<html><body>American Enterprise Institute Monday, October 1, 2007 [Edited transcript from audio tapes]   <TABLE cellSpacing=1 cellPadding=1 width="100%" border=0> <TBODY> <TR> <TD> <DIV class=BodyText>9:15 a.m.</DIV></TD> <TD> <DIV class=BodyText>Registration</DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText>9:30  </DIV></TD> <TD> <DIV class=BodyText>Introduction: </DIV></TD> <TD> <DIV class=BodyText>Christopher DeMuth, AEI</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText>9:40  </DIV></TD> <TD> <DIV class=BodyText>Panel I:</DIV></TD> <TD> <DIV class=BodyText>The Science on Women and Science: What the Data Say</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText>Moderator:</DIV></TD> <TD> <DIV class=BodyText>Christina Hoff Sommers, AEI</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText>Panelists:</DIV></TD> <TD> <DIV class=BodyText>Rosalind Chaitt Barnett, Brandeis University </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText>David Geary, University of Missouri </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText>Richard Haier, University of California Irvine Medical School</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText>Elizabeth Spelke, Harvard University</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText>11:50</DIV></TD> <TD> <DIV class=BodyText>Break</DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText>Noon </DIV></TD> <TD> <DIV class=BodyText>Luncheon</DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText>Keynote Speaker:</DIV></TD> <TD> <DIV class=BodyText>Simon Baron-Cohen, Cambridge University</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText>1:30 p.m.</DIV></TD> <TD> <DIV class=BodyText> Break</DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText>1:40</DIV></TD> <TD> <DIV class=BodyText>Panel II: </DIV></TD> <TD> <DIV class=BodyText>Stereotype Threat: The State of the Research</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText>Moderator:</DIV></TD> <TD> <DIV class=BodyText>Christina Hoff Sommers, AEI</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText>Panelists:</DIV></TD> <TD> <DIV class=BodyText>Joshua Aronson, New York University</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText>Amy Wax, University of Pennsylvania Law School</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText>3:00  </DIV></TD> <TD> <DIV class=BodyText>Speaker:</DIV></TD> <TD> <DIV class=BodyText>Charles Murray, AEI</DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText>3:40  </DIV></TD> <TD> <DIV class=BodyText> Adjournment </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR> <TR> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD> <TD> <DIV class=BodyText> </DIV></TD></TR></TBODY></TABLE> Proceedings: PANEL I:  The Science on Women and Science:  What the Data Say Christopher DeMuth:  Good morning and welcome.  My name is Chris DeMuth, I m President of the American Enterprise Institute and I m delighted to welcome you all here today for this conference on Women and Science.  This is a conference of AEI s Brady Program in culture and freedom, and I d like to take the occasion to thank the W.H. Brady Foundation, and, particularly, Elizabeth Brady Lurie for their wonderful gift that made it possible for us to establish this AEI project, and many conferences such as the one we are holding today.  The subject of women and science and, in particular, the relatively low rates of participation of women in the quantitative sciences and engineering fields has been a subject of tremendous amount of attention and vigorous controversy in recent years.  Some of these controversies have been a little esoteric for us plain hum-drum Washington politicos, but women in science is now also the subject of a great deal of federal grant making, and when something becomes a subject of grant making, that is something that we in Washington can really understand and get our arms around, so we are now paying attention also.  Today we will be focusing not so much on the politics of women in science, but to the science of women in science, and I m very grateful to my colleague, Christina Hoff Sommers, who has conceived and organized this session from the beginning, and who will be our host and moderator throughout the day.  She and I are also grateful to our colleagues, Charles Murray and Jurgen Reinhoudt for helping with many aspects of conference planning over the past several months.  But, most of all, I m immensely gratified that this conference has attracted so many exceptionally accomplished individuals from academic research and the professional fields who have been paying attention to these subjects and are with us today, both as members of the panels, and of guests in the audience who I know will be full participants in our discussions.  We have a rich and full agenda, and I know I am certainly looking forward to a vigorous and productive series of discussions.  So I would like to turn things over now for a few initial words, and to moderate our first session on The Science on Women and Science:  What the Data Say, to Christina Hoff Sommers.  Christina?  [Applause].  Christina Hoff Sommers:  Thank you, Chris.  Good morning and welcome to the American Enterprise Institute.  I have the agreeable privilege of moderating this intriguing conference.  In the Fall of 2006, the National Academy of Sciences, in cooperation with two other very prestigious organizations, professional organizations, they released a report entitled, Beyond Bias and Barriers, Fulfilling the Potential of Women in Academic Science and Engineering.  The report claimed to find widespread bias against women in every field of science and engineering.  The authors included Donna Shalala, the former Secretary of Health of HHS, and now President of Miami University.  She, along with several others, concluded that it was not lack of talent, but unintentional and outmoded institutional structures that are hindering the access and advance of women.  The time for action is now.   Now, the literature on sex differences is very complex, sometimes contradictory, there are all sorts of disagreements, schools of thought, feuds, researchers don t seem to agree on many things.  For example, women are very successful in getting PhDs in certain fields, and less in others.  Sixty-five percent of the PhDs in education go to women, 54 percent of degrees in social science, 47 percent in the life sciences, then it goes down to 26 percent for physical sciences and 17 percent for engineering.  Interestingly enough, in the 1960s, women were about 5 percent of veterinaries, and today they re approaching 80 percent of enrollments in our schools of veterinary medicine.  I heard one disgruntled farmer complaining that now that women had taken over the field there was a bias in favor of cats.  [Laughter] He had hogs and cattle and he couldn t find anyone to take, obviously prejudice.  But, anyway, maybe someone today can explain to us why this is happening.  Fortunately, for us, we have a group of fantastic scholars here today.  Some are sympathetic to the NAS report, others are not, and they re going to be able to shed light on this debate.  The NAS report calls for workshops for academic personnel to help them overcome hidden bias, it also calls for federal agencies to implement stringent Title IX compliance reviews of math, science and engineering programs.  One federal agency, the National Science Foundation, is already in full swing with initiatives designed to bring justice to women in academic science.  Now what I set out to do several months ago was to assemble the best researchers I could possibly find in the area of sex differences, and I think I have been wonderfully successful.  Each of our speakers is going to hold forth for about 15-20 minutes this morning, on this morning s panel.  I m going to be strict about time limits because I want to allow plenty of time for audience members to ask questions and make statements.  Our speakers have just far too many awards and degrees and books to list them, you ll find them in your programs, so I ll just say very generally that our first presenter will be Professor Rosalind Barnett, from Brandeis University.  She s a senior scientist at the Women s Study Center at Brandeis, my alma mater, and welcome to AEI.  We also have Professor Elizabeth Spelke.  She is the Marshall Berkman Professor of Psychology at Harvard University.  She s a star in the field of cognitive psychology.  She was profiled in The New Yorker.  Very excited to hear her speak.  We will hear from Richard Haier, a Professor of Psychology and Pediatric Neurology at the University of California in Irvine.  And our final speaker will be David Geary.  He s the Curator Professor at the University of Missouri s Department of Psychology, the Department of Psychological Sciences, and he has chaired that department for many years.  He specializes in cognitive developmental psychology.  So I think Professor Barnett, we ll begin.  Rosalind Chait Barnett:  Thank you very much for inviting me.  The title of my talk, which you ll find out in a minute, is From Stone Walls to Invisible Walls.  Briefly, the question of where are women in science is two parts, of the women in science where are they, and why aren t there more women in science, which I ll touch on briefly, then I ll share with you my particular focus.  But, first, I m going to take a few minutes to talk about what I think are three very over-hyped stories that get too much attention, and then I m going to take you on a long journey over time from the Renaissance to the present to put this whole question in a broader framework, and then I have some concluding comments.  My particular focus will be on three subjects, women s access to education in science, women s opportunities for employment in science, and women s appropriate recognition for their scientific work.  However, as I mentioned, I m going to take a few minutes to touch on these three over-hyped stories that suck up all the oxygen from what I think is the real story of women in science.  The first, women innately don t have what it takes to succeed in math.  This argument is often made based on the following considerations -- that women have almost never made it into the ranks of most accomplished mathematicians and scientists, and women s brains, hormones, and motivation are deficient.  These arguments continue to be made in spite of the fact that numerous peer review studies fail to find any evidence of large gender differences in math and science ability, including a major med analysis of four million students, which revealed essentially no gender differences in math aptitude.  Here you can see this on the slide.  I don t know if this has a laser pointer.  Oops, it does not, too bad.  The point being that the big story here is the enormous within gender variability, which overwhelms the mean difference, which shows a slight advantage for males, but the big story is down here, and that within gender variability is lost in the war of words about gender differences.  Unless you think this is a dead story, Kathleen Parker did a piece just two weeks ago in The Washington Post Writer s Group, repeating the same tired truths without any supporting data.  Second over-hyped story, in my view, males are over-represented at the upper tail of distribution in math aptitude scores, and, therefore, are more highly represented in leadership positions in math and science.  However, studies show only a weak relationship between scoring and the upper tail ability and eventual success in math and science careers.  In fact, of the college educated professional workforce in math, science, and engineering, fewer than one-third of the men had SATM scores above 650, which is the lower end of the threshold typically presumed to be required for success in these fields.  And in one long-term study, after taking into account relevant experiential and preferential variables, sex accounted for only 1 percent of the variants in science and math career outcomes among students who scored at the extreme tail of the math aptitude test.  Those scores in the top of the math aptitude test alone do not, by themselves, tell the whole story of success in math and science careers.  The third story is academic success is zero sum game, when girls succeed, boys fail, and vice-versa.  These two graphs are going to show you, really suggest that s not the case.  Here is a graph of a male and female s aptitude score, age 13, on the NAEP, which is the Nation s Report Card.  This is a sample of 172,000 fourth graders, going from 1973 to 2004.  As you can see, the steady increase of both boys and girls in math scores, the gap between them is about stable, has not been much changed, boys and girls are both getting better.  In the next slide, you see the same thing for, oops, where did it go, so these are two, very much the same kind of thing.  Now I m going to go back to my own focus, the long-term picture of institutional, cultural and organizational factors that have made it appear that science is a male preserve, and have made it extraordinarily difficult for women to assume leadership roles, or even to be visible in math and science.  To develop my ideas, I m going back in time to the Renaissance, and then I m going to jump ahead, finally touching on data from the 19th Century to the present.  As you will see, much has changed, especially with respect to women s access to education and employment in the sciences, yet, regrettably, their advances are not often recognized or rewarded, continuing to fuel the belief that science and math are still male preserves.  I believe this journey through time is pertinent to our understanding of the question of this conference, namely, where are the women in science?  As I hope to show, this question has long been asked, and the rationalizations for the situation of women today are strikingly similar to those offered in the past.  Perhaps this historical glimpse will help us move past the simple reasons put forth today and lead us to a better and fuller understanding of why we re asking this question in the way we are asking it.  Let s go back -- so many major scientific and mathematical discoveries were made during this particular period, that is widely viewed as the foundation of modern science.  The first thing to notice is that this period was marked by an amazing abundance of male superstars in science.  And, just briefly, Copernicus, Galileo, Fairmount, Kepler, Descartes, Newton, Leibniz and da Vinci.  Question -- during this extraordinary period, what were the women doing?  Ironically, as they watched the lives and rites of their husbands, sons and brothers expand, their lives contracted.  During the height of the Renaissance, when science was flourishing, respectable women had only two life options, they either entered into a marriage, often arranged, or they went into a convent.  Many women, wealthy or not, chose the convent.  Why?  Because a fear of dying in childbirth.  There was a very high rate of women s death in childbirth.  To put these options into context, during the 18th Century only approximately half of the eligible women in Venice were married.  So, what was convent life like?  Walled-off as they were, women had no possibility of participating in the intellectual and scientific life of the times, these wild, wonderfully-rich time.  Convent life thrived throughout Europe, where patrician girls were sent to be educated and kept secure until a good marriage was arranged for them.  They were generally taught poetry, music, embroidery, and other skills useful for managing a household.  Indulge me for just a minute, just imagine how different the story of science today might have been if half the sons of the ruling classes were, as youngsters, sent away to spend their lives behind stone walls while their sisters were free to pursue their intellectual interests.  To give you a feel for how prevalent convent life was, consider that during the Italian Renaissance, Venice, a city with a population of 86,000 have 50 convents and 3,000 nuns.  To put these numbers in context, I live in a town with a population of about 10,000, roughly one-tenth the population of 16th Century Venice.  We have one grocery store, one drug store, and one auto mechanic shop.  We would have to have five convents in my little town to have the number of convents proportionate to the number in Venice during the Renaissance.  Is it any wonder that the pursuit of science was then and has continued to be deemed a male pursuit?  Social role theory teaches us that when occupations are sex-typed, as science was during the Renaissance, it is human nature to infer that there must be something inherent about men that predisposes them toward math and science, something that women do not possess.  Before moving ahead, I want to share two brief vignettes.  Galileo, arguably the most illustrious of the Renaissance scientists, had three illegitimate children -- two daughters and a son.  Of the three, his eldest, Virginia, was the only one who mirrored his own brilliance, industry, sensibility and virtue and was, in his words, a woman of exquisite mind.  But he deemed her unmarriageable because he had not married her mother.  She was illegitimate, as were all the other two children.  At age 13, he placed her, and her 12 year old sister, in a convent, where they lived out their lives in poverty and seclusion.  In contrast, his son was legitimized by Fia, by the Grand Duke of Tuscany, and went off to study law at the university.  The second vignette has to do with Leonardo da Vinci, who was also an illegitimate son.  He was born to a peasant woman.  Had he been born a girl he would have been deemed unmarriageable, and, surely, been sent off to spend his life behind convent walls.  Let s jump ahead to the 17th Century in America.  There were two remarkable characteristics -- development of schools for sons of wealthy families, and strong limitations on education for girls.  A recurring rationalization, limiting girl s access to education, was that learning would have serious negative effects on them.  Here are some of the arguments.  As the brain develops, the ovaries shrivel, which was medical wisdom at the time.  Education will undermine their health and that of their future children.  Education will decrease their willingness to do housework and obey their husbands.  Education will lead to their inclusion of men s activities, as you re taking over men s jobs.  Echoes of these concerns can be heard today in the rationalization that women can t manage both a career and family, and don t have the physical stamina to pursue demanding work at the highest levels in math and science.  So, while the stone walls at boys schools lock girls out, the girls open their own schools.  In the 19th Century there were dame schools.  These were informal instruction held in the homes of women who lived nearby where girls were taught basic reading and writing, embroidery, and other feminine skills.  From the mid-1820 s to the present in America, there was a rapid spread of education for women in the U.S., mainly these dame schools, and, later, to women s academies, which were characterized by having all female faculty -- they were all single women, they could not be teaching if they were married, but, yet, they provided rigorous training in science, certain sciences.  These schools became a female enclave within the American scientific community.  They were not intended to open new careers for women, but, rather, to make these women better mothers for the American republic to raise moral and patriotic sons.  Despite the odds, women pursue their interests in science.  The early 19th Century saw a groundswell of popular books and textbooks written for women by men and women on such scientific subjects as botany, chemistry and geology.  There was an enormous audience hungry for such books reflected in this following.  These are the sales figures.  Conversations on Chemistry, which was published in 1806, went through more than 15 editions in the U.S. before 1860.  Familiar Lectures on Botany in 1829 went through at least 17 editions and sold over 272,000 copies by 1872.  Introduction to Botany had at least nine English editions by 1841.  But, we have women s employment in science.  Where were they employed?  The women s colleges had been the largest employers of women scientists, but they did very little research, these women faculty, because they had very heavy teaching loads, which was true until relatively recently, and they had no incentive in terms of advancement.  These women had the best jobs that were available in science at the time.  They rarely had PhDs.  Why?  Because they were not admitted to matriculate in any graduate schools until the 1890 s, and progress was slow and uneven in granting matriculation rights to women graduate students.  For example, Princeton, Harvard and NYU refused to matriculate women students to their graduate schools until the 1960 s.  As the women s colleges upgraded, they required PhDs for the new faculty.  Since there weren t many women faculty, and they wanted to upgrade their images, they started hiring male faculty, married men faculty, and so the number of women who had positions eligible for them were decreased.  Women retirees were replaced by men as women s colleges wanted to have a different face than older, single women.  And, so, even as women got more PhDs, employment became a problem.  Administrators also had fears that women were bad risks, again, the stories of being vulnerable, they couldn t hack it, they didn t have what it takes.  To induce male PhDs to teach at women s colleges, they had to offer them special incentives, including reduced teaching loads, good salaries, research support, and living quarters for their family.  So while women couldn t marry, men could marry and the schools provided support for them.  So, not surprisingly, men started to produce research results, lending support to the belief that women s accomplishments in science do not match those of their male counterparts.  Having finally gained access to higher education in science, women s next hurdle, then, was to find employment in science.  It was hard to get early data on employment, so I m going to jump ahead to some recent data.  These are data from 2002.  As you can see, women had many more opportunities of places to work, but only 42 percent of PhDs are working at four-year colleges.  Even though the greatest growth in employment, and some of the most remarkable developments in medicine today have come from the private sector, analyses of gender differences in science are based largely on studies of the Academy, as is the theme today.  In fact, the number of U.S. life scientists working outside the Academy grew from 83,000 in 1980 to 181,000 in 2000.  While all employment settings provide both obstacles and opportunities for women scientists, arguably the most extreme examples are universities, which epitomize hierarchical organizations, and biotech firms, which epitomize flatter networked organizations.  Biotech firms have strong ties to universities, they have flatter project-based employment structures, and cutting edge approaches to science.  Here is an example of a networked organization, where every person has many, many colleagues, and they provide better workplaces for women.  Why?  Because these organizations rely on partnerships to be successful, they re more flexible and transparent than hierarchical organizations.  Moreover, there are fewer chances for sexism to thrive, because advancement is based on input from a wide range of people, rather than a few, as is typical in hierarchical organizations.  Here is some data based on a study of, I think, 2000, women, I had that a little bit later on, you can see here for male scientists, the odds of promotion to supervisory positions do not differ by the organization, but are very, very different for women.  Women were eight times more likely than their counterparts in hierarchical organizations to supervisory positions.  So organizational context mattered, they affect career outcomes for women scientists.  However, the heavy focus, even today, on the situation of women scientists in the Academy reflects a persistent idea that any PhD worth her salt obtains a university position; other options are considered second best.  Despite all the odds, there were many outstanding women who were never given an opportunity to make a career in the Academy, and women scientists today who are holding high level leadership positions in government and industry.  Just to give you an idea, Margaret Mead, probably one of the most notable women scientists of the 20th Century, reportedly was never offered a tenure academic position.  She was hired by the American Museum of Natural History, where she worked out of an attic in 1927, and was not promoted to full curator until 1964, by which time she was world famous.  Rachel Carson, world famous biologist and bestselling author of Silent Spring, never had a faculty position.  Barbara McClintock, 1941, just before the University of Missouri denied her tenure, she moved to Cold Spring Harbor Laboratory where she won numerous prizes, including the 1983 Nobel Prize in Physiology or Medicine, and, notably, she was the first and only woman to receive an unshared Nobel Prize in that category.  Finally, Dian Fossey did her pioneering work in Rwanda as an independent researcher.  She only got her PhD after the major phase of her work was completed.  There are a number of women who are almost Nobel Laureates.  Physicist C.S. Wu of Columbia, performed crucial experiments proving the theory that won her colleagues, Lee and Yang, the 1957 Nobel Prize in Physics.  Biochemist Viola Graham, who helped, there is no picture of her on the Internet, but this is her article with James Sumner of Cornell, for which she shared the 1946 Nobel Prize in Chemistry with two other men, not with her.  Geneticist Esther Lederberg, who helped her then husband, Joshua, with microbial research that won him and two other men the Nobel Prize in 1958, again, not including her.  Marguerite Vogt, the colleague and close collaborator for 20 years on DNA tumor viruses and cell growth.  Renato Dulbecco, who shared the prize with two other men in 1975, again, not including her.  Columnist Anna Schwartz, who co-authored several books with Milton Friedman, and worked for decades on the detailed economic data that formed the basis for the work that won him, alone, the Economics Prize in 1976.  I won t go through the rest, and some of them you know, some of them you don t know.  Ruth Hubbard, the one on the bottom here, she had worked for years on the chemistry vision before marrying her husband, George Wald, in 58, who won the 1967 Nobel Prize for working the same area, and many assumed that they had always collaborated and that he deserved most of the credit for her earlier independent work as well.  Apparently, once she married a scientist of greater reputation, a woman s own independent work, which all too easily be dismissed as a small part of his.  And, today, we have many women scientists in leadership roles outside of the Academy, including the Chairman of Dow Corning, the Chief Executive Officer of [indiscernible], and President of [indiscernible].  We have women, 4 out of the 27 institute directors at NIH are women.  So this list is obviously, not exhaustive, it s merely meant to illustrate that there are many organizations other than university that provide opportunities for women scientists to achieve prominent positions of leadership.  Although much still needs to be done, women scientists today need appropriate recognition for their scientific work.  I m sure you all know the story of the MIT senior female faculty who were marginalized and whose plight was made very well known in the national press.  Compared to male faculty they did not receive equitable salaries, laboratory space, and so forth.  Why does this bias and recognition persist?  Many reasons, I m sure, but, just to give you one idea, one of the processes that has just recently been operationalized, two Swedish scientists noted that female scientists applying for prestigious fellowships at the Swedish Medical Research Council during the 1990 s had been less than half as successful as male applicants.  So the question was, does a peer review system evaluate men and women on an equal basis?  So what they did, the researchers got access to the MRC reviewer subjective scores, and found that they gave, this is the three criteria they rated the applicants on -- scientific competence, relevance of the research proposal, and quality of the research -- and they found that the reviewers gave the female applicants lower average scores than the males on all three parameters, but especially on scientific competence, which is typically related to the number of, and quality of their scientific publications.  So the inference is that women earned lower scores because they were less productive.  But, were they?  So the researchers developed six objective measures of scientific productivity from the applicant s CD, which they checked, and these measures included the total number of papers published in high impact journals and the total number of first-authored papers in high impact journals.  Do the men and women with equal scientific productivity receive the same competence ratings?  Absolutely not.  Here s the graph that shows that.  As you can see, the most productive group of female scientists are those with 100 total impact points or more was the only group to be judged as competent as men, although only as competent as the least productive group of male applicants, the ones whose numbers had fewer than 20 impact points.  To tell you this somewhat differently, their analyses found that to be awarded the same competence score as a male colleague, a female scientist would have to produce approximately three extra high impact papers and journals, such as Nature or Science, or 20 extra papers in excellent specialist journals.  Thus, a female applicant had to be 2.5 times more productive than the average male applicant to receive the same competence score.  And this was done in Sweden.  This study provides direct evidence that a peer review bias system is subject to sex bias.  Clearly, as scientists [indiscernible] and other human beings to the effects of agenda prejudice.  If changes aren t made to this, according to the researchers, a large pool of promising talent will be wasted.  So, clearly, women scientists have broken through the stone walls.  Much has to happen before the invisible walls come down.  Thank you.  [Applause].  Elizabeth Spelke:  Thanks to the American Enterprise Institute, and to Dr. Sommers for organizing this event.  I m very happy to be here.  I have not spent my life studying the topic of women in science, but I have been thinking about it a lot over the last 2� years, since then President of Harvard, Lawrence Summers, made three famous suggestions.  First, he suggested that men may have higher intrinsic aptitude for math and science, both in general and also at the highest levels of ability.  Second, that men may show a profile of motivation that is better suited to high intensity work, long hours of work in the sciences and other technical fields.  And, third, that because of the forces of free market economics, gender discrimination is probably not a major source accounting for the paucity of women in science.  Now, I think all three of these suggestions are very worth discussing, I hope we discuss all of them today.  I want to focus my 20 minutes on the first two of them, and ask, what is the intrinsic aptitude of males and females for science, and what are the intrinsic differences in their motivational patterns that might be relevant to their success in science?  To start with questions of intrinsic aptitude, I think first it s important to note that although science has a long and illustrious history, as we ve just heard, measured against the human scale of historical time, it has a very short history, measured against the scale of evolutionary time.  There hasn t been time for the human brain to evolve special purpose capacities expressly for doing symbolic mathematics and science.  Instead, when we learn and practice science and mathematics, we bring to bare core cognitive systems that evolved for other purposes and we harness them for this new function.  