manywomen and minorities receive (I wish thatsomeone had mentioned to me when I wasyounger that life, even in science, is a popularitycontest — a message that Larry Summers mighthave fo und helpful as well). It is incumbent onall of us who are senior faculty to keep a look outfor highly talented young people, includingwomen and minority students, and help themin whatever way possible with their careers. ■

Ben A. Barres is at Stanford University School ofMedicine, Department of Neurobiology, FairchildBuilding Room D235, 299 Campus Drive,Stanford, CA 94305-5125, USA.e-mail: barres@stanford.edu

1. Summers, L. Letter to the Faculty Regarding NBER Remarkswww.president.harvard.edu/speeches/summers/2005/facletter.html (2005).

2. The Science of Gender and Science. Pinker vs. Spelke: A Debatewww.edge.org/3rd_culture/debate05/debate05_index.html (2005).

3. Lawrence, P. A. PLoS Biol. 4, 13–15 (2006).4. Baron-Cohen, S. The Essential Difference: Men, Women, and

the Extreme Male Brain (Allen Lane, London, 2003).5. Mansfield, H. Manliness (Yale Univ. Press, New Haven,

2006). 6. Gould, S. J. The Mismeasure of Man (W. W. Norton & Co,

New York, 1996).7. Steele, C. M. Am. Psychol. 52, 613–629 (1997).8. Spelke, E. S. Am. Psychol. 60, 950–958 (2005).9. Leahey, E. & Guo, G. Soc. Forces 80.2, 713–732 (2001).10. Xie, Y. & Shauman, K. Women in Science: Career Processes

and Outcomes (Harvard Univ. Press, Cambridge, 2003). 11. Valian, V. Why So Slow? (MIT Press, Cambridge, 1998).12. Wennerås, C. & Wold, A. Nature 387, 341–343 (1997).13. Rhode, D. L. Speaking of Sex: The Denial of Gender Inequality

(Harvard Univ. Press, Cambridge, 1997). 14. Carnes, M. et al. J. Womens Health 14, 684–691 (2005).15. Gneezy, U., Niederle, M. & Rustichini, A. Q. J. Econ. 18,

1049–1074 (2003).16. Fels, A. Necessary Dreams (Pantheon Press, New York,

2004).17. Pinker, S. New Repub.15 (14 Feb, 2005). 18. Moody, J. Faculty Diversity: Problems and Solutions (Taylor

and Francis, New York, 2004).19. Petsko, G. A. Genome Biol. 6, 1–3 (2005).

“Simply raising expectations forwomen in science may be the single

most important factor in helpingthem make it to the top.”

— Virginia Valian

At school, girls and boys show similar levels of ability in the sciences.

J. W

EST/

ALA

MY

13.7 Commentary NEW WF 11/7/06 11:53 AM Page 136

Nature Publishing Group ©2006