Now, there s been great progress made over the last 10 or 20 years in understanding what those systems are and what their properties are through converging lines of research of five kinds.  First, research on the cognitive capacities of our close primate relatives, research on the emergence of cognitive capacities in human infants, research on patterns of variability and universality in cognitive capacities across cultures, and especially research on the cognitive capacities that children bring to bare when they first learn mathematics and science, and that adults bring to bare when we engage in mathematical and scientific reasoning.  And to make 20 years of research, to collapse that into a ridiculously short 5 minutes, I think we see evidence from all of this work for three fundamental systems at the core of human, mathematical and scientific reasoning -- a system for representing and reasoning about objects, one for representing and reasoning about number, and one for representing and reasoning about geometry -- all three of which emerge spontaneously, quite early in human infants, therefore, very likely have, in part, a genetic basis.  What s more, over the course of the preschool years, we see children spontaneously using these systems to develop systematic knowledge of object mechanics, to develop an understanding of symbolic number and symbolic arithmetic, and to develop, to master a whole host of representational devices that are crucial to math and science, particularly things like maps and measurement.  So, with that research as background, I think we can look at Summers first suggestion in the form of two questions.  We can ask first, do boys out-perform girls at tasks tapping any of the three core systems?  And, second, do boys have superior abilities to harness these systems, for example, for learning symbolic mathematics, learning to use maps and other symbolic geometrical devices?  Let me see if I can quickly give you a flavor for a very large body of work that does bear on these questions.  Let s start with objects.  Well, there s, for as long as 70 years or more, there have been many studies looking at children s spontaneous attention to objects, and engagement with objects.  In the late 1970s, Eleanor Mccabe [ph.] and Carolyn Jackson reviewed this literature, asking, do we see any general sex differences in spontaneous attention?  Their answer was no.  Girls and boys are equally interested in objects, equally interested in people.  But, more recently, over the last 30 years, we ve gone beyond patterns of spontaneous attention and asked, how able are young children to process the properties of objects, retain them over time, and analyze them?  For example, as in this first case here, how well does a baby who looks at an object, able to hold that object in mind when it moves out of view, how well can they remember it and its properties, or, when babies see two objects, for example, one going inside another, how well do they understand that mechanical and spatial relationship?  How many objects can babies keep track of at once if they see multiple objects disappearing one at a time into a container?  Now, all of these studies provide evidence for interesting patterns of abilities and limits, and we can look at that evidence and ask, is there evidence for sex differences, systematic differences between male and female infants in these abilities?  The answer, in a word, is no.  There is broadly, highly convergent patterns of performance by the two genders in infancy.  What about later in childhood?  This system doesn t go away.  It continues to exist in children and in adults, and one task that s become very popular for tapping this system of object representation.  In adults, it s a so-called task of multiple object tracking.  I don t have time to really describe this, but, just sufficient to say, it s like a shell game where you re confronted with an array of objects, you have to keep track of some subset of them, and they all move around independently, and you have to keep your attention on them as they move.  It s a very demanding task for us as adults.  We tend to fall apart if we have to keep track of more than about four of these things.  And it s possible to design child-friendly versions of these tasks, and ask, how good are children at keeping track of multiple objects over time?  We ve been doing this with children aged 4-8 years so we can go back and ask, are there gender differences?  Here s average performance of boys and girls, highly equal.  Here s performance at the high end, the kids who are best at keeping track of the most objects, the most reliable, and what you see in both of them is no tendency for boys to be outperforming girls on object tracking tasks.  Let me move to number, lots of evidence that starting very early in infancy, perhaps at birth, babies are sensitive to approximate numerical magnitudes.  They can tell the difference between an array of eight objects and an array of 16, though not between eight and nine.  They are sensitive to ordinal relationships between objects, and they can even engage in a form of non-symbolic arithmetic.  If you show an infant an array of eight objects go into a box, and then a second array of eight objects that go in the box, they know there s approximately 16 things there, and they react with surprise if you empty the box and reveal only 8 or 32.  So, all these abilities are present at an early age.  Are there sex differences?  No, not in any of them.  We can go on and ask what happens to these abilities over development.  There s a very rich literature in cognitive science and cognitive neuroscience providing evidence that this sense of approximate numerical magnitude is critical for our reasoning about symbolic arithmetic.  What s more, we ve shown recently that if you test for this sense of magnitude in children just at the start of school, they have it, of course, but variability in their performance on these non-symbolic tests predicts how well they succeed at mathematics at the end of the kindergarten year.  So we can ask, at this age, do we see a difference between males and females?  Again, the answer is no, not on average, and not at the highest levels.  Well, finally, let me turn to geometry, and I wish I had more time to tell you about two lines of research; one showing that very early in development infants are sensitive to the geometry of the surrounding spatial layout, and they use this sensitivity in their navigation.  I decided I didn t have time to talk about that.  But, second, infants are also sensitive to the geometry of visual forms, basic geometric relations like distance and angle, and equally sensitive to this geometry in studies of infants.  What happens later in development?  Let me mention findings from just one study conducted, published last year, that attempted to survey in children aged 6 to 10 years, and in adults, both in our culture and in a remote Amazonian culture, sensitivity to a broad range of geometric properties.  It s probably hard to see on this slide, but here s a subset of the things we tested for.  The task is very simple.  On each individual trial, people see six different geometric displays, five of which share a property, like perpendicularity, the sixth does not, and their task is to indicate which is the outlying figure.  Of course, I ve indicated the outlying figures here, they re not indicated for the subjects in the task.  And at cross-trials we test for a variety of Euclidian properties, like angle and distance, and also topological properties like chirality and degree of connectedness.  So, looking across sensitivity to all these different properties, do we see a difference between male and female 6 to 10 year olds?  Answer, no, not on average, and not at the highest levels.  If we look just at the people who are doing best on this task, we do not see a preponderance of males.  But, I want to spend 1 more minute on this task and show you what the trial-by-trial performance looks like.  Now we tested for somewhat more than 40 different geometrical relationships.  What you see here is trial-by-trial, the performance of girls in green, and of boys in blue.  Notice, first, that there isn t a single geometric property that is easy for one gender, and hard for the other.  Rather, what you find is highly convergent performance across boys and girls.  But if you look hard, you also see a couple of properties that trend toward showing a gender difference.  In particular, by a too lax statistical method, we see two properties where females tend to do better than males.  I give an example of one of them there.  And across this whole test, we see exactly one property where males tend to do better than females.  Now that may look familiar to people with any background in discussions on sex differences.  That property, the one property males do better at is what is tested in tests of mental rotation.  So if you think spatial ability depends only on mental rotation, one of the 40 things we tested for, you should conclude that there is a male advantage at space.  If you take a broader view, that advantage seems to go away.  Well, what about sex differences in the capacity to harness to these abilities to do higher-level things, like symbolic arithmetic, and map reading?  Let me quickly go through the evidence here.  Now, the symbolic system of number is mastered by children roughly over the ages of 2 to 4 or 5.  They learn to count, and they learn to use counting to do elementary, simple arithmetic.  There s a lot of variability in the speed at which different children acquire these abilities.  And that, too, is predictive of their success in elementary school mathematics.  Are there gender differences?  Answer, no, highly similar patterns of performance over this age range in mastery of the counting system.  What about map reading?  Beautiful research over the last 20 years shows that it s at about age 4 that children first start understanding visual representational devices like maps.  So we focused on that age, and presented children with a simple, purely geometrical map.  So, in this task, children are shown a piece of paper, it s got three elements on it in a particular geometrical relationship, and they re told that they will be able to find a toy at a location indicated at one of those elements.  Then they turn around, they have to leave the map behind, turn around, face an array of three objects ten times bigger, three-dimensional array, and they have to find the object that corresponds to the position on the map.  This is a purely geometric map-reading task.  And we can ask, are boys better?  And the answer, again, is no.  So, in summary, we ve looked at three core systems and see no advantage favoring males, two emerging cognitive skills and also see no advantage.  My conclusion from this work is that the state of the evidence to date does not favor the hypothesis of a male advantage in intrinsic aptitude for math and science.  What about gender differences in intrinsic motivation?  Are women biologically predisposed not to like math and science as much as men, or not to work in patterns that are most conducive to high levels of achievement in math and science?  Well, in contrast to studies of cognitive abilities, it s harder to answer this question from studies of motivational patterns because most of the motivational variables that psychologists have looked at systematically don t have the right kind of properties to be good candidates for a genetically determined difference in motivation.  Within a single person, most motivational variables change considerably over development and change when you move from one situation to another, or change when you move from one culture to another.  Certainly variables like, what do people want to do, if you ask people what they want to do you find enormous differences across cultures, and within one culture across historical time.  But there is one motivational variable that does look like a good candidate for a genetically determined factor that could be predictive of success in science and other things, and that s the variable of self-regulation.  Now, the systematic study of self-regulation began in the early 70s with work by the psychologist, Walter Michel, who took 4-year old children and presented them with the following choice.  He showed them two treats, one little treat, like a single marshmallow, another big treat, like a whole pile of marshmallows, and he said to the kids, okay, you ve got a choice, you can have the single treat any time you want, any time you decide you want that treat you ring this bell and I ll come into the room and give it to you, or you can have the bigger treat, but, if you want to have the bigger treat you ve got to wait for me to return.  Okay?  So, ring the bell you get the little one, wait for me to return you get the big one.  And now he leaves the room.  The kid is alone in the room and he simply measures, how long does the child wait to get the whole dish of marshmallows instead of the single marshmallow?  Now, this simple measure is astonishingly predictive of later achievement.  It predicts how well these kids studied at age 4, this measure predicted how well they would do in high school, it predicted their SAT scores.  In a follow-up, when these people were in their 30s, it predicted their highest level of educational attainment, and in a very recent study focusing on 8th graders, this variable, measured in a more age appropriate way for 8th graders, out-predicted IQ in predicting the child s ultimate school achievement that year.  So children were tested toward the beginning of the year in their capacity to self-regulate, ala Michel, and also in IQ, and then at the end of the year they asked, how well did they do in school, who got into the competitive high schools, both variables predicted, but self-regulation predicted more.  So, we can ask, are males better at self-regulation?  Well, most studies find no sex differences, but there was a meta-analysis recently.  The results of that meta-analysis are that sex differences are extremely small, but they do seem to exist, and they don t favor males.  There is a slight advantage for females in tasks of self-regulation.  So, stepping back from this work, I think what we see is evidence that boys and girls are equally endowed with the core cognitive abilities at the root of science and mathematics, they are equally likely to show at least one motivational pattern that does seem to be the only one that I found that actually seems to be predictive of later academic success.  And in light of those findings, we shouldn t be surprised by the statistics that show that today girls and boys perform equally well in math and science subjects, both in high school and in college.  There is no evidence from any of this work that males are better suited to become scientists.  So I want to end with a question, which is, why do so many people believe that that s the case?  I think that there are many potential answers here, but I want to leave you with just one suggestion.  The suggestion is that as we heard from Dr. Barnett, there are many more male than female scientists practicing in academia today, and, at any given time in history, it can be extremely difficult to distinguish what s typical of the practitioners of a profession from what s necessary to success in that profession, and it s easier to see this if you go back to a different historical time.  The time I want to take you back to is the 1930s, the institution, once again, is Harvard, the person is E.G. Boring, who was essentially the leader of the Psych Department at Harvard for 20 years.  Boring was an extremely tolerant person.  He has claimed that the only thing he was intolerant of was intolerance.  He was opposed to anti-Semitism.  He had many Jewish students.  But when he wrote letters of recommendation, he rarely recommended his Jewish students for the top jobs.  He would say things like, this student has done some really high-quality work, but, unfortunately, he s not cut out to be a scientist.  Now, why was he not cut out to be a scientist?  Well, Winston has gone back over the letters written both by Boring and by his colleagues describing some of these students, and they point to the following negative characteristics -- too talkative, too aggressive, too eager, and, my personal favorite, Christina, you should like this one too, gesticulates excitedly.  Now, what s going on here?  Boring emphasized throughout that every person has to be judged on their merits, but, what s being judged is who is suitable for science.  Now, in Boring s time at Harvard, science was the province not only of men, but of Christian men from genteel backgrounds, who let him finish his sentences, who didn t get too excited about their hypotheses, didn t push them too hard, and he came to the not unreasonable conclusion that that kind of dispassionate, calm consideration of the evidence was a central ingredient for science.  Okay?  Now, a lot of time separates us from the time of Boring.  Fortunately for Christina and me it s okay to gesticulate a little when you talk, and you can still be a scientist.  Since Boring s time, the gates of the academic community have opened wide to people from all different religious backgrounds, to people from many, many different cultural backgrounds, so that a far wider range of talent has gotten in, and, not coincidentally, with this opening of the doors, and the increase in scientific talent, the second half of the 20th Century in the United States saw the greatest time of progress and flourishing in science than I think has ever been known in human history.  What we need to ask now is, are we at the end of this process, or can we do still better in science by opening the gates still wider to the most talented people?  Beyond bias and barriers, the report at the center of the discussion today argues that we can do still better, and I hope that we do.  Thank you.  [Applause].  Christina Hoff Sommers:  And now Professor Haier.  Richard Haier:  Well, good morning.  I can t see everyone over to my left, if you can t see me, I was going to say, if you visualize George Clooney you won t be that disappointed.  [Laughter].  My wife told me that.  She s on medication.  [Laughter].  Well, I m glad to be here this morning.  This is an extremely interesting topic.  I have not been a sex difference researcher very long.  I have found some things that differ between men and women in our studies, I m going to tell you about some of them today.  I have one slide, further on, where I m going to leave the podium and go over and try to make it work, it s a little animation, and I can t do it from here.  But, I want to start by telling you about Daniel Tammet.  Anybody know Daniel?  Daniel has just written his autobiography called Born on a Blue Day.  Daniel is autistic, so, he s got those genes, he is also a savant, he s got those genes, he also is normal to above-average intelligence, which is extremely unusual.  Most of the time when you see a savant they re mentally retarded, often profoundly, except they can do one mental activity in an extraordinary way.  Daniel is very good with numbers.  He s very good at learning foreign languages.  He learned Icelandic in a week when they took him, the BBC, I think it was, took him to Iceland and put him with a tutor and a week later he was interviewed on Icelandic television fluently.  Really quite extraordinary!  He also, for a charity, decided that he would memorize the digits of Pi, you know, 3.14 and then the sequence of numbers goes infinitely.  How many digits do you think you could remember if I gave you like $10,000.00?  Could you remember 500?  No.  Daniel remembered and he recited on television with people checking 22,514 digits of Pi.  22,514, it took him just over 5 hours.  He just kind of read it off his memory.  What is his brain like?  Wouldn t you just love to know what that brain is doing when he s memorizing those digits, when those digits are stored and when he s recalling those digits?  Really extraordinary!  The brain, of course, is very complicated, it has lots of different pieces to it, each piece does something a little different, and it is an important thing to understand how the brain works and what is the relationship between the brain and cognition.  I think we all understand that cognition happens in the brain, it s fundamentally a biological process, and it s extraordinarily complicated.  This cartoon says, separated at birth, the Mallifert twins meet accidentally, and they are in the patent office with the same invention.  We know from studies of identical twins and identical twins reared apart that scores on IQ tests are very heritable, somewhere between 50 and 80 percent of the variability of IQ scores is probably a genetic, it s really an important finding because if something is genetic, it means there is some biology underneath it.  If something is biological, it may or may not be genetic, but if something is genetic, genes work through biology.  So, knowing that there is some component of heredity in intelligence really validates the idea that there is some underlying biology of intelligence.  This can be studied with brain imaging.  This is an image from a technique called positron emission tomography, where you inject radioactive sugar into a person and basically see where it goes in their brain while they are doing some interesting task.  So a PET scan like this, will look different in the same person, depending on what their brain is doing during the procedure.  So if you want to know where is silent reading, you have the person do reading.  If you want to know where math is, you have a person do math.  This is a slice of the brain just as if I took a knife and sliced right through my head and brain like this, and then looked down.  So here is the front part of the brain, here is the back part of the brain, here is the right, here is the left.  The colors show the amount of sugar activity, the harder a part of the brain is working, the more glucose, that s sugar, the more glucose it uses, that s where the energy from neuron firing comes from.  And, you can see the parts of the brain that are most active here in white, reds and oranges.  And, if you re ever on Jeopardy and the final category is positron emission tomography, the scale here is micromoles of glucose per 100 grams of brain tissue per minute, which all I mean to tell you about that is this picture is quantified, you can do statistics on it, it s real science, so you can tell how hard the brain is working.  Now, back some 20 years ago, we decided to use PET scanning to ask the question, where in the brain is intelligence?  So if you have people working on difficult reasoning problems during this procedure, you can see what part of the brain lights up.  This would be an example of the kind of problem we used.  There are some symbols up here arranged according to a rule or a pattern.  You have to intuit that rule or pattern and then pick from the eight multiple choices down here, the one and only one choice that goes in the lower right-hand corner to complete the pattern.  Anybody want to yell out the answer?  How many people don t have a clue?  So we have a few supervisors here from various activity.   If you look at the top row, you see you have four diamonds, two are white, two are black.  Look at the next row, four diamonds, two are white, two are black. Look at it this way, four diamonds, two are white and two are black, and, therefore, the answer is?  Number three because now when you put one black diamond in the corner here you have two black diamonds and two white diamonds, so you have four, the same this way.  So the pattern is maintained.  And I m sorry to tell you, but this is actually an easy item.  The actual test we use is considerably more difficult, it s a timed test, people work on it for 32 minutes while that radioactive sugar is labeling the brain.  And what we found was really quite surprising to us.  No one predicted it.  We have two brain slices from one person on the right, and the same two brain slices from a different person on the left, all on the same metabolic scale.  The person on the right scored 33 out of 36, the highest score in the sample.  The person on the left only scored 11.  Do you notice anything unusual here?  This person who is not doing too well, that brain is really working hard.  The person over here is doing very well, much lower metabolic rate really throughout the brain.  This led us to the idea that it s not how hard your brain works that makes you smart, but how efficiently your brain might work to make it smart.  In 1988 this got a lot of attention.  It was one of the first studies of its kind.  This picture was published in Newsweek.  There were cartoons about it.  It s probably way too simple a thought to say that intelligence has to do with brain efficiency, but it s been a pretty good concept and a lot of research on it.  We next went to a study of the computer game Tetris because we wanted to know what happens when you learn something.  Does your brain get more efficient?  We found people who had never heard of Tetris back in the early 90s when it was introduced and we did a PET scan on them before they practiced, here, and then the same person after practice.  This is after 50 days of practice.  50 days, they got extremely good, they were going so fast in this game you could scarcely believe a human being could do this.  What happened?  They are processing more stimuli, it s faster, it s harder, and metabolic rate in the brain actually went down.  Moreover, it went down the most in those subjects that had the highest scores on the intelligence test.  The smartest people became most brain efficient.  Now this is interesting.  And then, I had an insight.  When I was in graduate school in 1971 at Johns Hopkins, that was the year of the now famous study of mathematically precocious youth, started in that year, and I actually worked on that.  I had a long-standing interest in mathematical precocity.  It occurred to me, we can resolve this, definitively, what we ll do is PET scans in men and women while they re doing an SAT math test, and we can see, do the people who do the best on the math test, are they the most brain efficient, and then we get to see any sex difference in this.  This would really be cool.  And, what we found, these are group composite images, 11 people in each group, and, to make a long story short, in men the higher the metabolic rate in these temporal areas the better they did on an SAT math test during the imaging.  In women we couldn t find any area that was related to how well they did.  And, by the way, in this study, the men and women were matched on SAT scores.  We had men and women matched for having SAT math scores over 700, and then in the average range around 500.  This is really interesting because, and it didn t show anything about brain efficiency, by the way, it showed the harder the brain was working here in the men the better they did on math, there was no such relationship in the women, women were just as good at math, how they did it, according to this study, remains a mystery.  This was published back in 1995.  I show it because it illustrates that even when you equate men and women for high-end performance you don t necessarily get the same brain result.  This was a big hit back in 1995.  Now I want to talk about a different kind of imaging, it s called structural imaging.  This is what a typical MRI scan is.  PET scanning is functional.  PET scan changes if you re reading or doing arithmetic, if you re awake or asleep, if you re alive or dead, the PET scan looks different.  Not so with a structural MRI.  You could put a cadaver in the MRI and you ll get a brain image that looks just like a living person, it s structural, there is no functional information.  Interpreting functional imaging in this business is very difficult, it depends on the task and so on.  Structural imaging is actually a little easier.  We re using a technique called voxel based morphometry, where we have algorithms to segment gray matter and white matter, we can quantify this voxel by voxel in the image.  We did a study where we correlated IQ scores with the amount of gray matter.  The red and yellow areas show where there are statistically significant correlations between the amount of gray matter and IQ.  You can see here that there are areas in the frontal lobe and areas in the left hemisphere and some areas in the right hemisphere, there are areas distributed around the brain, they re not just in the frontal lobe, and these are areas, this is a sample size of about 47 people, there are areas where the more gray matter you have the higher your IQ score.  Now, if IQ scores are kind of meaningless, why would they be related to this kind of gray matter structure in the brain?  So, this is interesting, but what really surprised us, and, by the way, in this analysis here, sex was statistically removed to have no effect.  This is a combination of men and women.  When we broke this out separately by sex, we found this.  And if you remember nothing else of my talk today, remember this slide, because we were really shocked at this, because, again, the men and women in this sample were matched almost identically on IQ.  They are exactly the same.  Yet, what we found in the men was more frontal areas than in the women, although the women had this nice language area that the men didn t have.  The men had this big area back here in the parietal occipital lobe, kind of a visual spatial area.  There was white matter involved in the women that was not involved in the men.  This was pretty striking because what this means is that there might be more than one brain design to lead to comparable cognitive performance.  Not all brains work the same way.  Teachers have known this from day one.  Not all brains work the same way.  That s what I believe these data are telling us.  Here s the same data just showed from a different viewpoint, so you can really see the frontal lobe differences between men and women.  The parts of the frontal lobe are different in the women than in the men.  Now, interestingly, those areas that I just showed, the area back here and parts of the frontal lobe, are highly genetic.  This is a study of twins done by the UCLA imaging group.  The red area shows the parts of the cortex that are most heritable.  Not every part of your cortex is equally genetic.  They re different.  Many of the areas that we see related to IQ are also highly heritable.  That s interesting.  Recently we ve published a review of all the brain imaging studies of intelligence.  There are 37 of them.  What we have concluded is that there is a combination of areas in the parietal lobe, particularly, and in the frontal lobe, with some other areas as well, that these are the areas that turn up most often as related to intelligence in these 37 studies, all of which use different imaging technique, all of which had different kinds of subjects, all of which had different measures of intelligence, yet, these areas came up quite often.  So we believe that the way information flows around the parietal frontal networks is related to intelligence.  I may come to define intelligence someday by some measure of the efficiency of information flow around these specific areas.  Here s our P-FIT model down here.  Here are 111 subjects, men and women together, showing where there is a correlation between gray matter and IQ.  We have these frontal areas, we have these language areas, we have these temporal areas back here, but when we take this group and break it down into, I think this is 57 males and the rest females, you see men have a different part of the frontal lobe, and men have this posterior area here that women don t have.  Again, the men and women are matched on IQ.  There is not a question of men being better, women being better, this is a full-scale, (WAIS-R) Full scale IQ scores.  I think this is interesting, and I want to point out that we now have another sample, completely independent sample of 100 people, and as we are sitting here today, my assistant is grinding those numbers through to see whether we can replicate this in that other sample of 100.  By the way, if I put these brain areas, the amount of gray matter in these areas into a multiple regression equation to try to predict IQ, the amount of gray matter in just five areas predicts IQ extremely well.  Now it depends on kind of cross-validating this.  So, Einstein s brain actually was bigger right here and had more glial cells, which are support cells for neurons.  I m not going to spend too much time on Einstein s brain, but, this is the slide I want to just see if I can make it work for a second.  Now watch that.  Now let me tell you what you re looking at, and then we ll tap it one more time.  This is a new imaging technology called magneto encephalography -- it measures magnetic fields millisecond by millisecond.  What you re watching is what goes on in the brain for a period of 1 second after a person presses a button, one button if a light flashes on the right side of the screen, a different button if the light flashes on the other side of the screen, left side of the screen.  What you re watching is a single trial, one button pressed, and following that button press, when the light comes on you re seeing how the brain responds to the light and decides to press the button over 1 second.  You ll notice, as we play it again, 100 milliseconds, it starts here, then we get all this here, and now we re in the front, there s nothing more back here, and then it fades away, by 900 milliseconds the brain is done.  Now, imagine doing this in Daniel while he s performing some of his feats.  What you saw in that 1 second of time to a very simple decision gives you some idea of the complexity of what we re dealing with here.  That was, is the light flash on the right or the left, and look at the brain activity that was going on.  We know almost nothing about how the brain works.  We know almost nothing about whether sex differences in cognition, even where there are some, and, in most places there aren t, but even where there are some, we know almost nothing about the brain.  I m going to conclude by just asking the question, wouldn t it be nice to have some brain imaging in extremely successful women?  Here we have Rosalind Franklin who was probably the real discoverer of the double helix and the DNA story.  She did imaging.  She was visual-spatial.  And, of course, we have Einstein himself.  Wouldn t it be just great to know how these brains work?  I think the only problem I had with the NAS report is it really kind of prematurely excluded the idea that biology has anything at all to do with this.  I think that s an overstatement and really not correct.  I ll stop with that.  [Applause].  Christina Hoff Sommers:  Mr. Geary.  David Geary:  Thank you for inviting me today.  It s been a very interesting group of talks and I m sure there will be a lot of interesting discussion.  I m going to move well beyond the Renaissance, or, well before the Renaissance, and start at the beginning.  What I really want to do is to give us a framework for thinking about why there might be biological sex differences, and then, of course, trying to relate those to sex differences in the modern day becomes a bit more complex, but, we have to start somewhere.  I m going to tell you a little bit about the evolution of sex differences, why we should expect some sex differences in humans, sex differences in these core knowledge areas that Dr. Spelke talked about a bit as these emerge during development, and then, how do we make these links to math and science, or, can we make these links to math and science?  So, Darwin was a brilliant guy, it would have been nice to image his brain as well.  This is from one of his books in 1871.  Here we see sex differences across a wide variety of species for the antelope species down there, the females don t have the horns.  His question was, is there mechanisms, or, are there mechanisms that operate in nature that can explain all of these different sex differences?  He said, yes, there was, and proposed that many of these differences emerge as a result of intrasexual competition, competition for access to mates, or intersexual choice, discriminative choice of mating partners.  His theory went unnoticed or ridiculed for about 100 years, but now we know that these patterns are ubiquitous in nature in all sexually reproducing species to some extent.  About 30 some years ago these intersexual choice, and intrasexual competition were linked to sex differences in parental investment, the higher investing sex, typically females, but not always, as I ll show in a bit, are typically engaged in less competition and are more choosy when it comes to mates, and, even more recently, the sex differences in parental investment have been tied to an even more fundamental difference in the potential reproductive rate, and I ll give an example of that in a bit as well.  Here I show a bowerbird, and, this is a bower, you re probably familiar with these.  This is the male, this is the female, and this is his stuff, bower bling, I guess would be the correct term for that, the scientific term for that.  These bower, these male bowers build these huge stick structures, they collect all sorts of stuff.  This is not a nest, this is just to attract the female.  Building these structures is a very complex type of activity, but, nevertheless, has evolved, it s one of the most complex behavioral forms of competition in nature that we re aware of, except for, perhaps, humans.  We see some interesting sex differences, and there are a couple of points I want to make.  One is, females mature at 2 years of age, males physiologically mature at 7.  They begin to reproduce at 10, if they reproduce at all, 15 percent of males sire about 85 percent of the offspring.  Now what s important here -- two points, one is, we have evolution creating a very long developmental period.  During this developmental period, in males in this species, for humans it s for both males and females, during this long developmental period we see a lot of practice going on, male bowerbirds watch mature males, they imitate the bower building activities, they practice fighting, they do all sorts of things.  It takes them at least seven years to get good enough to even be competitive in bower building.  This is probably related to prenatal exposure to male hormones, but practice during development is very important.  Testosterone, post-natal testosterone levels are related to the energetic features of competition, but not to the skill.  So males with a lot of high levels of testosterone gather a lot of sticks and a lot of bling, but their bowers aren t any good, necessarily, it depends on what they ve done during the developmental period.  So, there s a lot of things going on here.  Biology does not mean development and experiences are not important.  In fact, biology can result in the evolution of sensitivity to experience.  Sexual selection also gives us a way of thinking about within sex variation.  Here we have another example from Darwin.  We have hummingbirds here, the male has this long symmetric tail feathers, and that s the female there.  Females choose these males because males, the offspring of these males have lower mortality rates than the offspring of males with shorter tails.  This leads to a situation of directional selection, in this case, the bigger the better.  Strong directional selection, if it differs for males or females, or if it is the same for males or females, leads to two changes, generally, higher genetic variance, in this case, in the males, related to the development of these tail feathers, and reduced canalization of these traits.  What that means is that not only has evolution resulted in more genes for the development of this trait, but also in increased sensitivity to the environment.  So, males of these species that are put in good environments have great tail feathers, so to speak, and those in poor environments suffer the most.  Females are in between.  So we have the evolution of greater sensitivity in males and females, but only for this particular trait.  So, evolution is, there s a lot going on there.  A critical test of Darwin s theory comes from species with sex role reversals.  These are very important.  Here we have pipefish, but we see the same pattern in other species, you get a number of species with sex role reversals.  This is a pregnant male, believe it or not.  If you ve seen this before, males have this ventral pouch here, females deposit eggs in there, and then the males fertilize them.  The males basically take care of the offspring.  Going back to reproductive rate, this means that females can reproduce faster than males, which is a reversal of the typical pattern.  And, so, the limiting factor in female reproductive success is the number of non-pregnant males that they can convince to accept their eggs.  In this situation, since it takes longer to incubate than to deposit eggs, we predict females are going to be more competitive than males, and that is, in fact, that we find females compete more for males, males are choosier, and here we see a female down here, they re larger, more colorful, and, behaviorally, more aggressive.  Now, what about humans?  What can we say about sexual selection in humans?  A lot of arguments that a lot of these things are just so stories, that people are just kind of making up arguments about what they think works based on whatever, but, we can do much better than that.  Comparatively, so across species, larger males, as strongly related to physical male-male competition and polygyny, 18 percent of primates are basically monogamous, and, in those species, there are no physical sex differences in size, or very small differences, and no differences in developmental, pattern of developmental maturation.  When we get bigger males than females, there s almost always an evolutionary history of males beating up on one another for access to mates.  Anthropologically, we see one-on-one in kin-based coalitional male-male competition is common in industrial societies, conflict at least once a year, rating [ph], warfare, so forth, occurs in about 90 percent of these societies, and the mortality rate of young males is 25 percent, very, very serious business.  Population genetic data, Y chromosome, as well as mitochondrial DNA, are very consistent with these patterns here.  We can look at the general principles, we can look at consistencies across species, we can look at what people do in the wild, and say that, yes, I don t think we can dismiss this as a way of approaching sex differences.  So, predicted human sex differences, there is many, most of which we cannot get in here, talk about today, but let s focus on male-male competition.  We d expect, as with the bowerbirds, that the emergence of these traits during development will be a result of an interaction between gene expression and experience.  For a species with a short developmental period, experience is less important.  We have a huge developmental period, about twice as long as common chimpanzees and bonobos, our closest relatives.  So we are designed to be open to experience, both males and females.  Children s early cognitive competencies and interest biases will interact in ways to facilitate the development of sexually selective traits, so traits that are related to these components of sexual selection.  We expect to predict sex differences in activity preferences and sex differences in later interests.  But, I want to have a caveat to that.  Darwin noted in 1871, there s a striking parallelism between mammals and birds and their secondary sexual characteristics, namely their weapons for fighting with rival males in their ornamental appendages and in their colors.  In both classes, when the male differs from the female, the young of both sexes almost always resemble each other, and a large majority of classes resemble the adult female.  In both classes, the male assumes the characteristics proper to its sex shortly before the age of reproduction.  What s important here, the point here, is that it is not the case that if it s biology and hormones we find early differences, and if it s culture and experience we find later differences.  It s not that simple.  In many species a lot of the evolved sex differences related to sexual selection don t emerge until the time of maturation, or emerge slowly during the developmental period, as we saw with the bowerbirds.  So, there s a lot going on there and we have to consider development and the evolution of the developmental period.  So, of course, it s a long way from sexual selection to math and science.  We saw a lot of important male scientists during the 14th Century, and so forth, but we didn t get any data on their reproductive success, so we don t know how Euclid did, and others, we know that Newton didn t do so well.  So, as Dr. Spelke said, we have to look at the core evolved biases and see how these interact with experiences during school and during development and, as related in this particular case, to math and science.  Where to look, I m going to skip that.  So, some hypotheses.  I would like us to consider that aspects of male-male competition during human evolutionary history, in particular, tribal rating, tool construction, which is typically, or, almost is very highly male biased, resulted in the elaboration of some physical systems, in particular, travel and novel territory.  This would involve the elaboration of three-dimensional visual spatial systems for navigation.  These would be more complex types of things than the types of items that Dr. Spelke showed us.  Tool construction, we would predict interest in objects, object manipulation, mechanical reasoning, use of weapons, implicit understanding of object motion and trajectory in three-dimensional space.  These are the primary or core abilities and motivational biases that may contribute to ease of an interest in learning modern day mathematics and sciences, mathematics and some aspects of sciences, but, I actually expect these to be very limited in terms of which sciences they influence.  In any case, are there sex differences?  Well, early on there s not much, but, during development, we do begin to see differences.  There are differences in interest in mechanical toys, some forms of construction play that emerge during the preschool years.  This emerges by the end of the preschool years into a very large difference.  This difference is related, at least in part, to prenatal exposure to male hormones.  One study recently showed that boys have a better intuitive understanding of tool use by 18 months of age, a follow-up study conducted by another group showed that skill development, as related tool use, was related to object oriented play in boys, but not girls.  So, the benefits of experience seem to differ there across the sexes.  Sex differences in three-dimensional spatial abilities and mental retention, we see, and, of course, that s not the only spatial abilities, but there are differences there.  This is looking at sex differences, average abilities, and, so, if there are no sex differences, we d have a cut right here at 50 percent.  Across a number of language domains, 65 percent or so of women do better than the average man, that s what that means.  In some areas of language production the differences are much larger.  I don t know if any of you have noticed that a lot of guys have these pauses in their speech.  Have you guys noticed that?  It s not that they re dumbfounded, I suspect, at least I hope not, I suspect it is a word retrieval difficulty, and, if we look at those types of things, 9 out of 10 women do better than the average man, that is, they have fewer pauses.  If we look at three-dimensional spatial abilities, 20 percent of women do better than the average man.  If we look at mechanical abilities, the difference is, 15 percent do better than the average guy.  So there are clearly women out there who do well in these tasks, but not as many as men, and women have advantages in other areas.  It s also important to consider not only sex differences, average sex differences, but also what you re good at.  So, what you decide to do, what you like in school, the occupations you may choose later on, is going to not only depend on what you re good at relative to other people, but what you re good at relative to the other skills that you have.  And, so, this is looking at intra-individual differences, and I kind of separate those out, but, looking at verbal, number, this includes quantitative reasoning, and spatial abilities, 23-24 percent of men, that s their best skill.  Some women, it is their best skill, but, about 6 percent or so; numerical reasoning, 9 percent or so of men, but 2 or 3 percent of women.  And, 20 percent of women, language skills are their best skills.  This is an amalgam of other types of things that aren t relevant to what we re talking about here.  So, intra-individual cognitive competencies are important as well.  Here, we see, some guys are just really good at spatial types of things, and that s the best skill for them.  Cognitive profiles of mathematics and science majors, this is a bit dated, but, it makes a point.  400,000 students looking at the 20 percent within their sex, so, high spatial males, a large number of them go into the physical sciences, engineering, fields, many few are women.  I suspect that this has changed, that this has gone down, and this has gone up.  There is certainly a lot of room for improvement.  So a lot of high spatial ability females aren t going into the physical sciences to the rate they can.  High verbal abilities, we see a lot of men going into the humanities and social sciences, rather than the physical sciences, and women going into humanity social sciences and education.  Sex differences in variability, we talked about the variability issue earlier, and here we see, I can t read that, but, I think that says quantitative reasoning, we have more men at the low end, this is 320,000 11-year olds, basically every child in Great Britain, more or less, at that time, this was published just last year.  We have more boys at the low end, more boys at the high end.  Non-verbal reasoning, we get a similar pattern.  What s interesting here is, we don t get this pattern for verbal skills, in fact, we get more guys at the low end, but more women at the high end.  And I actually predicted something like this a few years ago in a book chapter on sexual selection in cognition, arguing that if females used language to manipulate social relationships, and so forth, as part of female-female competition, we might find such a pattern.  But, what s important here is the variabilities do not extend to everything, but, extend to specific types of things.  Sex differences on the SAT in this country are amazingly robust, surprisingly robust.  As performance goes down, the gap stays about the same, about one-third of a standard deviation.  If we control for three-dimensional spatial abilities, most, or all of this gap, disappears.  So, the sex differences in complex spatial skills that I talked about later may contribute at least to some aspects of whatever is being measured by the SAT mathematic scale.  The mathematically gifted, again, this is the extreme of folks, this isn t the threshold of what it takes to enter these sciences, but, to do very well, Camilla Benbow and David Lubinski have looked at these individuals now for several decades, assessed them when they were 12-14 years old, and those that do well on the SAT mathematics at that age have an enhanced visual spatial working memory, and an enhanced kind of representation, or access to number information, and they do very well professionally, and they re over-represented in number of PhDs, MDs and so forth in the sciences and engineering.  Again, this is the high end, this is not, kind of, an entry-lever types of things.  There are differences in occupational choices and tradeoffs that they re finding in this sample for very talented women, they are over-represented in these fields as well, but, they re less likely to trade off family life and social life for academic achievement, on average.  We saw Einstein s brain a little while ago, and, so, I was going to skip this, but I think I m going to go ahead and read it, since I have 2 minutes anyway.  So, the words of the language, and, this is a description, Hadamard asked a number of mathematicians and scientists kind of descriptions of how they made their discoveries and how they worked, Einstein was among them, and this is a part of his description.  He says, the words of the language, as they are written or spoken, do not seem to play any role in my mechanism of thought.  They re cyclical entities, which seem to serve as elements in thought, are certain signs, or more or less clear images, which can be voluntarily reproduced and combined.  There is, of course, a certain connection between those elements and relevant logical concepts.  (My eyes are strained, so I m trying to read here.)  Very, very interesting description given the enhancement of his parietal region, the visual spatial aspects of his brain, and, of course, the conscious control of thought voluntarily produced, and so forth, it is a good sign of attention control and what s being measured by IQ.  In any case, I want to conclude, I want us to at least consider that sexual selection provides an explanation of many sex differences, this has been demonstrated across hundreds of species, the sex role reversal species provide critical, potentially falsifiable tests of these predictions, a hallmark of a good scientific theory, and, in fact, they withhold, the theory stands up.  Mathematics and the sciences are, of course, dependent on the insights of a few individuals, and the retention of this knowledge across generations, and the teaching to children in each new generation.  Sex differences and this knowledge cannot be the direct result of sexual selection, but may indirectly be reflected through primary or core cognitive systems, and motivational systems, upon which modern day mathematics and sciences are built.  My hypotheses are that sexual selection have elaborated some folk s physical competencies, and associated motivational biases more in males than in females, which indirectly contribute to some sex differences in performance and interest in mathematics and physical sciences and engineering.  Now, I want to add here, these are differences that emerge slowly over a long developmental period.  Clearly we don t know fully the interactions between any kind of core knowledge biases and how these may play out in schools.  We know that both males and females have extended developmental periods, and, therefore, are going to be more plastic or open to lots of different types of cultural-based learning than is the case in most other species in which sexual selection has been demonstrated.  But, nonetheless, I don t think we can dismiss this.  Christina Hoff Sommers:  Thank you.  [Applause].  If it s alright with panel, I m going to go directly to the audience, who may have questions, and we ll start with Steven Rhoads , an expert on sex differences.  We can have another panel just drawn from our audience.  Steven Rhoads:  Should I address it to particular people, or the panel in general?  I guess Professor Spelke and Professor Geary, I d be interested, in particular, in my reading of this as an outsider, it seems to me the most persuasive view biology has something to do with it, is, comes from the nurturing of women, which I don t think is only socially constructed.  In other words, it is related to testosterone, women with more testosterone are less interested in dolls, or their kids, more interested in staying home with the kid, and I just think that that s, if you look at the recent Pew study, still, after three generations of feminism, I think, in the textbooks, you still have, what, 29 percent of women with kids 15 and under, not 2 and under, 15 and under, something like that, would want to stay, work full-time, 77 percent or so of men want to work full-time.  I just have to think that that s something biological and women s attraction to nurturing, which has got to inhibit kind of the single-mindedness that some scientists have about not minding working 70 hours a week and so on.  I guess the other side of that, I think, is if you look at where the PhDs occur, of course, developmental psychology, women are two-thirds or so of the PhDs, mechanical engineering, not much at all, and I guess I wouldn t have predicted that the barriers of patriarchy would have been left here.  These guys with the pocket calculators, the stars, and so on, they re the real patriot, you re not coming in here, boy, you may take over the law schools with all these gladiators, and you may take over the med schools, you know, what could be more hierarchical than seeing, you know, parade around the doctor and then the resident and intern.  You might have guessed in 1960 that these are going to be the places that are never going to tumble, but, mechanical engineering, I mean, I just.  Okay, so that s kind of my two-prime question.  Elizabeth Spelke:  One thing that Dr. Geary and I don t disagree on is that biology, well, genetics, as well as ontogenetic biological processes are important for development, and one thing that I don t think anybody can disagree on is that there are biological differences between men and women.  The question is, do those differences play into differential aptitude for science?  Now, you ve appealed, in part, to differences between what men and women say they want to do, say how they want to spend their time, and it s, of course, altogether possible that genetic differences will influence those choices.  But we also know that cultural expectations and norms have an enormous influence on those choices, because, to see that, we only have to look at what people say in answer to that question has changed over the decades.  You pointed to examples with veterinarians.  It s changed enormously.  What people think of as possible lives for themselves is very influenced by what they see as appropriate, what they see as possible and open to them, and these are very open to change.  This doesn t mean that genetics doesn t play a role, but it means that the task of teasing apart genetic influences from other influences, is an extremely difficult one, and we would be wrong, I think, to conclude, at this point, from what people say today, that what they say is reflecting their underlying genetic makeup.  I want to also get to the plausibility argument that I think Dr. Geary made for why general principles of sexual selection might favor men in science.  I think the evolutionary work on sexual selection is wonderful and important, and I think it may apply to humans in interesting ways, but I don t get the implications for science.  It seems to me that if you took a Darwinian perspective and looked at human prehistory, you might wonder that anybody would ever do science at all.  The women should be devoting all of their time to nurturing their precious relatively few numbers of offspring.  The men should be devoting all of their time to fighting each other.  Nobody should be thinking about Fermat s last theorem.  The question is, how do, if these differences do condition how girls and boys develop, and how men and women choose to divide their time, how does that play out in the context of modern societies?  Even if it were the case, that men were, in general, more competitive, is there reason to think that greater competitiveness leads to greater science, particularly today, and a time of great collaboration and inter-disciplinary work, and so forth, that I think is less than obvious.  Even if it were to turn out that women were more focused on family, and, by the way, there is reason to think that the difference between men and women in competitiveness and the difference between men and women in valuing family is more, is not a black and white difference in absolute quantity -- more a difference in the kind of style of evaluation.  It s not the case that women aren t competitive, and it s not the case that men don t care about their children.  But, in any case, even if there were these general patterns, how does that impact on how people decide to spend their time, the kinds of intellectual problems that they find interesting and rewarding to look at?  That s where I don t see the connection to the evolutionary approach at this time.  David Geary:  A couple things, first, we re pretty sure there s a relationship between evolution and reproduction.  That, I think, has been established.  But, a couple things, one, with the Benbow and Lubinski study, some of the tradeoffs we see in these very high ability men and women, women make more of the family, social-oriented tradeoffs than men.  In terms of life satisfaction, there is no sex difference there.  So men are working more, on average, than women, and women are spending more on social relationships, including family.  But, humans are very interesting.  There s only about 3 percent of a million species where guys do anything in terms of care of infants, and investment, or, even protecting of offspring.  Humans are one of those species, which makes us very, very different from even chimps or bonobos, or other species.  So we have male paternal investment.  So the sex difference there is actually smaller than lots of other species.  We expect there to be less differences in humans than in other species.  When males invest in offspring, they become an important resource for females.  Women like guys who are going to invest, help out, and so forth.  We predict female-female competition, so we expect females to be actually more competitive with one another, and, generally, in humans, than we would, perhaps, in other species.  So, there s a lot of nuances going on there.  With respect to the interest things, I think what we would predict is that, more generally, I didn t talk about it today, but there is a pretty robust sex difference in interest in objects, mechanical types of things, at least that emerges during development, and interest in living types of things.  So, I think what I would predict is that women are going to segregate into more living-oriented sciences, medicine, psychology, vet schools, law schools, well, I guess law is not science, and men are going to go more into the physical sciences.  So, the question is, to get more physicists, do we have to take away from the number of medical doctors or number of vets?  There s a tradeoff there.  I think science is broadly defined.  I think there d be smaller differences, as you narrow the definition, I think that s where you re going to find more differences.  Christina Hoff Sommers:  Question here.  Male Audience Member:  A brief comment, and then a question for Dr. Spelke.  The first comment, tacking on to what Dr. Geary just said, is, given that the science, the largest by far, sciences, of the life sciences and the social sciences, there is no deficit of women overall, whereas in the physical sciences and engineering there is a tremendous difference.  It seems to me it really would be more productive to spend time, and have conferences, on the deficit of women in those areas where there really is a deficit, not just in science broadly.  But, let me get to my specific question for Dr. Spelke.  I found your anecdote about Dr. Boring very interesting, in which he assumed that certain Jewish characteristics of aggressiveness or being talkative or gesticulating wildly would make one unsuited for a career in science.  I m curious if you could point to specific characteristics associated with women that might, unfortunately, be assumed to be associated with success in science that we, in an irrelevant way, use in making those kind of ill-formed judgments.  Elizabeth Spelke:  Thank you, I d be happy to do that.  Let me give you two examples.  One is the example of spatial ability.  I think Dr. Geary and I would agree that the claim that males are better, in general, at spatial abilities is not quite right.  Rather, you can decompose spatial abilities into different kinds of abilities in the work that we ve done, sensitivity to different kinds of geometrical relationships, and attention to different kinds of geometrical information in a navigation task or other tasks.  And what you find is that males and females show different patterns.  I think there s a tendency, then, to ask, okay, which of these patterns make for better science?  Let s look at Einstein.  He says, I turn things around in my mind, so it must be that of all of those spatial abilities out there, it s the mental rotation that s really predictive of science.  I think that the mental rotation is predictive of science in the same way that not using your hands when you talk was predictive of science in the 1930 s in that it s typically associated with the people who typically are scientists, namely men.  But there s no reason to think if you start from the full panoply of abilities, and ask, how do they get used by students as they re learning math, as they re learning science, in school?  There s no reason to think that mental rotation is more important than any of the other abilities that show a different profile, many showing no sex differences, and a few of them showing a female advantage.  I think we tend to selectively focus on those capacities that are consistent with the idea that the prototypical scientist is a man.  Let me give you one more example.  The SATs have already come up in Dr. Geary s discussion.  There s an interesting phenomenon here.  The phenomenon is that girls do less well, both on average and at the upper extreme, on the SAT for mathematics.  The other phenomenon is that girls do better in school mathematics, both in high school, and in college they do equally well, as boys.  If you take SAT scores, and predict performance in academic mathematics in college from high school students SAT scores, what you find is that the scores under-predict the academic performance of girls, relative to boys, equating for institution and for difficulty of the math class, and also equating for ability.  So you take two students who get equally great grades in math at college, a male and a female, and go back to their SAT scores, the girls SAT scores will be lower.  Now, why is it that in discussions of mathematical aptitude we hear again and again about SATs, which show a gap, and not about the actual measures of mastery of mathematics in, as new material is presented in new classes, which don t show the gap?  I don t think that there s a deliberate or malevolent pattern of bias going on here.  I think there s a very natural psychological tendency to take what s typical of people in a group and emphasize that relative to other things.  Rosalind Chait Barnett:  I just add one point about mental rotation, there s a new study, it s actually, the history of studies like this, but it just came out from Canada, looking at mental rotation, and just to the long and short of it is that rather than assuming, because boys had an initial advantage in mental rotation that that was somehow a reflection of genetic or other kinds of innate differences, these researchers decided to see whether that was changeable, teachable, and they exposed kids who didn t have a prior advantage in rotation to 10 hours of training on a certain video game that was associated with visual spatial attention.  In 10 hours alone, they were able to reduce the sex difference remarkably, and that difference persisted over five months.  So the fact that boys and girls play different games, which is definitely the truth, there s a chance for boys and girls to develop differently in terms of their facility with mental rotation, but that can be overcome with specific training.  So I think reliance on that is somehow a reflection, a difference in the amount of rotation as a reflection of something genetic or innate characteristic is challenged by that and other studies as well.  David Geary:  A couple things, one is, to be successful in any area of science you have to have a lot of different skills, complex visual spatial skills, probably only one of them, and depending on the area of science that you re in.  The three-dimensional mental rotation is interesting because it predicts performance on the SATM.  And, related to Dr. Barnett s comment, other studies, in general, have shown that, yes, training improves performance, generally both males and females, overall, the gap tends not to change depending on the measures being used.  In the solving of complex mathematical reasoning problems, males spontaneously use more spatial strategies and representations of the problem than females do.  If you give females training on that, the gap in problem-solving performance decreases considerably.  So, yes, a lot of these things are changeable, but, I think spatial skills are related to some areas of mathematics and science, but, probably not all of them.  Debra Rollinson:  My name is Debra Rollinson, and as the infamous radical liberal who suggested we apply Title IX to our universities because they don t hire the women they train in science and engineering, I ve had a great opportunity to travel the country the last seven years, giving talks, what I call my uppity-women talk.  And as part of that, I like to challenge some of the myths.  One of the most pernicious myths, both Dr. Rhodes and Professor Spelke have already addressed, this single-minded 70-hour week, round the clock, 110 hours, you know, I ve worked longer, I ve worked all night, this pecking competition we have in science.  My favorite anecdote about that is with Rutherford, who popped into Cavendish Labs late one night in the early 20th Century and saw a student hard at work.  He said, were you at work this morning?  The young man, expecting to be praised for working hard all day, said, why yes, I was.  And Rutherford said, well, when do you think?  We haven t, and as a quantitative physical scientist, and, I especially appreciated the focus both Professor Spelke and Haier brought to their talks, we have to challenge these myths because we haven t done the control experiment, and we re starting to do the control experiment, working in teams.  You don t have to have people slaving around the clock, if you ve got a team you ve got a broader-based effort.  We honestly don t know if the science won t be deeper, better, more creative, if we actually open it up to new ways of doing science.  That makes it much easier not to be a man with an infrastructure supporting your ability to work around the clock.  I do have a, I d like Professor Haier to riff on something, and perhaps Professor Spelke to pop in.  It seems to me, again, as a physical scientist, that the brain cognition is a state function, and men and women, and men within the male varietal, and women within the female varietal, have different mechanisms to come up to the same piece of cognition.  In physical science, that is a thermodynamic function, and it really doesn t matter which part of your brain you re working to get there.  Richard Haier:  I can t remember the last time I was asked to riff.  I think I know what that means.  Our data, and other data, suggest this idea that you can have different brain structures or architectures to get to the same cognitive point.  I think this is a very interesting and exciting possibility.  Is it true?  I don t know.  We have to find out.  Brain imaging has only been around for about 20 years, and, it hasn t really been applied very much to what I would call individual difference issues.  I think it s exciting and I don t know where it s going to go.  Elizabeth Spelke:  Yes, just quickly, I agree with you that brain imaging work is extremely exciting, and I also agree with you that we have an enormous amount, still, to learn from it.  One point from your data though.  I think I m correct, tell me if I am, that, the data that you showed us do not show that there are qualitative differences in the brain regions that are recruited by men and women on these tasks.  They don t show, for example, when you don t see, colors lighting up in some part of the women s brains, that that area is not being used.  Rather, what they show is that in some cases the relative size, the variability in the size of the area, or the variability in the degree of activity of the area, is associated with differences in intelligence or performance in one sex and not in the other.  So you re getting differences in one case, and not in the other.  But that, in itself, doesn t tell us whether, when any man, or any woman, faces one of these problems, are they using different parts of the brain to do it, or are they using the same parts in ways that may or may not vary, as you vary measures of intelligence or performance.  Richard Haier:  It would be impossible to see that in structural imaging, you would need functional imaging, like the little animation I showed, and it is correct that whatever we learn about group differences cannot be applied to any individual so that let s imagine a few years from now, by the way, if you re the parent of a high school kid, you will appreciate this, we don t have the SAT anymore.  What you do in high school is, you go down to your local imaging center, lie in an MRI for about 8 minutes, take a nap, listen to your Ipod, and on the way out you get an MRI report that predicts how well you are going to do in certain subjects in college.  Is this far-fetched?  No, it is well within my imagination, and I think it is theoretically possible.  It is a matter of empirically testing this, and I would guess it would be impossible to get a grant to do this.  [Laughter].  And, moreover, the prediction equations would be different for men and women, perhaps, maybe not, it s an empirical question.  And, if they are, so what, what do you conclude from that?   Christina Hoff Sommers:  I d like to ask a question of the panel about the National Academy of Sciences study.  I think, some years ago, some researchers contrasted, they said that the literature on sex differences tends to have an alpha-bias or a beta-bias, and you can usually identify it right away.  There are people that absolutely, the alpha-bias, they exaggerate difference, and then people with the beta-bias, they deny, explain away difference.  And you can usually tell the literature on sex difference if you re dealing with an alpha or a beta.  The popular culture loves to exaggerate difference, and it s all over, the men are from Mars, women are from Venus sort of school of thought.  However, there are wonderful exceptions, our panel today, getting a rich variety of perspectives, and there are two recent papers, a paper and a book, one is called, The Science of Sex Differences in Science and Mathematics, which came from the Psychological Science and Public Interest.  It just came out.  It s a wonderful consensus statement.  The best scientists, many that are in this room and are on the panel, came together and just agreed on what we can say that we honestly know.  It wasn t always easygoing.  They had to sort of cancel out one another s bias.  I put a copy of the summary statement in your booklet so you could read the report because it s really quite, quite fine.  And there s also a book called, Why Aren t More Women in Science, and it s just come out, it s called Top Researchers Debate the Evidence.  We have several of the authors here, I m happy to say.  And, again, it s just remarkable in the way that it represents the different points of view, and you just come away enriched, and, no matter what, if you have an alpha or beta-bias you come away a little uncertain and it gives you pause.  It s just so different from this National Academy of Sciences study.  Oh boy, this is so self-assured, it s ready to move, take action, fund, workshop, everybody is going to workshops to be re-educated on gender.  It just seems it s way ahead of the science.  The tone is so different from these.  This is totally beta.  I just don t understand.  To me, I m mystified how that happened, and maybe you agree that it did happen.  Elizabeth Spelke:  Can I respond to that?  I think that the National Academy s report is answering a different question from the question that is being answered by that book and the article, which I m eager to see, I haven t seen that yet.  The question being asked by the National, the question being asked by the report is not, are there differences between men and women?  The question that is being asked is, are there barriers and bias to the progress of women in science?  The evidence that bears on that question comes not primarily from studies of cognitive development or motivational differences.  It comes from studies of institutional behavior that look at, for example, the composition of the talent pool and who gets selected, that look, as in the study you were talking about, at the criteria for selection.  Those are the points of departure.  The analyses and conclusions in that report are consistent with either an alpha or a beta view on sex differences.  You could believe that the sexes are very different, and that science will be most enhanced if the most qualified scientists of either sex have equal access to the institutions and the resources, or, you could believe that the sexes are very similar and come to that conclusion, or maybe a different conclusion.  But I think the questions are different, and that s why, that accounts for the difference in the focus of the report.  Richard Haier:  I must have read a different report.  I mean, the National Academy of Sciences report comes to very bold assertions about sex differences and cognition, if they exist at all, don t matter, and we don t need any biological research on this.  So I think if the report said what you just said, fine, but I don t think this report says that.  You re nodding your head yes.  Joshua Aronson:  As a person who s research is heralded in that NAS report, I feel very uncomfortable because I don t think we can be that confident about the biases holding one back.  I thought your point was excellent, that, you know, [cross-talking].  Elizabeth Spelke:  [Inaudible].  Joshua Aronson:  Oh, I m sorry, I m Joshua Aronson.  Christina Hoff Sommers:  He ll be on a panel this afternoon.  Joshua Aronson:  I thought your point was excellent, why hasn t medicine shut its doors to women?  I think that there s, I don t believe that the biases are holding women back so much, perhaps self-imposed biases, but I don t think that s the story.  And it s too confident.  It s simply too confident.  Christina Hoff Sommers:  Yes, it is too confident, on this topic, but, maybe you don t think so, other audience members?  And say who you are, please.  Constance Holden:  Constance Holden at Science.  I just want to ask Dr. Barnett and Spelke, since everyone on the panel seems to agree that boys and girls like different games, and we know we tried to make girls like trucks and boys like dolls, it didn t work, why isn t it logical that boys and girls would like different scientific games when they grow up?  Rosalind Chait Barnett:  I can start by answering that just briefly, we can talk about it more later, because other people might not be as interested in the details, there s a lot of evidence that suggests that the way in which toys are marketed and packaged affects children s preferences for those toys, and there s some dramatic data showing that if you take a gun and holster, which is a typical male gender toy, if you modify that gun and holster so it s now made full of rhinestones and colored purple, boys and girls both agree that it s a toy for girls.  There s also very early influences of parents perceptions of how good or how appropriate it would be for boys and girls to play with cross-gendered toys.  Kids are aware of that very early on and it remarkably shapes the time that they will play with toys they find equally attractive but that have been gendered.  So I think that s an extremely cultural and says more about that than innate differences.  Elizabeth Spelke:  Could I respond to that also?  I actually think that we don t know to what degree differences in the expressed preferences, either of children or of adults, are due to genetic differences and to what degree they are due to culture.  I do think, though, that we can do better than simply look at spontaneous preferences for one or another kind of manufactured toy in asking, what are the core capacities that need to be nourished to do high-quality science?  We can do better by looking at the actual cognitive processes that go on when children deal with objects or with number or with space.  These are highly abstract activities, and it s far from obvious that the degree to which you would enjoy waving a gun around is going to influence any of them.  So that was why I tried, in my talk, at least, to focus primarily on getting at the underlying cognitive abilities and not the preferences that you see either in a child or when you simply ask adults how they like to spend their time.  Male Audience Member:  I have a question for Richard.  Christina Hoff Sommers:  Oh, sorry, did you just want to say one thing?  Male Voice [Richard Haier?]:  The preferences are, I disagree with Dr. Barnett, the preferences are very well established, they re very large, they have a heritable component to them, and they are influenced by prenatal exposure to male hormones.  That said, we don t know what the downstream or upstream influence of those are.  We can find out.  We need to acknowledge these differences and say, okay, is there something there, and then let s find out what is going on downstream, if anything.  It may not matter.  I m suggesting a hypothesis that it reflects in inherent interest bias that may have a downstream effect.  Male Audience Member:  My question for Richard is, supposing that women and men are recruiting different parts of their brains to do the math problems, what s your hypothesis for why that might be?  Richard Haier:  I would turn to an evolutionary person to speculate about that.  I thought you were going to ask me what s the implication of that for education.  That s an interesting problem.  I m actually going to give a talk about that next month at the National Academy of Science Symposium.  I ve been doing a lot of thinking about that.  One of the difficulties in applying any neuroscience or psychological result to education, well, there are a number of difficulties.  One, the results tend to be over-generalized, and may have no implication for education at all in a practical way.  What I think is a more difficult problem, how many of you are educators, by the way?  Quite a few.  I am worried that we don t have enough people in the education field who are sophisticated in statistics and science to think about the kinds of studies that we re talking about, make some judgment about, you know, which ones might be relevant, and then design programs and experiments to see if they are relevant.  I m worried that we re going to go through a terrible period in education that will, we ve been through a terrible period where a lot of fads that have originated in kind of the fringe science areas have taken hold in education, wasted an awful lot of time and resources, and proved not to be effective at all.  So, the implication that men and women may recruit different brain areas is not important as the concept that each person, male, female, it doesn t matter, that each person needs to be evaluated in some way so education can be optimized for that person.  I m a little concerned that we spend too much time talking about men and women and how that generalizes.  Christina Hoff Sommers:  Can I just go to the audience because there are so many, you ve had your hand up for a long time.  Yes.  We will have time this afternoon for more questions.  Lois Tett :  Thank you.  I am Lois Tett, independent writer and from Majority Women on Businesses.  It was said in the conference that the brilliant years of a baby being born is from 0 to 5, and that some cultures begin educating or talking and communicating with that baby when it s in the womb.  If we accelerated when the baby is born, we ve been talking to him, he will be smarter, and we expose him, early, very early on, from various toys, games, etc., not just, say, dolls for girls, and trains for boys, but things that are mind-expanding, I would think, it was suggested that we d have more scientific minds coming out of this.  Christina Hoff Sommers:  Good comment.  Do you agree with that?  Richard Haier:  There s no evidence in terms of empirical studies, but it wouldn t hurt to do those things.  Christina Hoff Sommers:  Yes.  I m neglecting that side, are there people back there, I can t see.  Dr. Jacquelyn Black:  I m Dr. Jacquelyn Black, and for the last 12 years I have served on the American Society for Microbiology s Committee on the Status of Women in Microbiology.  We have done a lot of investigations as to how our microbiology interested women are doing.  In academia, where I also function, coming in at the instructor level, we see 80 percent female and 20 percent male.  This is a definite interest.  Talking about where you get your interests, it doesn t seem to help.  They come in, 80 percent.  What are the results when we reach full professor?  It s, again, 80/20, 80 percent male, 20 percent female.  What has happened between the 80 percent of interested females who successfully got their PhDs and got started in academia, and now they have dwindled to 20 percent at the full professor level?  This, I think, is very, very crucial, and perhaps more so than the interest, because the interest does exist here, but the glass ceiling closes in.  Christina Hoff Sommers:  Yes, in the back.  Mary Juhas:  I m Mary Juhas.  I m from the National Science Foundation, Directorate for Engineering.  I have a PhD in material science.  I work on aerospace materials.  It s a very boy-oriented area.  When I was young, I m the eldest of three daughters, and when I was going through school, what the numbers showed was that the women who tended to go into the mechanical engineering, materials engineering, electrical engineering, very male-dominated areas of engineering, came from single-gender families, or they were the eldest child.  I fit right into that.  I had guns, and, I did have guns.  [Laughter].  But I had Barbie, and I had a baseball, a catcher s mitt, and my sisters and I played that.  And, by the way, my youngest sister graduated from the Air Force Academy in the first class to admit women.  So there s your data there.  What I would like to talk to you about and get your reaction about is, the public understanding of engineering.  Girls are not brought up to think that engineers help people, that we make the air cleaner, the water cleaner.  More people lived on 9/11 because engineers made those buildings safe and structurally redundant.  They don t know that we build drug delivery systems that keep Grandma and Grandpa alive, that we do tissue scaffolding and all these other things.  What happens is, at a very young age, boys and girls are taught that you have to be good in math to be an engineer.  Now, I was told in 7th grade, that I would never, ever, do anything except become a forest ranger or something, because I didn t have math skills.  I just worked my patutie off.  Then when I found linear algebra I realized that I could do the spatial thing too, just as easily as the guys could.  So, this is really my plea from NSF and from all engineers, both men and women, is, to get the word out there that engineers help people.  I really believe that this is a way that we will be able to cast a broader net, that we engineers work in teams.  Girls do very well in teams.  They don t know that until they come through the door in engineering classrooms that you work together, and you re not Dilbert, working in isolation.  [Applause].  Christina Hoff Sommers:  Yes.  Ruta Sevo:  I m Ruta Sevo, independent writer.  I appreciate all the information we ve gotten, and the emphasis on cognitive psychology and biological differences, brain differences.  I think the disjoint in addressing the report from this perspective is the giant leap being made from innate skills at which there are not many differences, and linking those to occupational choice and interest.  There are a lot things that happen along the way.  I would say that what happens is social cultural effects that constrain and guide choices regardless of innate ability.  I don t think you could take IQ and SAT scores and link them to certain occupations.  I think one of the biggest barriers to women in science and engineering is completely independent of those fields.  It has to do with societal constraints on the fact that women have to do two jobs, that the nurture has been delegated to women and it hasn t been supported, and, so, almost every occupational choice you can look back and say, why are women in veterinarian science as opposed to others?  It s because the conditions of work allow for those two jobs to exist.  So if you took away the nurture obligation and said, okay, everybody, society will take care of families, now go into the field that you naturally feel you re oriented to, you ll probably find a much different distribution, and cross-culturally, you see that happen because there are different supports for different occupations and then different preferences.  The guidance for career interest and choice really is not tied to biology and IQ or SAT scores.  It s actually tied to parental influence, educational experience, and then the other experiences that come along between the BA and the PhD, and those are societal.   Christina Hoff Sommers:  Let s just get as many comments as we can.  Yes.  Phoebe Leboy :  Phoebe Leboy from the Association for Women and Science.  I d like to pick up on that and ask a frank question.  Given the data that you all have presented on differences, yes, they may exist, but they seem relatively modest, is there any of you who are prepared to say that these kinds of things we re looking at should have an appreciable affect on the ability of women to go into any scientific field, and, if so, could you explain what they are?  How are you defining their ability?  Simon Baron-Cohen:  Simon Baron-Cohen, and you ll be hearing from me shortly, but, just picking up on this point, I agree that one way of summarizing what we ve heard is the differences between the sexes in cognitive tests are quite modest when you re doing population studies, they are small differences, they re quite subtle differences.  One question is whether it s the same size of difference when you go to the extremes?  So, most of the studies that have been considered are looking at groups of males and females, and we find small differences, does that make a big difference to later occupational outcomes?  But, Richard mentioned Daniel Tammet, who had this extraordinary memory where he could memorize Pi to 22,500 decimal places.  These are people at the extremes.  One way to look at this, is, if we took whatever the relevant dimension of aptitude we are interested in, some people would say is it mental rotation, and some aspect of maths, if you look at people in general, comparing averages, there may be very small differences, if you re just looking at the extreme tails of those distributions, we may see something very different.  So if we take a very uncontroversial example, like height, in the general population males are a little bit taller than females.  It s only a couple of inches and it doesn t probably make a whole heap of difference.  But if you look at the extremes, if you go up to the sex ratio of people at 5 10 , there re 30 males to 1 female.  This is a point from Steven Pinker, if you go two inches further on that extreme, we re just talking about height, the sex ratio of people over 6 tall, it s 2,000 males to 1 female.  So, it s past the nature of the normal distribution, that once you start looking at the tails of the distribution, a small difference in the means can make a big difference at the tails, and part of what we re interested in, is, people who are at the tails of the distribution.  We re not talking about people who are just applying to university to get into science, we re talking about the Nobel Prize winners, and the people who are making major contributions.  I m very interested by the fact that if you take the Olympiad winners in mathematics, you know, what is the sex ratio there?  They are usually selected at 16 years old, so it s not as if it s about other distractions, like family pressures and nurturing, and so forth, these are at 16 years old, what s the sex ratio of people who are winning the mathematics Olympiad in different nations?  If we take the Field Medal, which is the equivalent to the Nobel Prize in mathematics, it s been running for a century, as far as I know there hasn t been a single female winner.  So this is the prize given, is that true, not true?  Female Audience Member:  A woman won it one or two years ago.  Simon Baron-Cohen:  Okay, so there have been one or two, and the question is, these are the extreme measures, if you like, and, is there something different happening at the extremes, which might help us understand some of the discrepancies that we re talking about in occupational ratios in different sciences.  Female Audience Member:  But if you take the Intel Science [inaudible] for the last two years, five of the last six, you wouldn t.  Christina Hoff Sommers:  Yes.  Female Audience Member:  Back to me again.  When I was [cross-talking].  Christina Hoff Sommers:  Oh, I shouldn t have done the same person twice.  Can we give it to someone else?  I m sorry to do that, I m getting confused.  Yes.  Amy Wax:  I m Amy Wax.  I ll be speaking later this afternoon.  I also want to put in a plea for a little bit more clarity and less muddle as to whether we re talking about averages or the right tail of the curve, because variance is a huge amount of the story and in what I ve read these two aspects of gender difference become conflicted whenever it s convenient for one side or the other to focus attention on average, as opposed to variance.  But I guess, I want to make a comment, and then ask a question.  My comment is, I listen to some of these panelists and others in the room and I really wonder whether I ve been living in an alternative universe.  I went to college in the early 70s and I started out in science.  I never saw any evidence that women are discouraged from going into science, certainly in elite circles quite the contrary, there are just a huge amount of encouragement, I think, for women, and many of the women of my acquaintance who were of very high ability as an undergraduate at Yale and in graduate school at Harvard, I think went, continued in science far longer than they really wanted to because of all the, now this is just anecdote, I realize, and I m kind of tired of anecdote in the field, so I ll just leave this and I ll go to my question.  There seems to be this idea that there aren t really any, there s no difference, significantly, in the abilities or the temperamental aspects of men and women that were to count for their wildly different success in the field, in sort of the advanced areas of the field.  It s got to be either discrimination or some kind of internalized expectation that is holding women back.  I detect a certain circularity.  I mean, if there are any outcome differences, this is taken as proof that there must either be discrimination, or some kind of internalized disability.  So I would say, what, apart from equivalent outcomes and success, would count as evidence that, in fact, there are self-imposed or self-selective based or ability-based differences that account for disparities in outcome?  I mean, what would count in people s view as bona fide evidence?  Christina Hoff Sommers:  Good question.  Elizabeth Spelke:  Thanks, I want to address, actually, both parts of the, first, your comment on extremes versus averages.  The reason I put up two graphs for each of the abilities I talked about, one was looking at average performance, the other at extreme performance, was for just that reason, we didn t, when we test for core abilities we don t see differences either on average or at the extreme.  To your question about what counts as evidence for discrimination, I completely agree with you and think this is consistent with what s written in the report, that the simple fact of seeing different proportions of men and women in a field, does not, in itself, tell us that there s discrimination.  The evidence for discrimination comes from audit studies where the same resume is sent in with a male versus a female name, from analyses of the behavior, looking at what are the performance criteria that are applied to women or men, how much do you have to publish to get a fellowship in the example that was talked about today.  That s the evidence for discrimination.  Then, the question, what causes it, is the question that s addressed by these further studies that ask, what do people want, how are they performing on tests, why are they making the choices that they re making, and so forth.  But, I completely agree, the evidence, and as the report states, comes from the objective measures of probability of success as a function of gender.  Amy Wax:   I too have looked at the report.  I ve looked at some of these studies.  I think that what isn t being discussed here today is both the paucity and the questionable quality of the very few studies that are being relied, the [indiscernible], which has no controls for quality, it d be just a bean counting study.  [Cross-talking].  It doesn t.  It s just how many papers did you publish in top journals.   Elizabeth Spelke:  No.  It takes account of the impact factor, it s not just a number of [cross-talking].  Amy Wax:  But the impact factor is quantitatively assessed, it has not been sent out double-blinded and sex-blinded to a committee to say, how good is this resume.  That did not happen in that study, and it s one study, and there are other studies in the literature that show with equivalent resumes the chance of getting tenure when sex-blinded is the same.  So there s a lot of equivocal material on this, and actually less there than meets the eye.  I mean, that s, I think, part of the problem here is that we have kind of recourse, we are falling back on the idea that if men and women make different choices, which they clearly do make radically different choices of career, of life course, and the like, and there s a whole bunch of studies that never get cited.  I ll give you one example.  There s a study out of Georgia of a very large sample, I think it s a national longitudinal survey that shows that the most intelligent women disproportionately drop out of the workforce.  Why?  Assortative  mating.  They married the most successful men.  So, the population of women who work and the population of professional women is disproportionately enriched for a somewhat less able group of people because very high ability women are staying home.  Now that s the sort of evidence that s out there, that never gets cited.  I mean, it s quite interesting that there s a kind of selectivity, I think, in the studies that people do talk about.  Christina Hoff Sommers:  We are just about out of time, and, one last question/comment from Susan Pinker.  Susan Pinker:  Yes, I have a question, I think, for Rosalind and Elizabeth.  I m a little bit flummoxed about the hand-wringing that s going on about women and science, because so much progress has been made in about 30 years, and I think there are more areas where there are more women now than men, than the reverse, so, not just veterinary science, but in pharmacy, in clinical medicine, as well as research medicine.  And, so, I can t, what I can t figure out, is, why would there be barriers in science in some fields and not in other fields?  That kind of, that s one question, and, related to that question, is, that, if women feel they can t go into fields because they don t see other women there, then why do we have, for example, so many, in fact, more female law students now than male, more female medical students now than male.  Why would they go into these fields in the first place if they never saw women there?  Rosalind Chait Barnett:  Well, beginning, this is a very big question, I try to show that women do better in flatter organizations, like these biotech firms, than they do in hierarchical organizations, and that has largely to do with the fact that people who comment on their ability for purposes of promotion and advancement, there s a greater range of people so that the women are less likely to be evaluated on the basis of stereotypes about women s ability in organizations where they have more exposure.  That s just a partial answer to your question.  Christina Hoff Sommers:  We really, we have to pause for a break, and then lunch, and one sentence for Mark from Professor Haier.  I ve given him permission.  Mark:  It s my understanding that the National Academy of Sciences is working on a second report on this subject that will be, quote, more empirical.  Christina Hoff Sommers:  Ah-ha.  So that s a hopeful ending to this session.  Thank you.  [Applause].  Lunch is served around the corner.  LUNCHEON KEYNOTE SPEAKER:  Simon Baron-Cohen Christina Hoff Sommers:  Sorry to break up these conversations.  Attention!  I think we re going to have to get started with the luncheon speech, if I could have your attention, please.  Our next speaker is Professor Simon Baron-Cohen.  He joins us from Cambridge, England, where he is at Cambridge University, where he is a Professor of Developmental Psychopathology, and Director of the Autism Research Center at Cambridge.  He s the well-behaved Baron-Cohen.  [Laughter].  Steve Pinker has called him one of the most brilliant, I m not going to gesticulate, I m thinking of Professor Boring, one of the most brilliant research psychologists of his generation.  I certainly agree with that.  I am honored that he can join us today at AEI.  Please join me in welcoming him.  Thank you.  [Applause].   Simon Baron-Cohen:  Well, I m sorry to interrupt your lunch, and I hope you won t hold me responsible for any adverse effects in digestion or otherwise.  I m going to talk for about 40 minutes, and, so, I feel a little bit indulgent, really, to be given more time than the panelists this morning.  So, the title of my talk, as you can see, is not just about sex differences but the link with autism.  This slide simply just acknowledges that the work I m going to be describing is the result of collaboration and, I think there may be a pointer on this thing, let me just see, there we go, so we have one group of scientists who are psychologists, and also doing brain scanning, and a second group who are biochemists and involved in looking at the role of hormones.  I ll be describing a particular study, which is a longitudinal study looking at the role of fetal testosterone measured during pregnancy and where we follow-up the children after they re born to see if there s any relationship between prenatal hormone levels and later aspects of child development.   But, this was the point I was making in the discussion earlier.  This comes from Steve Pinker.  When we look at populations, and if you find a difference in the means, which, as you all know, right there in the center, that might look like a very small difference, and the question is, what s going on at the tail, and Steve has very use for a little magnifying glass to say it s actually not very easy to see what happens at the tail, but, one of the statistical properties of the normal distribution is that the distribution falls off according to the negative exponential of the square of differences from the mean.  So, he then, very usefully, gives us, zooming in on what s going on in the tail, and what you can see, putting it into plain language, is that something that looked like a very small difference in the middle of the distribution, by the time you get to the end the difference has really become much more divergent.  I gave you the example earlier of height.  At 5 10 , let s say we were just sampling at this point, you have 30 males to 1 female, just two inches more, so you re just sliding along that X axis, two inches further and this X ratio has gone to 2,000 to 1.  So, it just means that we have to really think carefully about what the ramifications would be for small differences at the center of the distribution.   Okay, so, why would anybody in their right mind make any link between the study of sex differences and autism?  On the face of it, these seem to be two very different sorts of topics, and, while one reason is that autism, which you ve all heard of, and this medical condition affecting social development, but also involving very narrow interests, so-called obsessional narrow interests, affects boys much more often than girls.  If you take classic autism it s about 4 males to 1 female, Asperger syndrome, the high-functioning subgroup that we hear a lot about these days, 9 males to 1 female.  So, something about being male affects the likelihood of receiving a diagnosis on the autistic spectrum.   I also want to, and, just sort of give advanced notice of the topic we re going to be discussion later, which is about hormones, and male fetuses produce about twice as much testosterone as female fetuses.  So, it s happening very early in development.  But, if you look at the individual differences within one sex, you can find a 20-fold difference, which is a lot of individual differences, and the consensus amongst developmental neurobiologists looking at these sex steroids, like testosterone, is that they may have organizational effects on brain development.   Tying this back to autism was the observation from the pediatrician, Hans Asperger, whose name is now given to that subgroup, he was writing in Austria in the 1940 s, and, in his clinical, anecdotal descriptions he came up with this observation -- that the autistic personality is an extreme variant of male intelligence.  So, part of what we ve been trying to do in our research, even though we re straying into quite a politically sensitive area, is, in order to understand autism we have to also understand something about the nature of typical sex differences.  Which leads me on to typical sex differences, and, you heard earlier about sex differences in the brain, and, in this slide I ve simply summarized a small number of the structural differences you can find in the size of different brain regions between males and females.  I m not going to go through each and every one of these things, but, suffice it to say that if you looked at regions which are larger in typical males than females, you find differences.  Let s just pick out the amygdala, in the limbic system, sometimes people refer to it as the emotion center, structurally larger in males than females.  And, again, there are a bunch of differences in brain structure in different regions, where, this time, they re larger in females than males.  And, again, just picking out one on the list, the corpus callosum, you can see it in white, just in the center there, it s the connective tissue between the two hemispheres, and, down below, you can see it s been magnified to show that it s the posterior section of the corpus callosum, which is thicker in females than males when you do these group studies.  So, I could have produced a larger list, people who are interested in this, there are good sources now, showing that the sexes differ in terms of neuroanatomy.  The question is, do those things have any relevance, do those differences have any relevance to brain function and psychological differences?  So we ve had a bit of discussion about psychological differences between the sexes.  This slide is really summarizing literally hundreds of studies by child psychologists, looking at spontaneous choice, that if you leave toys out on the carpet, videotape the children, and later look at those videos to see which toys do they spontaneously choose to play with, what you find is that between the ages of about 1 and 5 years old more boys than girls will spontaneously choose to play with toy vehicles and constructional toys like LEGO, and more girls than boys will spontaneously choose to play with dolls, but, particularly creating social stories, often with an emotional content.  So, on the right side of that slide, you can just see an example of a dataset from Melissa Hines s work where she puts out the so-called masculine, typical toys, like toy vehicles, and you can just see the differences in the amount of time that girls and boys play with these male-typical toys and female-typical toys, like dolls, girls showing more interest, more spontaneous contact time with dolls than with vehicles.  So one interpretation of these spontaneous toy studies, looking at children at play, is to say that maybe early on in development little boys are having their interests, their attention, attracted more easily by systems of a variety of kinds, whether they re mechanical or constructional, and that maybe little girls, early on in development, are having their interest focused much more on people and on the emotional lives of other people.  Bear in mind that these are just group studies, the findings don t apply to every individual of each sex, we re just seeing differences on average.  Now, this is an interesting slide to interject because here the same paradigm, spontaneous toy choice, has been used with another species, in this case vervet monkeys, where they put out the toy cars and the toy dolls and look to see whether the two sexes in this species show any differences.  From Alexander and Hines data that was published a few years ago, you find that the male monkeys spend more time pushing along these toy cars, and the female monkeys spend more time examining and interacting with these dolls, which just hints at the possibility that even though culture is important, we heard earlier about the toy industry targeting our kids in different ways, and the effects of cultural experience, it may be that biology is also an important factor to include.  Now we get to the data to do with adult human occupation and sex ratios.  Broadly, what s been found is that there are a whole cluster of occupations where males outnumber females, mathematics being a very clear case.  If you look at higher levels, at university level, in most cultures where it s been studied, the sex ratio is about 13 males for every 1 female.  But, computer science, physics, and engineering are not far behind mathematics, tool-making being about three-quarters male.  And then there s a second cluster of occupations, primary school teaching, counseling, where you see the opposite sex ratio, more women than men.  Here we have to be much more careful about talking about spontaneous choices because we don t all have the luxury of choosing our occupations.  There s a lot of different factors that can determine which field a person ends up in.  But, sitting sort of side-by-side with the child toy data, it s tempting to say that in the adult workplace we re also seeing the same pattern, namely that males are ending up working in fields which have to do with systems of one kind or another and females, again, on average, are more likely to end up in fields to do with people, and, particularly, the emotional lives of other people.  You can see on the right of that slide, the graph you saw earlier, showing SAT math results from 1972 through 1997, males and females, and with, although fluctuations year-by-year, a very typical pattern of sex differences on the SAT math test.  So, what I m going to do next is just tell you a bit about some research we ve been doing looking at two psychological processes in terms of sex differences, and its link to autism.  The two processes are empathy and systemizing.  I m going to start by just defining these.  Empathy, we know what that means, it s a word in the English language, but I ve defined it here as the drive to identify another person s thoughts and feelings, so to put yourself into someone else s shoes, to imagine their perspective.  But, it s also got that second component, namely, to respond to somebody else s thoughts and feelings with an appropriate emotion.  It s quite important to have both components in empathy because if it was just about recognizing what someone else might be feeling, we wouldn t necessarily want to call that empathy, after all, a psychopath might be very good at spotting that their victim is in pain, but not have the appropriate emotional response.  So, empathy, presumably, has to have this affective reaction as well.  Systemizing, a very different kind of process, the drive to analyze a system, any kind of system, or to build a system, and what you can see I ve listed there is the common types of system that are all around us in the environment.  Mechanical systems, like computers and car engines, abstract systems like mathematics, natural systems like the body or a plant or the weather, and even collectible systems like your collection of CDs at home, which may be more or less systematically organized. And, basically, what systems seem to show is that they follow rules.  So when you try to systemize, when you re trying to understand how that system works, you re really trying to identify the rules that govern the system so that you can predict how that system works.  Are there sex differences in empathy and in systemizing?  Well, this is one test of empathy.  It s a test of individual differences in the adult population, where you look at photographs of faces and you have to judge which of these four words around the photo best describes what the person in the photo is thinking or feeling.  So, it s quite a tough test because all you have to go on is the region around the eyes, subtle emotional expression in the fact.  They are black and white photos, quite degraded in quality, and all four words describe possible states of mind.  Whilst I ve been talking, I m hoping that you ve been looking and making your choice and agree that she looks a bit dispirited.  This is a test that we ve given to men and women in the general population, and what we find out of a maximum of 25 such photographs is that women are scoring higher than men, it s a small but statistically significant difference, when you compare groups of men and women.  We ve also got a group of people with Asperger syndrome, who are scoring much lower than the other two groups.  This is, of course, no surprise, since part of their diagnosis is that they have difficulty with empathy, with reading emotions.  But here we have some quantitative confirmation of their disability. Part of what we can take away from this one study is that it s confirming, in a quantitative way, that women are slightly better than men, as a group, in reading emotions in the face, especially of a more subtle nature, and that people with autism, or, on the autistic spectrum, may be showing an extreme of the male typical profile.  Here s a different kind of test, which we can think of as much more relevant to systemizing.  In this case we ve taken a physics test, all the items come from a test which is used to select potential engineers so that people who haven t studied engineering yet, but they re being looked at for their possible aptitude into engineering.  We ve just given these little problems to people in the general population, men and women, and you can see that adults in the population, men are scoring higher than women on this test of what s sometimes called intuitive physics.  We ve also given it to children with Asperger syndrome in a comparison group.  You can see that children with Asperger are scoring higher than children without a diagnosis.  This is all the more striking when I tell you that the control group are teenagers, whereas the group of children with Asperger syndrome are aged 8 to 11 years old, so they are in primary school, being compared to an older comparison group, and yet out-performing that group.  So, despite their disability, in psychiatric terms, and in social terms, they re actually performing at a higher level on this test of systemizing.  This is a different kind of psychological test, which many of you will have seen because it s been around for about 50 years.  It s a favorite in psychology.  It s called the embedded figures test, where you have to find this target shape as quickly as you can in the overall display, up above, and we think of this as a test of attention to detail, which is presumably a prerequisite for systemizing, because, to understand the difference between one system and another you really have to pay attention to the small details.  If I trace out the solution there, you can see the front of the cube, in yellow, and then you can see it receding into 3-D, and this is a timed test, and what you can see here is the average number of seconds that it takes people, men and women, to find the target shape.  So, this is actually back to front, that this is, this should be the female data and the male data, it s just a typo, that males are showing faster speeds at finding the target than females.  Here we ve got the group of people with Asperger syndrome, who were even faster than both other groups.  So this is, again, giving us some evidence of sex differences when it comes to attention to detail of at least a physical design, and people in the autistic spectrum showing an extreme of the male profile.  Now we can drop down one level to look at brain activity, because this is functional magnetic resonance imaging, whilst men and women are doing those two tasks, looking at the photographs of the eyes and looking for that target shape in the embedded figures test.  So now we re looking at whether the two sexes are using different parts of the brain.  If we start off with that test of empathy, the eyes test, what we re reporting here is regions of the brain, which show greater activity in women than men.  You can see that bilaterally, in the medial prefrontal cortex, women are showing more activity when they are decoding somebody else s facial expression to infer another person s state of mind.  On the embedded figures test, women are showing more activity, this is the back of the brain, in the visual cortex, again, bilaterally, more activity whilst they are trying to find the target shape hidden in the larger design.  So the sex differences we see at the psychological level are mirrored by sex differences that we see at the level of brain activity.  I should say that, I haven t shown it on this slide, that when you give people with autism or Asperger syndrome, brain scans whilst they re doing those two studies, on the eyes test women were showing more activity in this area, prefrontal, medial prefrontal, and that s exactly the same region that people with autism show under-activity, compared to controls.  And, on the embedded figures test, when they are performing at great speed finding those small details, they are also showing less activity in those visual areas.  So I can pause now to present you with a model, which is that I ve been talking a little bit about two dimensions in the population, empathy and systemizing.  This relates back to the whole discussion about normal distributions and what happens at the tails.  If we imagine that zero means you re exactly average for the population, and on empathy, as you go up this way, you re above-average, and down this way, below, and on systemizing, out this way, you re above-average, and, across this way, you re below average, what we ve been finding in a range of different psychological tests, is that more males than females seem to be in this pink zone on the graph where their systemizing is slightly higher than their empathy, and more females than males seem to be in this light blue zone where their empathy is at a slightly higher level than their systemizing.  These are just group studies, as I ve said.  There are lots of people who are in the white zone who show no discrepancy between their systemizing and their empathy, they are as good, or as bad, in one domain as in the other.  What we ve been finding is that people with autism seem to be down here, in the dark red zone, that s to say they are below average on empathy, but their systemizing might be anywhere from zero, that s to say, normal, through to above-average.  Well, that s just the model, the question is how well does the model fit the data, or the data fit the model?  One way to do this, to test the model, is to have standardized instruments, which can measure empathy and systemizing in a parallel way.  So we ve used questionnaires.  The empathy quotient, where you go through the questionnaire, and you say whether each statement, whether you agree or disagree with that as a statement about you, what you can find is that women score higher on this empathy quotient than men, people with Asperger are scoring even lower.  So that same pattern we saw on the performance tests is coming out in self-report.  But we ve also got a systemizing quotient, the S cube, which asks you how interested you are in a variety of systems.  You can see one is a mechanical system here, another one is an electrical system.  You just say whether you agree or disagree with many of these statements.  Now we ve got the opposite pattern, it s just a self-report instrument, but males are scoring higher than females in their self-reported interest in a range of systems.  People with Asperger are scoring even higher than the other two groups.  So now we ve got data on individuals who have filled out both questionnaires, and now we can plot them, so each dot you see in this scatter plot is a person who has taken both the empathy quotient and the systemizing quotient.  It s a very busy slide, but if I talk you through it, we ve got women in red, and we ve got men in blue.  Here is empathy, and here is systemizing.  And then we ve got people with Asperger in green.  Just eyeballing the data, you might agree with me that there are more red dots up here, and more blue dots down here, and then more green ones down here, suggesting that there may be differences between the groups, that it s not a sort of random distribution.  Very important, you can see the title here is that we are plotting individuals according to whether they are showing better empathy than systemizing, or better systemizing than empathy.  Each individual, we really need to just regard them as individuals.  So we ve got some men in blue here, who are indistinguishable from a lot of the women in red in terms of their good empathy being better than their systemizing.  Equally, we ve got some women down here in red, you can see, who are indistinguishable from many of the blue individuals, the males, in terms of their systemizing being better than their empathy.  So, it s very important, once we start looking at individuals, to realize that you can t predict the nature of someone s profile, their brain type, based on their sex.  That would actually be stereotyping.  You really need to look at the individual to see where they lie, and if we now, not just eyeball the data, but count how many men and women fall into each of these profiles, empathy greater than systemizing, we do see more women than men, 44 percent versus 17 percent, in a general population study.  The opposite profile, systemizing better than empathy, we see more men than women, 54 versus 17.  And now the extreme of the male profile, where your systemizing may be intact or even above-average, but your empathy is below average.  This is where we see the majority of people with Asperger.  So, the data don t fit the model precisely, but they re certainly in the right direction, suggesting that we ve got some justification in saying that this profile, empathy better than systemizing, is more typical of women than men.  And this profile, systemizing better than empathy, is more typical of men than women.  And, just to underline the point so that we really don t sort of have any misinterpretation of the data, you can see that really all we re talking about is the sexes diverging, it s not like the majority of women show this profile, it s just that more women show that profile than men do, and equally fair for men, you know, we re just looking at patterns.  Now we can sort of try and explore biology because there isn t really any question that post-natal influences culture experience are likely to be shaping sex differences.  We don t really need to get into a big argument about whether culture plays a role, the real question is whether biology also plays a role.  One way to study this is to look at newborn babies to see whether there are any sex differences at birth before culture has had much opportunity to shape sex differences in the mind.  So here we ve got a study of babies aged 24 hours old, so this was the youngest we could find them, and, we asked the obstetricians in the local maternity hospital if we could study them minutes after birth, and they asked us if we would kindly wait for mother and baby to settle down for one day.  So, it s not a pure test of the role of biology.  They re 24 hours old, but we waited for the baby to be awake, the baby is flat on their back, and they are presented with two objects, one at a time, a human face, or a mechanical mobile object, and they are filmed for how long they look at each object.  It s a test of attention.  And, I should say that the researchers who did this study asked the parents not to tell them the sex of their baby so that they could remain blind, and, by the time the videotapes were coded later for eye movements, it s actually very difficult to tell whether this was a male or female baby.  But, let s jump straight to the results.  That, more boys than girls, this is percentage, who looked longer at one object or the other, more boys than girls looked longer at the mobile, the mechanical mobile, and more girls than boys looked longer at the human face.  There s a percentage of each sex who looked equally long at both.  So, again, the results don t apply to every individual in the group in each sex.  As far as I know, this is the only study of its kind.  I would love to see some attempts at replication, but, this was published now, six or seven years ago, and they are very time-consuming studies, to study newborn babies, but it would be good to see whether these results replicate.  But, assuming that the results have some validity, if we re finding a sex difference in attention to a social object versus a mechanical object on the first day of life, one implication is that whatever the role of culture later on, prenatal biology must also be making a contribution.  So now we can move to hormones, and, in particular, fetal testosterone, because I m going to finish by telling you about this, I think, quite unique study, where we ve been taking advantage of women who, during pregnancy, are having amniocentesis.  Here is the fetus, surrounded by the amniotic sac, and bathed in amniotic fluid.  As you know, during amniocentesis, which is only about 6 percent of pregnancies, a long needle is introduced into the amniotic fluid, and that fluid is sampled, usually because the fetus is thought to be at risk of some developmental condition like Downs Syndrome, so the fluid is analyzed for chromosomal abnormalities.  Here we are asking the mothers for their consent to also analyze the amniotic fluid for fetal hormones, in particular, testosterone.  And that s because, from animal research, it s very important background, animal research tells us that if you inject testosterone into the amniotic fluid, let s say this is now not a human, but it s a rat, and it s a female rat, and you introduce, you raise the levels of fetal testosterone in the amniotic fluid, some of that will be taken up by the fetus, into the brain.  And when you look at the female rat s brain, after they re born, post-mortem, her brain has been masculinized, it much more resembles, structurally, neuroanatomically, the brain of a male rate.  And her behavior, if you looked at her behavior, these treated female rats, if you let them run through mazes, usually that if you let male and female rats run through a maze, the males will run faster through the maze, find their way through more quickly, but the female rats who have been treated with testosterone become as fast as the typical male rats.  So the fetal hormone is having an effect on both brain anatomy and behavior.   As I said earlier, the fetus is producing its own testosterone, in males it s produced by the testes, in females by the adrenal glands, gets taken up by the blood, goes into the brain, so it crosses the blood brain barrier, and the way that the hormone does its work is by binding to androgen receptors, all of those regions, which I showed you earlier, which are dimorphic between the sexes, are rich in androgen receptors, and, so, there s a very good chance that some of the differences we see in brain structure between the sexes arise because of the effect of fetal testosterone and its activity through being bound to androgen receptors.  Once it s bound, it can modulate neurotransmitters like serotonin and GABA, and there are some theories suggesting that testosterone also affects connectivity in the brain by preventing apoptosis, or selective cell death. So, lots of reasons why we should expect, let me go back, this hormone, fetal testosterone, may be a good candidate mechanism for understanding sex differences in the brain and behavior.  So what we re doing is, well, this slide just tells you that we ve succeeded in measuring fetal testosterone in the amniotic fluid, and, so, here we ve got boys and girls as fetuses, and this shows you that hormone levels, as you might expect, boys are producing more than girls, but what I wanted to draw your attention to was the huge individual differences.  We ve got some boys whose fetal testosterone is so low that it s in the female range, and we ve got some girls whose testosterone levels are high, and are in the male range.  These just show the outliners.  The question is, if you ignore somebody s sex, and just look at their hormone levels, does that predict anything about the kind of behavior they show post-natally?  The bottom of this graph just tells you about something that s considered to be a consensus, knowledge in development neuroscience, and, particularly, the sex steroids, which, this is a timeline, his birth, if you look at production of testosterone, there is a surge in the production of testosterone between about 10 and 18 weeks of pregnancy.  After birth there is another little surge, about three months after birth.  If we continued up here, we d see a third surge, which is puberty.  We know what this surge in testosterone is doing at puberty, because that s to do with the timing of puberty and the secondary sexual characteristics.  But, one assumption is that this early surge in production of testosterone, prenatally, may have a role in brain organization and shaping differences between the sexes in terms of brain anatomy.  This is the study.  We measured the hormone levels during pregnancy because this is the time when women are having an amniocentesis anyway.  By good fortune, it coincides with that surge in prenatal testosterone.  But we re just taking advantage of the fact that this is a time when we can study hormone levels in development.  Then, after the baby is born, we have called in the infants, their 1st and 2nd birthday, to look at behavior like eye contact and rate of language development, vocabulary size.  When they go to school at 4 years old, we ve looked at quality of social relationships.  Now these children are 8 years old, so we ve been following them.  We can look at more entrusting things like their empathy, their systemizing, and we ve given them tests like the embedded figures test.  So what d we find, just cutting straight to the results, we re looking at the children, the first point when we ve seen them, which is at 12 months old.  We filmed them whilst they re playing with toys.  We later look at the videotapes for how often does the child lookup from what they re doing, at their mother s face.  The mother is seated at the side of the room.  We instruct the mother not to initiate any social contact with her child, but just to respond if the child initiates.  We re looking to see how often does the child lookup from what they re doing at their mother s face.  And what you can see is that girls look up much more often than boys.  There s a big difference in the rate of eye contact in a 10-minute play session.  Now, we can see if there is any relationship with fetal testosterone.  Now we can look back at fetal testosterone.  There s the scale.  This is the amount of eye contact.  Girls are in red, boys are in blue, and no surprise that there are more red dots down this end of the scale because girls are producing less testosterone than boys.  But, is there any relationship between these two axes?  The answer is yes.  There is an inverse, that s to say a negative correlation, which means the higher your fetal testosterone the less eye contact you make at 12 months old.  We then saw these children at their 2nd birthday, and now we re asking the parent to fill in a checklist about vocabulary comprehension and production.  So this is the McArthur, which many of you all know, or the British equivalent of it, and you basically ask the parent which of these words, in a long list, does your child know and produce?  What you can see is that girls have a larger vocabulary than boys at 2 years old.  It s quite a significant difference.  We ve actually known this for a long time, that girls talk earlier than boys, that boys are more likely to end up in clinics for language delay and for language related disorders like stuttering.  What we haven t really known is, why is there a sex difference in rate of language development.  So now we can look back at their fetal testosterone levels to see if there is any relationship between vocabulary size and testosterone levels.  And, again, the slope of the line is downwards, meaning that the higher your fetal testosterone the smaller your vocabulary at 2 years old.  So here we re finding this hormone shows some association, not just with a measure of social interest, eye contact, but also language development.  I m jumping ahead now to the fact that they re 8 years old, because now we can give them tests of empathy.  Here we ve asked the parents to fill in the empathy quotient about their child, now that the child is 8.  You can see that we ve got girls in red here and boys in blue and then people with autism in green, and here s the empathy score, basically saying that girls at 8 years old are scored by their parents as having better empathy than boys.  So that s just the instrument, the EQ.  But now, let s relate it back to fetal testosterone, once again, the slope of the line is downwards, meaning that the higher your fetal testosterone score, the lower your empathy score at 8 years old.  That result is mirrored, very closely, by not just a questionnaire, but, this performance measure.  So you remember that we had this test looking at eyes and judging what someone else is feeling or thinking.  But now we ve adapted the language on the test to be appropriate for an 8 year old.  So you have to choose which of these four words best describes what the person in the photo is thinking or feeling.  Hopefully you agree that he looks interested.  We ve given this to the 8 year olds and found not just a sex difference with girls being better than boys at this test, but also that there is this negative correlation, the higher your fetal testosterone, the lower your score on this test of empathy.  What about systemizing?  This is the obvious other question.  So now we ve got the systemizing quotient, which is adapted for children, parents are filling it in to report how interested their child is in a range of systems, like machines and mathematical systems.  We ve got boys in blue, girls in red, small, but significant difference, and in green people with autism or Asperger, suggesting there are differences in strength of interest in systems, boys having a stronger interest than girls.  When we look back at fetal testosterone levels, for the first time you can see it is flipped over, the line is going upwards, meaning that the higher your fetal testosterone, the stronger your interest in systems of a variety of kinds.  That is mirrored on that test of the embedded figures that we saw early.  So there s the test, where you have to find the target shape as quickly as you can in the overall design.  If I trace it out, you can see that the triangle is hidden there.  It s a timed test.  Boys are faster than girls on this test, compared to girls.  Children with autism are super fast on this test, excellent attention to detail.  And when we look back in this study of fetal hormones, again, the line is going upwards, meaning that the higher your fetal testosterone levels the faster you are at finding the target shape hidden in the overall design.  There s one last measure, which is a measure of autistic traits, which we ve used in this study of fetal testosterone.  I should say that if we go back to these studies, none of these kids have autism, they are only in the study because their mother had an amniocentesis, but we are able to look at individual differences in psychological outcomes in relation to fetal testosterone levels.  But, if you take the kids in these samples, and now give them a measure of autistic traits, so this is not the same as looking for a clinical diagnosis of autism, but just how many traits does the child show, you can use measures like the autism spectrum quotient where you just go through these statements and each statement counts as one autistic trait.  What you find is that there are individual differences on the EQ, autistic traits.  Here s the general population, showing that even without a diagnosis there is huge variability in the number of autistic traits people have.  In green we ve got people with a firm diagnosis of autism or Asperger who are scoring much higher on this spectrum of autistic traits, but it is confirming this new view of autism, which is that it s on a continuum, that autistic traits are distributed right through the population.  So now we can look at any relationship between fetal testosterone and the number of autistic traits.  What we find is a positive correlation, the higher your fetal testosterone the larger the number of autistic traits a child is showing.  So, I am going to finish with a number of conclusions, and then hope we can open it up for a provocative discussion.  The first point is that different studies do suggest that there are sex differences in the mind.  They are often small and subtle, so you don t find them in every study, but where you do find them it s worth paying attention to them, sometimes you need large sample sizes to identify them, but rather than sweeping them under the carpet we have to examine why they are arising.  The second point is that culture is important, but it doesn t look like culture can be a complete explanation for the sex differences that are found because not just from the evidence from the newborn baby study, but also from this study of fetal prenatal testosterone.  So that s the third point in the discussion, that, sex differences seem to be associated with fetal testosterone, and I didn t say, but, in the fetal testosterone studies, when you look at the patterns of correlations just within one sex, so, just within the boys, you still find the same patterns.  We re not just seeing more evidence of sex differences at the hormone level, we really are seeing hormonal effects.  The last couple of points there is that autism spectrum conditions may fit an extreme of the typical male profile, so, this gives us a new way of thinking about medical conditions like autism, rather than thinking of them as categorically unrelated to the rest of us in the population, we may start to think about autism as simply an extreme variation of individual differences that exist in the population.  And, finally, as a prediction, fetal testosterone levels may be important in the cause of autism.  We haven t studied that yet.  To do that you would need a much larger sample size than we have had available in these amniocentesis studies, but we are now collaborating with the Danish Biobank, who have 90,000 samples of amniotic fluid.  They ve stored in the deep-freeze every amniotic sample from a woman who has had amniocentesis since 1980.  So, there s the opportunity to look at actual cases of diagnosed autism and their fetal testosterone levels that were collected at the time.  But, because autism is a genetic condition, it runs in families, we don t think this is purely a hormonal, likely to be a purely hormonal story.  We know that there are a whole bunch of genes that influence your hormonal levels.  We re studying about 25 genes that are involved in regulating testosterone levels.  These may turn out to be important genes for understanding the cause of autism.  I d just like to finish by inviting you to visit our website if you want more details of any of these studies.  Thank you very much.  [Applause].  Christina Hoff Sommers:  Do you just want to, I can call on people.  [Inaudible].  Simon Baron-Cohen:  Sure.  Female Audience Member:  I just want to know, how do you know what the baseline truth is on the eye, emotional exam?  Simon Baron-Cohen:  Well, it s a very valid question, and, that eye test, the correct answer doesn t lie in the stimuli, you know, the correct answer is determined by consensus.  So you have a large panel of judges and you just simply look to see which of the four choice words on each photograph is picked by the majority.  So it s a consensus-based approach.  But, you don t know what the mental state was of the person in the photo when the photo was taken.  There is no way of really knowing that.  Female Audience Member:  I actually have several questions, but, I ll just do one.  So, when you show a face to a baby versus a mobile to a baby, does a baby know the difference between what s a mobile and what s a face?  Simon Baron-Cohen:  No, we re not expecting them to tell us the difference, we re simply looking at how long they look at each.  So, this is, it s a standard way of studying infants, is to look at looking time, and that s because little babies have very little control over their bodies, they can t talk, but one thing that they do have control over is their eyes, and their eye movements.  Female Audience Member:  Right, but how do they know what, I mean, I guess, how can you interpret do they know what they re looking at?  I mean, could it be round versus square, could it be beige versus?  Simon Baron-Cohen:  Absolutely, and, I think what you re saying is, are there certain properties of the stimulus which might be attracting more boys to look at the mobile and more girls to look at the face, was it color, was it size, was it luminosity, was it the way they moved, you know, lots of different stimulus properties.  I think you d need a whole series of experiments to tease those things apart.  We tried to control for some of those factors.  So, we had a human face as the face, but the mechanical mobile was actually made to be the same size as a human face, and the same color, but was not face-like.  So we tried to match the two stimuli for size, shape and color, but, obvious difference is that a human face moves by itself and it has emotional information, it moves with animacy, and a mechanical object moves in an inanimate way, so it may be patterns of motion that were capturing the boys and girls attention differently.  Female Audience Member:  Were the faces moving?  Simon Baron-Cohen:  Yes, I mean, in that experiment, I don t know who was asking the question, but in that experiment both types of objects move, they obviously move differently because living things move differently to mechanical things.  Female Audience Member:  Did you do any further research on the females who had high fetal testosterone levels, and the males who had very low testosterone levels below the medium for females?  Simon Baron-Cohen:  At the moment we ve just looked at these correlations, so we re trying to set aside sex, or, what in the states you call gender, but, in Europe, we call it sex.  So we re setting aside what sex the person is.  And, we use the word sex, by the way, just because that describes your chromosomal makeup, and, that s how we designed individuals into being male or female, on their assumed chromosomal makeup.  We re just looking at, if you ignore their sex, do you find patterns, correlations with hormone levels?  I think what you re asking is, if we now start looking at individuals taking, for example, males who are low in testosterone, or females who are high, so they are atypical for their sex, are there any interesting patterns?  And, we haven t done that yet.  As they re getting older it s going to be very interesting to see whether you can predict patterns of interests academically or occupationally from their prenatal levels.  It s one of the, you know, I hope that I can be involved in research long enough to see how the results turn out as these kids go into adulthood.  Male Audience Member:  Could you just run through what the magnitude of some of those correlations were between fetal testosterone and some of the outcome measures, because they were obviously very different, some of them?  Simon Baron-Cohen:  They are, and, across a range of measures, the fetal testosterone variance is really only accounting for about 20 percent of the variance in outcome.  So, we re not saying that this explains everything, but the fact that there s any pattern at all means that this may be an interesting biological mechanism to explore further.  This is important to underline because some of us who study biological variables are sometimes misunderstood as saying that biology is the only important factor.  I m just saying, you know, that depends of repetition, for the nth time, that cross-culture is important, but some of these biological factors should also be considered.  Female Audience Member:  Correlation, of course, isn t cause, and, if the recent implication that autism may be linked to women who were under fertility drugs to get pregnant, how would that integrate with the findings you ve got?  Simon Baron-Cohen:  Well, first of all, I don t know that finding.  Female Audience Member:  It s been reported in the U.S. press, probably within the last half-year or so, again, it s implication that that might be a link.  Simon Baron-Cohen:  Right, yes, I don t know if it s been in the scientific press or just the news.  Autism has, for whatever reason, caught the public imagination, and every week there s a new story about how autism is associated with something like how much TV you watch.  So, we tend to wait until things, until the dust settles, and things go through peer review.  But, if it was the case that women who were taking, is that what you were saying, fertility treatment drugs affecting hormones, that that was influencing rate of autism?  Well, that would be important, I mean, personally, I d be surprised at that, a result like that, and that s simply because the placenta is a very effective barrier for protecting the fetus from maternal hormones.  So, the mother has her own testosterone, but it doesn t tend to cross the placenta because of an enzyme in the placenta, aromatose, which converts it to estrogen.  So, because testosterone at high levels is neurotoxic, the fetus needs to be protected from increasing levels.  The same is true the other way, actually, the fetal testosterone doesn t cross to affect the mother.  Female Audience Member:  [Inaudible] egg/sperm not the fetus.  Simon Baron-Cohen:  Well, the implication is that it s the hormone being produced by the fetus themselves rather than what the mother may have drunk in the water or taken pills during pregnancy or whatever, that it may be genetic factors affecting hormone levels that are more important.  But, let s wait until the studies have been done.  Female Audience Member:  Can you comment a little bit, please, about the hereditary aspects of autism, and what kinds of in-depth studies have been done on this?  Simon Baron-Cohen:  I think the consensus is these days that autism has a genetic basis.  I know in the states there s still a very powerful lobby suggesting that vaccines have damaged the child s brain, and mercury levels in the vaccines may have produced heavy-metal toxicity, some environmental component.  But, the heritability of autism is not really questioned by the majority of scientists.  That s because if you look at twin studies, if you look at identical twins where one has autism, the likelihood of the co-twin also having autism is much higher than in non-identical twins.  In identical twins, it s about 60 percent concordance rate, in non-identical twins it s only about 5 percent.  There s a massive difference.  And, these days, not a week goes past without a new genetic association at the molecular level being reported in autism studies.  So, new genetic differences are being found between cases and controls, autism and people without.  So I think it s only a matter of time before we see so-called susceptibility genes for autism being identified, but the idea that autism is genetic is no longer really in question.  Elizabeth Spelke:  I want to, just a quick comment, and then a question.  The comment is to second your suggestion that the, your study of newborn infants be replicated, for three reasons.  One is, for, obviously, understandable reasons, this is a study that s gotten enormous attention.  I think I ve seen it referred to in Op-Ed pieces, in numerous newspapers, and, so forth.  So it s clearly an issue that s perceived to be of great importance to people.  The second reason is that the finding is at odds with findings from a very large literature looking at older infants, 3, 4 months, and on, spontaneous attention to people and objects, which hasn t shown that sex difference.  The third reason is that in that particular study you don t have control over the stimuli that are being presented.  I believe the face is a live person, the first author on the paper, and she is manipulating the object, which means there really isn t, you can t be sure that the different babies are seeing exactly the same thing.  I think it would be quite easy, actually, to redo this with videotapes and have that control, and hope very much that you will.   Simon Baron-Cohen:  Well, I mean, it s one thing for us to try and replicate our own result.  It s always much more interesting if an independent research group attempts the replication.  So I m hoping that will happen.  Just picking up on some of the points.  So, it s difficult, when you re testing newborn babies, to have a lot of opportunity for stimulus control, and experimental control.  I mean, you know, because you ve done this kind of stuff.  Elizabeth Spelke:  Well, you can use videotape, so you re at least sure that all the babies are seeing the same socially responsive [cross-talking].  Simon Baron-Cohen:  We decided to go for, not for videos, because, in a way, touching on issues that David Geary raised, that, if this has any evolutionary significance, ecological significance, you don t really know what it is that gets lost in a videotape from a live stimulus.  What we re talking about, and the question earlier, is, what is it that the baby sees when it sees a face?  We don t know which aspects of the face are the thing that might trigger the interest of someone who likes faces.  Is it color?  Is it movement?  Is it something else?  So we went, in order to, as a first test of the sex difference, we went for live stimuli.  The second point you were making is about the lack of sex differences in some earlier studies.  Well, we went for quite a large sample because, I think sex difference is maybe quite subtle, so you may need adequate sample size to have the power to reveal the difference if it s there.  So we were going for, in that study it was 100 newborn babies.  Most infancy studies and developmental, you know, cognitive developmental studies, are often much smaller than that, so don t necessarily allow for a test of males and females with enough individuals in each group.  Elizabeth Spelke:  Okay, let me move to my question, which is to ask you to relate some of the very interesting data you showed us to the issue of sex differences in aptitude for math and science.  I think it s extremely interesting that both your eyes test and the embedded figures test are showing these effects of fetal testosterone levels.  Is there any reason to think that good ability to read eyes is a handicap in a scientist, or, that good ability to find a triangle in a baby carriage is an asset?   Simon Baron-Cohen:  In a way, some of the findings I ve presented are, they re not directly related to adult outcomes in terms of the topic of today s conference, they re just really food for thought.  We have looked at scientists, so students who are studying science, compared to students in the humanities, in terms of the eyes test, and in terms of the embedded figures test, may be no surprise, but we find that students who are majoring in science do better on the embedded figures test and do worse on the eyes test.  We don t really understand why someone who is good at science should be disproportionately less good at this measure of empathy.  Is it, you know, it s very hard to disentangle cause and effect, and all we ve actually got is correlations, as was a point made earlier.  But, it s an interesting question as to whether there may be some kind of tradeoff, that the better you get in one domain, the worse you get in another, and the fact that we re finding a single molecule, in this case, fetal testosterone, showing the association with both tests of empathy and tests of systemizing, may mean that they share some common mechanism.  But, there s a lot of questions to disentangle in this area.  Rosalind Chait Barnett:  I have at least one question, and I hope we can get the second one in as well.  I m going back to your data on vocabulary, the mother s reports.  It seems to me, and you have to help me understand this, that there s a real confound there.  First of all, you don t measure anything about the mother/daughter or mother/son relationship, maybe that mothers are reading more to their daughters, talking more to their daughters, so that your, the children who are high in fetal testosterone tend to be boys, and they tend to have lower vocabulary, but it may well be an artifact of the fact that mothers are not spending as much time in direct interaction.  So I think that that s a huge piece that needs to put in there before you can make any sense.  Let me just relate it a little bit to the second one, which is the toy choice.  I don t know if you control for prior experience with those toys, in terms of how much time kids spend, it s very likely that girls are spending more time, anyway, with dolls, because that s what their parents are buying for them, and boys are spending more time with mechanical things.  So unless you control for prior experience, or novel toys, then it s hard to make a deduction.  Simon Baron-Cohen:  Yes, I think both points are well made, and, let s start with the second one first, because then I can remember it.  I think that the toy choice study doesn t really tell us about culture versus biology.  I mean, it s telling us that by the time they get studied, between 1 and 5 years old, you see sex differences in pattern of toy choice.  But, you re right, this could be the result of prior experience.  They are already 1 year old, which is quite old, really, they ve had a lot of experience.  So, all I was really illustrating there was that when you leave children to show spontaneous behavior you do see some sex differences, and we need to understand where those come from.  But, your first question, which is about, did we look at, let s say, on the language test, vocabulary side, did we look at parental variables, what are parents doing with their children, not just hormonal variables?  We had the opportunity to look at some social variables, so we looked at birth order, number of siblings, whether one or both parents were working outside the home, and we controlled for each of those factors.  We didn t look, I think you re absolutely right, we didn t look, in fine detail, at how much time parents were spending reading to their child.  I mean, you re right, there may well be, and I ve tried to underline this point, there may well be post-natal experiential factors that contribute to your ways of language development.  We know that first-born children talk earlier than later born children.  That s unlikely to be a biological mechanism, it s probably to do with how much attention the child is getting, and their lack of, as it were, sibling attention from the parents.  So, I think post-natal experience is, I m not questioning that that s important, but, just taking your idea about, do moms read to their daughters more than they read to their sons, or talk to them more?  There are lots of studies, as you know, showing that parents inadvertently treat their sons and daughters differently.  They use more emotional language towards their daughters, for example.  We looked just within sex, so this is quite an opportunity to look just at hormone effects.  If you just look at the boys, it doesn t rule out the possibility that the kids who are doing best on language development also had more reading experience, but we were finding some association with prenatal testosterone.  Like you, I would like to focus everything on post-natal experience, but, to do that would be to ignore certain findings that are coming up in the studies.  I m trying not to ignore some of the biological evidence.  I m trying to keep in mind both biology and culture.  Christina Hoff Sommers:  We don t have too much time, but we have a question here, here, here, here, and then we ll break up in time for the next session [inaudible].  Female Audience Member:  So, realizing that there are only so many variables that you can include in your very interesting study, I wonder if you are assuming that all of these subjects are heterosexual.  And the reason I ask that is because if you were able to superimpose sexual orientation later on, it could be rather interesting.  What I m thinking is, the person who does my hair and the person who I hired as an interior designer happen to be men.  Now, having said that, when I started a gay/lesbian/ bisexual/transgender group in engineering at a large research one institution, people started coming out of the woodwork, and they were very highly intelligent men whom, I m sorry, and I m generalizing here, but I would describe as being extremely empathetic, but also tested very well.  So, I was just wondering if, these may be data that are screaming at you that are out at the, sort of the outliers.  Simon Baron-Cohen:  Sure.  So, sexual orientation is very, is potentially a very relevant variable to bring in, and, obviously in our longitudinal study the kids are only 8 years old, so, we might wait a few years, but that s the beauty of the design, is that we can follow them as they go into adolescence, and some of these differences in sexual orientation will become apparent in the next five years, and we ll be able to see whether sexual orientation has anything to do with some of the other measures we ve been looking at, like empathy or systemizing, and, indeed, the same mechanism of fetal testosterone, or is under some completely different, under the control of some completely difference mechanism.  So, I sort of feel I ve got myself into quite a lot of trouble already just looking at male/female differences, but, you re right, the study does allow you to look at other variables that could be important.  Female Audience Member:  That was precisely the point that I was going to make.  You should be looking at different hormones.  It has been well-shown in rat models and also with human cases that high, or elevated levels of adrenaline during the third trimester of pregnancy will result in gay male offspring, whether they are rats or humans, not everyone, but a higher percentage, and, also, that this will cause, if you give estrogen to these feminized or gay men, that the estrogen will be metabolized in the female pattern in the brain rather than the male pattern.  You also have X, Y non-responders who do not respond to testosterone, who develop, by default, into the female pattern.  You should look at what theirs are.  And you should look at women who have adrenal hormones due to adrenal tumors being elevated during pregnancy and what happens to their female offspring.  There are so many possibilities that you should be looking at, at other hormones.  Simon Baron-Cohen:  Yes, I completely agree, and we are doing that, I mean, I gave you a little snapshot here, but we are looking at certainly the other sex steroids, like estrogen, and people think that actually a lot of their sex differences in the brain may be more easily related to estrogen than to testosterone levels.  But that s another whole study.  And there are other hormones outside the sex steroids, exocytosis is a very important, new candidate, for explaining differences in sociability, and the medical conditions that disrupt your sex steroids are also very relevant.  So we have looked at congenital adrenal hyperplasia.  So, in particular, girls, who, for genetic reasons, are over-producing fetal testosterone, and, we re looking at their behavior.  And the converse, people with androgen insensitivity syndrome, so these are chromosomally male, but they have no androgen receptors, so they can t use their testosterone.  They look female, in terms of their body, they re being raised as girls, they believe that they are female, but they have a Y chromosome.  So, they re a very interesting group to test, specific hypotheses.  Female Audience Member:  And I would differ, perhaps, with you, with regard to the testosterone being only produced by the fetus, and having an effect on the fetus, because, within the past month there has been a study, and I m sorry I didn t think to remember now, where they tested, during pregnancy, every woman that goes in for regular pregnancy medical care gets a urinalysis.  They check testosterone levels in the urine of these women and in the offspring of women with high testosterone during pregnancy they did find some difference, and I cannot think, at the moment, what it is.  So I would look not just at fetal, but at the maternal.  Simon Baron-Cohen:  Sure, but remember, that even if you find the mother s hormones, let s say, testosterone levels are high, and so are the fetuses, that could be for genetic reasons, because mother and child share half of their genes.  So we don t know that it s because the hormone is crossing the placenta, it could be because they are genetically related.  So, it s just to explain the complexity.  Christina Hoff Sommers:  We have time for two quick questions from you, and then two quick answers, and we ll move on.  Male Audience Member:  My youngest son was a subject in an experiment within the first 24 hours after he was born.  He grew up to be extremely athletic.  I ve always wondered if they returned the right baby.  [Laughter].   Simon Baron-Cohen:  But he calls you Dad.  Male Audience Member:  Good enough.  My question is, I understand that there are times during human gestation where the brain produces 500,000 new neurons every minute.  This is extraordinary, and it would be very odd if this rate of production wasn t terribly sensitive to things like testosterone.  I m wondering if the testosterone surge corresponds with these bursts of neuron development?  Simon Baron-Cohen:  That s a fascinating question.  It s probably one that has already been addressed in animal research.  We re trying to do something a little bit new in human research by using amniocentesis, but, this field, as you know, is actually quite well established in animal research.  So, it would be worth looking to see if those studies have already been carried out.  Christina Hoff Sommers:  Final question.  Male Audience Member:  All right.  I had a question regarding the parental differences, or, a comment, regarding parental differences in treatment of boys and girls regarding vocabulary.  It is the case that parents do engage more social contact with girls than with boys, but, at least part of that difference is due to the responsiveness of boys and girls to that parental contact.  Boys, little guys are much less responsive to their moms than the little girls are, and perhaps that never changes as they grow up.  But, in any case, I think those differences in responsiveness are entirely consistent with this view, that there may be an early different in attention to social stimuli and reaction to that.  I think, also, that the testosterone effects may be underestimated here because what you re getting is a single point within this curve, and what you really want is a total area under the curve to get not only the amount at a single point but also the total cumulative amount during that time.  So, these are probably conservative data, would be my guess.  Simon Baron-Cohen:  Thank you.  Christina Hoff Sommers:  Thank you, and let s give a warm.  [Applause].  Just a 5 minute break and come right back for our next panel, maybe 10.  PANEL II:  Stereotype Threat:  The State of the Research Christina Hoff Sommers:  We re going to get started, take your seats.  Welcome to this afternoon s session.  We re going to be talking about stereotype threat, and Josh Aronson is here.  He s an Associate Professor of Applied Psychology at New York University, where his research on social and psychological, he researches social and psychological influences on motivation and confidence.  In 1995 he and his colleague, Claude Steele, published groundbreaking laboratory studies on stereotype threat, and, according to this research, when someone is exposed to negative stereotypes about their race or gender, this can have a debilitating effect on performance and achievement.  Together with Claude Steele, he now co-directs a national taskforce on the achievement gap.  We re delighted and honored to have him here.  He s going to explain how a stereotype threat might affect the performance of women in math and science.  He will be followed by Amy Wax, who is a Professor, a Robert Mundheim Professor of Law at the University of Pennsylvania Law School.  You might want to take some time to ponder her resume, it s kind of amazing.  I think if it were written by a Hollywood scriptwriter you d tell them to tone it down a little bit.  She got her B.S., Suma Cum Laude, in Molecular Biophysics and Biochemistry at Yale, and then her M.D. with Distinction in Neuroscience at Harvard Medical School.  After that she had a residency in Neurology, completed that, went to Columbia Law School, where she was Editor of the Law Review, and since then has pleaded 15 cases before the Supreme Court, and I think we re going to send her to solve the problems of the Middle East, or, I don t know what you ll do next, but, there seems to be.  I m looking forward to hearing her reaction to research on the stereotype threat.  Professor Aronson will begin.   Joshua Aronson:  Thank you so much for having me.  I m honored to be part of this conversation.  I really don t have a strong opinion about sex differences.  What I think is so fascinating about this area, I m a social psychologist, we study contextual influences on behavior, and one of the most amazing things to me is that, something that William James said many, many years ago, is that, we re as many different people as we have social contact, so we can be very different.  Lots of you may notice this in your everyday life, that you feel smarter or funnier or more charming with some people than you do with others.  That s essentially a social psychological finding, and it s one that is relevant for what we re talking about today, that human intelligence, rather than being this thing in the head that you carry around with you from situation to situation, actually is impacted by the social environments that we encounter.  So what my take is on the sex differences, I think that everything I ve seen today convinces me that there s a mixed bag of evidence for it.  I think you would agree.  It s a bit of a Rorschach test that what you came in here thinking is probably what you still think because there s evidence on each side of it.  What I think, with great conviction, despite not having a strong opinion on whether biology is the real reason for the paucity of women in sciences like engineering, is that nature may give us these little differences, people have talked about these subtle differences, and we are biological animals, but we are also cultural animals.  We evolved for culture, some people say.  And, what a culture does is takes those differences, one of the things a culture does is sort of blows them up, you know, so, like you see the average body shape of a woman, that is made by nature, but then you see a Barbie doll, which is made by culture, and we hold up the Barbie doll, and I m fascinated by the fact that, for example, girls look to Barbie dolls and evaluate their beautiful bodies that nature created and feel bad about their bodies.  That s the kind of cultural effect that fascinates me.  And that s sort of what stereotype threat is about, particularly given the fact that Barbie, one of the things she says when you pull the string, is, math class is hard.  [Laughter].  I define stereotype threat as the potentially disruptive psychological discomfort arising from the awareness of a negative group or personal reputation in a situation where the reputation is irrelevant, and, thus, confirmable.  We all experience this in some way.  If it s strong enough, stereotype threat can interfere with intellectual performance, and what I mean by strong enough, that depends on the nature of the situation, and the person, there are big individual differences in this, no two people react exactly the same.  Some examples of stereotype threat, you know, an African American taking an IQ test, perhaps they ve read the bell curve, they ve heard that Black people don t do well on these, they feel nervous that the stereotype might be true, they might be judged to be less intelligent.  Or, if they re experiencing difficulty, the stereotype activated in their brains may create the feeling that I m not smart enough to do this.  A girl called upon in math class may wonder whether her answer is going to mark her as one of those girls who can t do math.  Finally, my favorite is George W. Bush in public speaking.  I think he s absolutely fascinating because in his 1994 debate with Ann Richards, he was described this way by James Fallows == Bush was eloquent, he spoke quickly and easily, he rattled off complicated sentences and brought them to the right grammatical conclusions.  He mishandled a word or two, million, when he clearly meant billion, stole when he meant sold, but, fewer than most people wouldn t in hours debate.  More striking, he did not pause before forcing out big words, as he so often does now, or, invent mangled new ones.  So, here s a guy who s very intelligent, and, by the way, I know his SAT score, much lower than the people I get to estimate his SAT score now, imagine, so, raise your hand if you think he got 900.  Raise your hand if you think he got 1,000.  Raise your hand if you think he got 1,200.  Charles Murray knows.  He got a 1330 by today s standards.  I ve surveyed thousands of people on this.  They give him a 900.  So he appears less intelligent than he actually is, and says, of course, things like, I know that people are working hard to put food on their family, childrens, plural, do learn when standards are high and results are measured.  So here s a very smart guy who appears to the world, and sometimes acts unintelligent.  What causes that?  Well, one person has argued that he has pre-senile dementia, a doctor.  I disagree, and so do his neurological exams.  What happened, is, if you do a search for trying to find out what George Bush s reputation was prior to the 2000 election, it was as a pretty smart guy, a little nepotism, maybe, but no one was talking about what a dummy he was.  During the 2000 election it started echoing through the media echo chamber, after he sort of failed an on-air quiz about international politics, this guy is kind of a dummy.  Well, imagine if everybody starts thinking that you are stupid and part of your job is to get up in front of people and convince them that you re not.  It s very unnerving.  It freights the situation with a set of unnerving expectations, which can make a smart person less smart.  That s the notion of stereotype threat.  Now, of course, that s anecdotal.  I can t prove it.  What we do to look at this phenomenon is to conduct experiments.  Experiments, we started looking at the Black/White achievement gap, test score gap.  The idea was that if you could reduce this sense of apprehension about confirming a stereotype, you could improve performance.  We had participants at an elite college, African American and White, taking a very difficult verbal test.  For half of them we gave them the test the way you normally get these tests.  This is a test of your verbal intelligence.  It determines important things about your future.  The other condition, we created an exotic condition in which you are getting the same exact test, but now you re being told that the researchers have no interest whatsoever in measuring your ability with the test, they want to talk to you afterwards about the psychology of problem-solving.  Still, you re encouraged to do as well as you can.  Here s what we find when we look at performance.  Higher scores equal higher scores, believe it or not.  When we tell them they re measuring your ability, we get a big Black/White gap.  When we tell them we re not measuring your ability, the Black scores rise a lot.  Now, we re talking about women here.  Spencer, Steele and Quinn did a similar study, in which they simply gave college women and men, matched in terms of their mathematics grades in college, a difficult exam.  This is what they found when they simply framed it as a test of your math abilities.  In another condition, they told the women, and, by the way, this test has never produced gender differences.  What you find is that performance on this test goes up a lot, and you may notice that men s performance drops in this condition.  That s a consistent theme in this research, that the advantaged group, the group that s supposed to be better, when they find out they don t have a particular advantage, their performance goes down a little bit.  There have been over 200 studies published on this phenomenon since we published our first paper in 1995, and it generalizes to all kinds of different groups -- Latinos taking verbal tests, old people who worry about the stereotype that old people are senile do worse on a short-term memory test, low SES students taking verbal tests, Black students taking tests in miniature golf, not a big problem sweeping the nation, I realize, but, if you frame even a stupid game like miniature golf as a test of kinesthetic intelligence, what you find is that the White golf players do much better, and the Black players do much worse.  However, if you frame the same exact task as a test of natural athletic ability, now the stereotype is going in favor of the Black students, they do better and the White students do worse.  Women taking tests of political knowledge, in that study if a male interviewer calls a woman to ask what she knows about politics, the woman scores worse than the same set of questions asked by a female.  White males taking tests of social sensitivity, something that apparently they fail at, could call it the Bill O Reilly effect.  [Laughter].  Sorry about that.  [Laughter].  And, just to show you that this can happen to anyone, White males taking math tests.  Now, this study I did a long time ago at Stanford.  The idea was to see whether this is a general process that anyone could experience, or whether it required sort of a history of stigmatization of internalized feelings of I m no good.   We believe that it didn t, but, for a stringent test of this, we went out to find the whitest and brightest males we could find.  It turns out they were Stanford male engineering students, mostly they had SAT scores above 700.  In one condition, we told them they were taking a very difficult math test, and, they scored at that level.  The other group was told the same exact thing, this is a test of your math abilities, however, the reason we are looking into this is we want to see why Asians are so good at math.  In that condition, these White males plummet almost a whole standard deviation in performance.  So, in other words, you walk them in the shoes of an African American or a Latino or a woman in math, and they experience the same performance detriment.  So, this argues strongly for the situational nature of underperformance, at least that there s an element of situation when we see people underperforming on tests.  It s not all what you know, it s what you know, plus what you can produce in a situation that may be loaded.  Some general conclusions from 200 published studies.  Impairment occurs both on tests and in terms of GPA.  Impairment on test results from anxiety, reduced working memory capacity, impaired self-regulation.  It s not typically a function of giving up on the test.  MRI studies show that women appear to use different regions of the brain when they re under a stereotype threat.  It affects both elite and non-elite students.  If you motivate the non-elite students, so that they really worry about not doing well, they will under-perform, typically the effects have been found only with elite students who care so much about looking smart that the stereotype threat undermines their performance.  It can arise as a function of simply mixing students together.  If you have a group of women taking a test, and you introduce men, the women will start under-performing as a monotonic function of how many men are in the room.  It s been found in field studies as well as laboratory studies.  Some studies look at how interested women are in mathematics and science careers after they ve read a study that talks about biological differences in math.  What you find is that the women show less interest in math and science as a career.  So these things can mediate more than just performance, but also attitudes about one s aspirations and a fit in the field.  Lots of people have been talking about mental rotation.  Biological differences in spatial ability are, seems to be the area in which those biological differences are most reliable.  For example, this is a classic task called the Vandenberg mental rotation task, you look at the left hand item, and you try to match two of the other items to it.  This is the part in the talk where I typically lose people because they re trying to solve the problem, so, I ll just get rid of it.  [Laughter].  I ll just leave you hanging.  I m Jewish, I m not very good at these mental rotation things.  [Laughter].  The point, again, is that this is incredibly biologically linked effect.  This is the one people point to.  What we wanted to find out is, whether that s the whole story in performance here, whether mindset doesn t also play a role here.  So we did a simple study with elite college students on the northeast of the United States, I m not telling you which college they re from, it s not NYU.  In the controlled condition, we asked them a question right before they took the test, which was designed to produce a thought about something totally irrelevant to the test.  And, so, that creates the baseline condition here.  In another condition, we do something that actually widens the gap.  We ask them questions that make them think about their gender.  And here, that gap widens considerably.  The good news is that if you remind women and men that they are students at an elite liberal arts college in the northeast, you close the gap.  So, my research assistants swear to me that no testosterone was administered to any of these students.  So, it seems like mindset here can matter a lot.  Larry Summers talked about the extreme end of the distribution, and I think that s been a point of confusion.  Thus far, I have not been talking about those students, and, we wanted to see if stereotype threat can affect even people at the highest levels that we could find.  We ve shown it in White male engineering students, but not with women.  So we went to a large university and found the hardest math class we could find.  If you go into a class like this, here s what you see, 90 percent males, many of them Asian, and then a small cluster of women.  Those women have made it through a lot of stereotype threatening situations.  And, so, we actually did this study with the cooperation of the professor.  We gave them a difficult test based on the material they were learning in class, but a very hard version of it, and then we did something, we gave half of the students the test in the way you normally got the test, the professor says, this is going to count towards your grade, the other half, he added the statement, that, by the way, I ve given this test a lot in the past, and the girls do just as good as the boys.  They re men and women, actually, so he didn t say that.  What is so fascinating here is that in the controlled condition, the women are performing just as well as the men, but when you tell them there s no gender differences on the test, the women out-performed the men, because these are women who are the cream of the crop, and, so, when they re in the normal stereotyped threat environment of that class where they re outnumbered, they re being suppressed to equality with men.  They are actually smarter than the men in this class based on their test score here.  Another way to look at this is to see what happens when you do an intervention where you teach people things that theoretically should reduce their sense of stereotype threat.  I sense that I m running out of time, so, 2 more minutes.  One of the really interesting effects in the sex differences literature is if you go overseas to Japan you find that the sex difference is still there, but it s very small, but the girls in Japan are like way better than the boys in America.  So, one idea that we had was that what do Japanese people have that Americans don t?  Well, one of the things they have that s related to stereotype threat is a very different attitude about what intelligence is.  If you ask an American kid, what is intelligence?  They ll say, somebody who doesn t have to work hard.  You go to Japan, by and large, they say, somebody who works hard.  Very different.  And Carol Dweck has shown over and over again that this difference in your conception of intelligence matters a lot.  So we, essentially, one of the things we did in our intervention, was teach these American kids in a middle school to be more like Japanese kids.  We changed their conception of intelligence.  We had another role model condition.  I won t go into that.  And a third, controlled condition, in which students learned about drug abuse.  So they had two different intervention conditions, and a controlled condition.  And then we looked at the end of the year at their statewide standardized tests that we got from the state.  What we found is in the controlled condition on the math test, gender effect, where boys are doing better than girls, in the malleability, Japanese, think like a Japanese person condition, a big closing of that gap, and, even better in the role mode condition.  So we intervened away this sex difference on the 7th grade math portion of the statewide test by attending to stereotype threat.  Finally, a field study conducted by the Educational Testing Service.  As you may know, if you take the AP calculus exam, before you start, you write down your race and your gender.  This has been shown in a couple of studies to take an otherwise non-evaluative situation and make it feel evaluative to a test taker and lower their performance.  So, ETS wanted to find out if it would make a difference in the real world.  What they found was, yes.  If you ask about gender before taking the calculus test, it hurts girls.  You can see it from the formula score.  Some of the participants in this actual study taking the real AP calculus exam had that information moved to the end.  And that s what happened.  The females do a little bit better than the males.  A re-analysis of this data by an independent researcher made the following statement based on the numbers provided in that study.  Women benefited substantially on the calculus test when demographics were asked after testing, rather than before.  This simple, small and inexpensive change could increase U.S. women receiving AP calculus, AP credit, by more than 4,700 women every year.  But you can also see the dilemma that ETS would be in because they would have to give it to fewer men because the phenomenon flips because men get an advantage out of being reminded of their gender before they take the test.  In conclusion, human beings are both biological and cultural animals.  Whatever the origin of group differences, they are not fixed.  Cultural ideas about those differences can exacerbate them or lessen them, and intervention can boost performance and nurture intelligence so that biology need not mean destiny.  Thank you very much.  [Applause].   Amy Wax:  Thank you for inviting me here to comment on stereotype threat.  Before I get started, I m under very strong time constraints here, I do want to thank my Research Assistant, Jason Levine, he s a Penn Law student.  He s in the back.  Jason, please stand up.  He s a brilliant, energetic, and numerate, and that is, best of all.  I am a law professor, so what am I doing here talking about stereotype threat?  Well, I am someone who uses social science to address questions of law and public policy.  And, in that role, I often think about issues of inequality, and inequality and gender.  So, I don t produce social science, but I am a consumer of social science.  I like to think I m an educated consumer.  And, as you know, the educated consumer is your most important customer.  Now, I noticed in looking at the social science of gender, a good deal of talk about stereotype threat, any many extravagant claims made for its significance as a kind of global, all-purpose explanation for observed disparities in performance.  Here I m going to confine my discussion to gender differences, and, in particular, gender differences relevant to math and science.  I recognize that the stereotype threat literature is much broader than that, but much of what I say will be pertinent to the literature generally.  Now, why is it that stereotype threat is such a darling of gender-minded psychologists, and there s been so much enthusiasm?  Well, stereotype threat is appealing because it doesn t posit any temperamental or cognitive [& ] differences between the sexes, but rather attributes observed disparities to an understandable response on the part of individuals to unjust and unjustified social signals and stereotyping, and it also promises something of a quick fix if you remove or dispel those signals, then the bell curves will magically come together and converge.  Sounds promising.  So I decided to investigate for myself.  My investigation led me to the conclusion that on the current state of knowledge, the claims routinely made on behalf of stereotype threat are vastly exaggerated.  Stereotype threat research is a classic case, to use a phrase coined by one of my colleagues, of over-claim syndrome, exaggerating the social significance of scientific observations, an exaggeration unjustified on the actual data.  To summarize, there is no basis, not yet, anyway, for identifying stereotype threat as an important, significant or substantial factor behind gender performance differences, and foreseeing it as anything more than a sort of sideshow, a small quirk, a small tail wagging a large dog.  In addition, there are contradictions and anomalies in the theory and the literature that seriously undermine its cogency at this time.  Now, don t get me wrong, I m not denying that there is such a thing as this effect, which scientists like Josh Aronson and his colleagues have observed.  What I am doing is I am raising questions about methodology that go to significance, how important is stereotype threat in explaining the patterns that we see in the real world?  Finally, in the last part of my talk, I want to propose how to redesign stereotype threat studies so as to correct some of these defects.  So, what are the problems that I see in stereotype threat?  Research.  Just to summarize briefly.  Relative magnitude, what is the size of the ST effect, compared to the gender gap in performance overall?  I submit that the stereotype threat literature provides us with no information on that very important question.  Baseline yardstick, what is the background measure of mass scientific ability independent of the influence of a stereotype threat, and, against which stereotype threat is assessed?  Threat in the air, this goes to methodology, are special interventions required to dispel stereotype threat, implying ST is just out there most of the time, or is special interventions required to create it, implying it s not out there?  The studies are all over the map.  Fourth, cherry picking, why do, and there s a lot of cherry picking in this field, why do women do as well or better than men in some math measures, and we have a lot of talk about grades and classes and that sort of thing, but do worse on standardized tests and in life performance and academic performance generally?  Finally, pervasive disparity, can ST account for the full range of non-math, science gender performance differences?   Okay, so to just go through these in something of detail.  How do we do gender stereotype threat studies?  Well, we do race and gender studies, and studies looking at all sorts of parameters pretty much the same way.  We, for gender, take a group of college students, or some group like that, usually in a restricted range of ability, we divide them into men and women, we jigger it so we have equal numbers of men and women, then we divide the men and women again, and we give one group the test under the so-called diagnostic condition designed to elicit stereotype threat, and the other group the test under the controlled condition.  Now, we have the problem that students in each group might differ, on average, in ability level, even with the restricted population that we are drawing from.  So what do we do about this, because we can t compare test results across groups unless we have some kind of baseline?  Obviously we can t give the test to the same person twice under different conditions for reasons I won t go into but should be obvious.  Well, we can do it in a number of ways.  We try to pick people who have taken the same math courses, who have gotten similar grades, stuff like that.  But, really, the best way of doing it is to statistically adjust for some sort of test of math ability.  And what is it that most researchers use?  SATs.  They control or they adjust for SATs, and then they report their data.  Now, what are the problems with doing that?  Well, there are a number.  Number one, it generates data that creates the appearance that men and women are roughly equal in quote/unquote true math ability, and, thus, creates the impression that any disparity in performance is due largely and entirely to stereotype threat.  Secondly, and this is in some ways a more important point for my purposes, this very design obscures, discards, leaves on the cutting room floor, critical information needed to compare the size of the stereotype threat effect to the size of existing gender disparities overall so we get a sense of the relative magnitude of the effect and the background gender disparity.  Now, the first point, right, I ll go to the next, the first point was made, and can be made, with regard to the race study, and here we have the classic 1995 paper, the point was made by Paul Sackett, in comparing the Black and White Stanford students we did not report the raw scores, Aronson and Steele, rather they controlled for SAT.  Now, you know, and I know, that Blacks and Whites have a significant SAT gap.  That SAT gap is about 200 points.  It s largest among the highest ability students, and here we re talking about Stanford students.  We don t know whether the Blacks and Whites in this study had the same SATs.  We doubt they did because we wouldn t need affirmative action if they did.  But we control that data away.  And all we re left with is the control data.  Now, they do say control by SAT, I was surprised that when Aronson showed this data in his presentation he didn t even include that it was controlled by SAT, which is amazing.  Joshua Aronson:  But they weren t.  Amy Wax:  But this was.  This says control by SAT, adjusted by SAT, unless I ve got the slide from the wrong study here.  But let me just, this is the wrong, you did publish this paper?  Joshua Aronson:  The data I just showed was uncorrected by SATs, so, the fact that Aronson didn t show adjusted by SAT reflected the truth.  Amy Wax:  Okay, well, I m talking about this paper, this one, which was adjusted by SAT.  So, let s just talk about this paper, and then we ll talk about the other paper in a moment.  All right, so, that s the 1995 paper, where Aronson and Steele control for SAT.  Now we have a similar paper involving gender that is also adjusted for SAT.  We have essentially the same problem in terms of information loss here.  We don t know what the absolute SAT levels were for this group.  We don t know how large a difference in ability we have because we re sort of extracting it away.  Now, even apart from adjusting from SAT, the studies that are presented where SAT is not controlled, we do have a background fact about the methodology, which is, that when we draw a sample of subjects for a particular study, there is no guarantee that that sample of subjects is representative of the distribution as between, let s say, Blacks and Whites or males and females, as a whole.  In other words, we may have Black students who represent a more able group than your average Black student compared to White students who represent perhaps a less able group.  The point I m trying to make is that we don t have a methodology that attempts to draw subject samples that are an accurate reflection of the background population.  So, because of that, we cannot actually know what the relative size of the effect is, relative to the background gap.  That is the point that I m trying to make here.  Well, what do we actually see as a set of background facts?  For males and females on the math SAT, I don t need to belabor it, we have, in many years of SAT administration, ratios ranging from 4 to 1, to 13 to 1 for the very top scores.  Here I m talking about 780 and above, 800 and above.  The really interesting question, the real reason we re interested in stereotype threat is to ask, how much of that is actually due to stereotype threat?  In other words, if we could get rid of stereotype threat, would these bell curves magically converge, would the ratio go from 6 to 1 to 1 to 1?  That is the question that we re really asking.  Now, let s go back to the problems here.  Baseline yardstick.  What is the background of math/science ability independent of the influence of ST?  What are we using to compare men and women, or, of the stereotype threat group to the non-group?  Consider the following statement in a 2002 paper by Marks and Roman.  ST studies show that eliminating stereotype threat, quote, protects women s math test performance, and allows them to perform at the same level as equally talented men.  Equally talented men.  Well, here s the question.  How do we know whether and when men and women are equally talented?  We need to know that if we re going to get information about the relative size of the ST effect versus the gap overall.  We definitely need that.  One way is to give a test to assess their ability.  The test that s usually given is the SAT.  It s often used as a control in these studies, not always, but often.  Now how do we know that SAT, though, is an objective test of ability, or whether it itself is tainted by stereotype threat?  This is a hall of mirrors.  Right?  We can t have it both ways.  I mean, either stereotype that explains SAT and like disparities, in which case it s tainted, or it s an objective measure of ability, and we can use it as a baseline.  Right?  Indeed, the confusion on this creates anomalies.  If stereotype threat explains the gender gap in tests like SAT, we should be comparing men to women with lower SAT scores, right, on the theory that women scores are artificially depressed.  I don t see the studies doing that.  Or, alternatively, if we compare men and women with the same SAT score, women ought to leap ahead of men if we dispel the threat.  I don t see that systematically in the literature.  A few minutes.  I am going to sort of skip over the cherry picking, I m going to go to pervasive disparity, and then I m going to propose how I think ST research should be done.  Pervasive disparity, we re told that it s a stereotype threat as a matter of math and science.  The excerpt we have from a study by Halperin, et al, says women are not affected by verbal stereotype threat, they don t think that they re inferior in that regard.  Well, I did a little study.  I looked at sort of the main magazines of opinion, which I read compulsively, Atlantic, New York Review of Books, New Yorker.  Guess what you see, men to women ratios of authors, 8 to 1, 9 to 1, 10 to 1, the same ratios you see at the right tail of the bell curve of the math test.  Well we certainly won t be very hard to attribute that to stereotype threat.  I suppose we can say it s discrimination.  But, I do find it hard to believe that a bona fide left-wing politically correct journal like the New York Review of Books discriminates against female writers.  Maybe they do, but, once again, we are seeing these vast differences in life performance, in drive, in production, in productivity.  I think it s very hard to explain those with stereotype threat.  So how should we be doing stereotype threat research?  Let s start with a hypothesis.  Addressing the unanswered questions, the math SAT disparity at the right tail of the bell curve is due entirely to stereotype threat.  Corollary, absent stereotype threat, and assuming male and female SAT test takers are similarly representative at the top, which I think is fair, the bell curves for male and female math SAT performance will converge if we got rid of stereotype threat.  So this is what we would see.  Now this is what we do see.  Here s what we ought to do.  Take a score, 750, ask how many men, what percent of men get that score, find the corresponding score that the same number of women get, or, rather, the same percentage, find, do a study that takes a sort of well-distributed representative sample from 716 and above, men 750 and above, and we ll have to jigger those samples a little bit, it gets a little complicated, but the concept is simple.  Give them a test under non-stereotype threat conditions and they ought to do the same.  This is what we should see.  And that is the study we should do because that will tell us how much of the disparity is due to stereotype threat.  If we do see a convergence, we know it s all.  If we don t, well, the good news is that the SAT is indeed an objective measure of ability untainted by stereotype threat.  The bad news is, the SAT is indeed an objective measure of ability untainted by stereotype threat.  Thank you.  [Applause].  Joshua Aronson:  May I cross-examine?  [Laughter].  Christina Hoff Sommers:  It comes in handy, I guess.  Yes, go ahead.  Joshua Aronson:  Could I ask you, what is the purpose of an experiment?  Amy Wax:  To answer a question.  And the question I have is, how much of the disparities, this is a quantitative question, how much of disparity we see in standardized test performance, productivity as scientists, promotion rates between men and women in science and math, how much of that can we quantitatively attribute to stereotype threat, and how much is attributed to other?  That s a very simple question.  Joshua Aronson:  And it s a question totally inappropriate to laboratory experiments.  They cannot give you that information.  Amy Wax:  Well then the stereotype threat literature, I think, is not very interesting.  Joshua Aronson:  Well, you re entitled to your opinion, of course, but let me tell you what an experiment is for.  An experiment is for testing a theory.  To test a theory.  And to get angry at an experiment because it does not tell you what goes on in the real world is a little bit like getting angry at your lawnmower because it doesn t peel onions.  It is not intended to give you that kind of information.  It is simply to test a theory.  And what the theory says is that human intellectual performance is responsive to environmental conditions.  Study after study shows that.  It doesn t matter if you control for SAT or not.  So I think you have a very, you re expressing a deep misunderstanding about the nature of this kind of science.  Now there is a way to test how much stereotype threat matters in the real world.  I think the study you want to do is to measure people s propensity to experience it, and then give them a real exam that they have to take.  That s one way to do it.  You can enter it into some kind of regression formulation, and then get the amount of variance accounted for.  You don t get it from experiments.  So I would say that rather than over-claiming, you are misunderstanding what the nature of social science research is, and I think that it s as if you didn t see the talk that I gave because I showed you three kinds of evidence that allows you to make a Gestalt type inference about whether this matters in the real world.  You test the theory about how human beings minds work, you do survey research that shows that people who measure high in stereotype vulnerability also do worse in GPA, that s been done by several researchers now, and then you do interventions to see if in the real world you can make the effect go away with targeted interventions.  So I think to say that there s nothing there, that it s over-claiming, is to sort of just disbelieve what you just saw.  And that s fine if you disbelieve it, but I don t think that, I think it s pretty sound evidence that something is going on there.  Now, I cannot tell you at this point how much of the gap is caused by stereotype threat, but I find that that s not even a terribly interesting question.  It s sort of like saying, how much of, what if I told you nine.  It s nine.  Would that change the way you behave towards this thing?  You ve written in the Wall Street Journal that it s parents, it s broken Black families, and I agree with you.  That s what you ve said in the Wall Street Journal, that unstable families.    Amy Wax:  I didn t say anything about stereotype threat and unstable families.  Joshua Aronson:  No, no, no.   Amy Wax:  You re talking about two different articles.  Joshua Aronson:  What you said is, look, stereotype threat is this trivial thing, it s a gravely flawed research, I ve yet to see how controlling for SAT flaws it.  I showed you the results without it.  The results are pretty much the same.  If you relax people during a test, all of a sudden Black students know more.  That is amazing to me.  That s absolutely amazing to me, and the fact that 200 studies show the same effect, that for you to sweep it away with something that you wrote in 2004 and you remain unconvinced by data just, it suggests to me that you ve got some kind of axe to grind about this stuff.  You didn t say, for example, in your paper, well, parental influence, that accounts for everything.  No.  I would have jumped all over you for that.  But you said that s what it was, but you didn t include all the other factors, and now you turn around and you say, oh, well, stereotype threat can t explain everything, therefore it s a worthless proposition.  That s ludicrous.  We always talk about lots of [inaudible].  [Applause].  Please, don t do that.  Amy Wax:  Can I get my slides back up?  I m sorry.  I actually have to disagree.  I asked a very simple, extremely basic scientific question, which, well, I started with a hypothesis, which is the way scientists start, which is, let s assume, or, is all of the disparity in standardized tests that we see and we ll just confine ourselves to the right tail of the bell curve, which is where the action is, and where the disparities are, how much of that is due to stereotype threat?  That s a very simple quantitative question, which I think it is fair to wonder about.  And, I mean, the reason to wonder about it is that it really very much affects policy, it affects where we direct our efforts.  Just, going back to the race context, we have a 200 to 250 point disparity in SAT scores between Blacks and Whites.  That s a fact.  I didn t cause that to be true and I didn t make it up.  Now, if that is all due to stereotype threat, than that points us in the direction of interventions that are radically different, from 5 percent of that difference is due to stereotype threat, or 3 percent, or 10 percent, or 20 percent.  I mean, I think it is not unfair to ask about relative magnitudes.  This doesn t strike me as rocket science.  We re always asking about relative magnitudes.  I mean, medical scientists ask about relative magnitude and they expect an answer.  If it is 5 percent, or 3 percent or 2 percent, then I think we ought to be directing our efforts elsewhere, such as, maybe, you know, as I tell my students who say, why do you think some people do worse on the LSATs?  I say, well, here s the reason, they don t know the answers.  That s a good working hypothesis.  I m glad to be proven wrong, but I think that s a good first cut.  Joshua Aronson:  So is your contention then that knowledge is all that matters?  Amy Wax:  I don t know.  Okay?  I don t know, but I would like to know.  Is it 80 percent knowledge, 20 percent anxiety?  Is it 50/50?  Is it 90/10?  That is what I d like to know, and even if you, the fact is that until you draw your experimental subjects in a way that is representative of the background population, you will never know.  Christina Hoff Sommers:  I think we re going to open it up for questions.  I just want to point out that, my hobby horse here, criticizing the NAS study, it talks about stereotype, I know, and I m sorry, but, it s influential and it s quoted everywhere, and it just, here we ve heard this brilliant exchange, and it s reduced to just, kind of an outright statement that this stereotype threat exists and they footnote all the research that supports it, and none of the research that questions it.  So, that s all I m saying, is that, what we heard here is not reflected here.   Joshua Aronson:  What exactly is the research that questions it?  Let s talk about that research, okay, because I think that, let s make these [cross-talking] and what data did they use to come to their conclusion?  Amy Wax:  We re not questioning it, we re questioning its significance, which is a completely conceptually different matter.  Okay?  Joshua Aronson:  No, you called it methodologically flawed, gravely methodologically flawed.  Tell me the methodological flaw in that experiment.  Amy Wax:  Because it doesn t answer the question that we need to have answered in order to know whether to incorporate this into our policy recommendations.  Joshua Aronson:  Again, you re asking an elm tree for pears.  Amy Wax:  What do we do about the gap?  What do we, you know, we have a practical orientation.  I m a lawyer.  I m a policy person, what shall we do about the gap, if anything, that has to be informed by social science research?  If this problem is that people don t know the answers, they really don t have the knowledge, then that points to a different set of measures than they re anxious, we re stereotyping them, we re sending them the wrong signal if we just tell them that they ll do just as well as everybody else, bang, we ll get this.  Joshua Aronson:  Well, you re cherry picking now, which, I hate to borrow your term, but there are three interventions out there based on this theory.  One reduced the Black/White achievement gap in middle school by 40 percent, without knowing how much of that achievement gap was caused by stereotype threat, an intervention based on the theory reduced the gap by 40 percent.  You know how much it cost?  Amy Wax:  Was your sample representative of the Black/White gap overall?  I mean, we don t know anything about your sample.  There may be a restriction in range.  There may be a huge restriction in range.  We don t know that Blacks are representative of Blacks in general, Whites are representative of Whites in general.  I mean this is basic sort of statistical stuff.  These are basic points.  This is not fancy stuff.  Joshua Aronson:  We picked at random, three schools in three different parts of the country, we do the study, they re middle income kids, they re Latino, and girls, they re normal kids.  The answer is not is this representative, but, is it replicable, can you do it in other, and, you re not saying, hey, wow, this is cool, we should replicate it and make sure, you re saying, I don t, I want a different kind of data, I want to be able to put a number on the amount, and that just seems like we re way past that point.  We ve already figured out how to intervene in a positive, inexpensive way, and it s sort of like you just don t like the answer.  Christina Hoff Sommers:  We ll come back to the panel.  Let s take some questions from the audience.  Yes?  Female Audience Member:  I don t understand how this works so consistently in one direction.  I would think if you re primed to believe that your sex, or the other sex, or whatever, does better on the test, that might, in some cases, inspire you to work harder, try harder, and wouldn t trying harder make you do better in some cases?  Why does it always make you do worse?  Joshua Aronson:  That s exactly why you find the GPA difference and test score gap difference with women.  When you try hard, when you add extra motivation to a timed high-stakes test, the typical response is to under-perform.  It s called the Yerkes-Dodson Law, that complex tasks, under high arousal, are undermined.  However, that same cue to women, that they may be falling behind, may make them study harder, do their homework, check their notes, take better notes in class, and that s the exact pattern you get with women, is that they do as well as men in their mathematics classes, but worse on the test.  So, that fits the stereotype threat data pretty well.  So, it s not a question of, it s a question of two different task domains that make all the difference for this kind of phenomenon.  Dr. Jacquelyn Black:  Dr. Wax, you said earlier this morning that you had not experienced any discrimination, any sex discrimination as you came through, I would not like you to go home from this conference thinking that that is typical.  I would like to give you one example of stereotype discrimination.  I walked into, at the University of Chicago, a graduate program, and the Professor walked in, and he looked at me, and he said, you, you re female. I said, yes, sir, as I looked around the room, I was the only woman in the room.  He said, why are you here?  I said, because I want to learn animal ecology, your course.  He said, there is no place for women in science, get out.  That s about as straightforward as you can come.  I said, no sir, being young and foolish, and I sat there.  He said, well, then, we will teach you that there is no place for women in science, and when you finally learn you will drop this course and go away.  I suffered through that course in many, many ways.  I was told that I must load all the fieldtrip equipment onto the bus, including the heaviest machinery there was, and none of the 20 men were to lift a finger to help me, when I learned that I could not lift everything, I would leave and go away, where all the other people who didn t believe in equality here.  I went on the fieldtrips, it was March in Chicago, it was snowing, I fell in the water sailing up a stream, I pulled one end of the stream every single time, while 22 men rotated at the other end of the stream.  I was soaked, I was freezing, I was shivering, I begged to go out to the bonfire on the shore.  He said, if you drop the course you may come out by the fire.  In the end, did any other woman from the entire division of biological sciences go into his area of science?  No.  A whole, several decades of women were kept out of that specialty of science.  Eventually this man became President of the American Association for the Advancement of Sciences, AAAS.  Do you think that he was going to help women in his position?  And what about those other 22 men who watched him treat me this way and get away with it?  Did this color their thinking?  I think there is a lot of stereotypic threat out there, some of which is, perhaps, a little better disguised than what I had, but it is for real.  [Applause].  Christina Hoff Sommers:  Next comment, or, could you repeat the question?  No no. Yes, in the back.  Constance Holden:  For Mr. Aronson, I mean, part of this, I think is the following, everyone in this room has taken a boatload of standardized tests, and we re processing this information anecdotally to see if it makes sense.  I d have to say, from my experience, the fact that I m filling out something before I take a calculus test as to whether I m male or female, that that s going to make a difference in how I do on that test, I find that very difficult to believe.  I also find it, in a way it s immensely heartening to know that from your point of view, if you just remind a man that he s a man, by having him fill out this circle, that he s going to do better on the test.  I mean, that is, if this were true, I would certainly endorse the idea, let s have the men and women take the tests in different rooms, and the men could fill out their gender before they take it, and the women could fill out their gender after they take it, and let s have everybody take the, you know, get the best score possible.  In a sense, the significance of this, it almost defies belief on its face.  You say one sentence to somebody before they take a test and it s going to completely differentiate the score.  Anyway, let me get to more of the pointed question.  A lot of what you re talking about seems to go to test-taking anxiety, that some people do well taking these high-level standardized tests, and some people don t do as well, and I suppose it could be true that women, on average, don t do as well, except it hasn t stopped them from largely taking over the legal profession and doing very well in medicine.  I think that s the reason why the biological seems to be such an attractive and likely answer for what s going on to women in science.  Why isn t part of the disparity, when you say different things to men, for example, particularly, when you say the score doesn t count, we just want to talk to you about taking this test, why aren t you seeing their efforts reduced, and that, to an extent is explaining the differences in the scores, in the gap?  Joshua Aronson:  Well, I think I want to comment on your first point.  I think maybe you ve hit on why Dr. Wax finds this research uncompelling.  We heard her list of incredible accomplishments, maybe she just never experienced this on a gut level.  I think that for non-intuitive research findings to be accepted, it s sort of, it helps if people have experienced this.  So, everyone who has experienced this sort of goes, yeah, yeah, and then there s people like you who go, I just don t see it.  If we have to rely, in social science research, if we can only believe in studies that feel intuitively right to us, we re in big trouble, because so much of human psychology takes place under the level of conscious awareness.  Every student, I did a study published in the same package of studies in which we took these African American and White students and we had them, we totally relaxed them because we told them we re not measuring your intelligence, but, for half of them we asked them to indicate their race right before the test.  So this takes a relaxing experience, where you re just problem-solving, and now it s a racial issue.  The people s anxiety went up and their performance went down.  After the test, I said, did it bother you that I asked you to write your race?  And maybe 2 people out of 100 said, oh yeah, I didn t really like that.  Do you know what most of them said?  They said, no, I do it all the time.  The ETS study did the same thing, and you saw the difference in the effect.  So people don t necessarily feel these things, that s why we can t go on the basis of our guts when we talk about research results because often they are not things that we can actually feel.  As to your other question, why don t people not try when you tell them their scores won t count, I m not sure what you re talking about, but I would like to see the data that you re talking about.  Amy Wax:  Can I comment?  First of all, what I feel and what I don t feel is completely irrelevant, although I will say that certainly I ve felt test anxiety and I am not denying that test anxiety makes a difference some of the time.  I m not denying that test anxiety is irrelevant.  So let s just get that straight here.  I m really making a very different, once again, quantitative point.  How much does test anxiety count overall in the sample populations, these large samples of test takers over years and years?  How much, how important is it, compared to how much you know, developed aptitude, and the like?  That is the $64,000.00 question.  We just, that is the information that we need.  I think that when stereotype threat morphs into anything like the claim that if only Blacks were unaware of or could somehow get rid of the notion of their own inferiority, we would see, you know, we would see Blacks and Whites with scores like, this is what we would see.  That 200-point SAT disparity would vanish.  Anytime anything in the literature approaches that claim, I think, this just is a distortion, and it s a really upsetting and disturbing distortion, the same for males and females.  The disparity at the right tail, you know, that s just stereotype threat.  Joshua Aronson:  Could the audience members who think I ve said that stereotype threat accounts for the entire gap please raise their hands?  All right, Amy, I think [cross-talking].  Amy Wax:  Small, large, middle-sized.  Female Audience Member:  Hold on a second, I do want to say, because, even though you didn t, and, by the way, I m enjoying this discourse because I hear what both of you are saying, but I want to tell you from a K-12 perspective, we have studied, I have studied your work, Dr. Aronson, and Steele, and the Steele brothers are twins, and they have different attitudes and beliefs about stereotype threat, but I will tell you from experience, people have read your work and said, see, it s their own damn fault why they are not excelling, these Black students are not excelling, and they have interpreted it very wrong.  So I agree with Dr. Wax, that there has to be quantitative factors, also understanding other mitigating factors that contribute to student achievement.  The other conversation that I keep hearing is that we lump gender and race as if they re two distinctly separate things, being one of both groups we need to start having a different level of conversation because you cannot generalize about all women and all people of color because there are women of color, men of color, and we need to start thinking very differently about how this information gets translated into K-12 education, because that is what s happening here.  Someone reads your work and says, that s it, he said it, it s gospel.  Joshua Aronson:  That s interesting, because you re saying that people read this work and say, it s all the kids fault?  Female Audience Member:  You bet.  I have had heated arguments with people, similar to Dr. Wax, saying, show me what else, show me why you would say this is the only factor, because they ve taken it as gospel.  This is the attitude, well, see, and I m talking, I ve worked in a very highly affluent community with highly educated educators on the east coast, outside of Boston, that truly believe this because you wrote it, you and Dr. Steele wrote it.  Joshua Aronson:  Well, we wrote something, we didn t write that the entire achievement gap was caused by stereotype threat.  We certainly didn t say that.  In fact, if you read my piece in Educational Leadership, I quote H.L. Mencken saying, anybody who thinks this has to do with one thing is out of their minds.  I am tired of people saying, for the record.  I am tired of people saying that we think stereotype threat is the primary cause of the achievement gap.  What I like about it is that it s probably the most actionable cause of the achievement gap, that s why I study it, not because I think it s the one thing, but I think it s the thing I find fascinating, the thing I ve experienced myself, the thing that does, indeed, cut across lots of different groups, because we may be all different, but we all belong to the same human race and we all want people to think well of us and we want to belong and we want to feel that people feel we re smart.  So, that ties us altogether, and we see, in every group we ve looked at that we ve exposed a stereotype threat, the same kind of effect, even White males at an elite college experience it.  Amy Wax:  Can I say something?  I think one of the dangers here is that stereotype threats, a very, sort of popular explanation for what s going on, and there is the danger that attention to one thing detracts from another.  It s easier to pay attention to stereotype threat than to do the hard work of actually reading, learning, studying, and perfecting your skill in math and verbal areas.  That s a lot of hard work.  I think we would rather pay attention to stereotype threat because there is a quick fix for it, just tell people that, you know, that gender differences don t matter, or this test doesn t show gender differences, jigger the setting in which you re giving the test, dispel anxiety, and that can substitute for the grinding work of really learning something.  Joshua Aronson:  I ve improved the test scores of lots of students by this knowledge.  Tell me what you ve done with your proper focus on the problem?  Amy Wax:  But how much, relative to the test score disparities that exist [cross-talking]?  Joshua Aronson:  Well, we saw the data.  Amy Wax:  What is the relative magnitude?  Are you getting a 5 percent improvement, a 2 percent improvement, a 20 percent, a 100 percent closing of the gap?  That is where, tells us where our efforts should go.  Male Audience Member:  Just a quick question.  Christina Hoff Sommers:  Last question, and then we move on.  Male Audience Member:  Professor Aronson, you didn t say anything about Professor Wax s hypothetical study, I mean, she thinks this would shed some real light on what concerns her, do you have any objection to the way she outlined it?  It just doesn t interest you?  What was your reaction?  Joshua Aronson:  It s not, the way she s outlined the study, if I understand it correctly, is to look at the top end.  Different ends of the continuum.  I don t know, I m excited about all research that sheds light on the subject, so I have no objection to it.  I was, I have to tell you, I was very caught up in the fact that she didn t seem to be taking in any of the, what I consider to be extremely helpful data showing that we can boost grades and test scores by understanding this stuff.  I mean, that seems to be the ultimate, that, when people do studies they do it so they can get smart enough so that they can go in and help people.  That s my perspective, is that you do the experiments so that you get informed so that you can go out into the world and understand situations in humans so that you can help them.  And I don t see anything in my own research that suggests that I m wasting my time.  Christina Hoff Sommers:  I just want to ask one quick question, which is that there s been a lot of research on self-esteem, and a lot of parents put effort into helping their children internalize very positive stereotypes about themselves, positive views, I m great, and we ve had years and years of self-esteem education, and American kids now have the highest, as a group, probably the highest self-esteem in the world, and yet, they do worse on many tests compared to kids in Korea who have lower confidence and sense of themselves as math takers for example.  Maybe your work doesn t bear on that, but, to me, it would just, I wonder if you thought about?  Joshua Aronson:  Well, self-esteem is involved to the extent that human beings want it, and one of the ways you get it is by fitting in with other people.  The worst blow to someone s self-esteem is rejection.  That has an evolutionary base.  And, you, so, these programs that say, everybody is wonderful, and they puff them up with praise rather than lead them to high achievement, which, incidentally, I also believe in high achievement and studying, that, real self-esteem, grounded in reality, grounded in achievement, is what you want, but you have to, what you have to be careful of, is, you have to attend to self-esteem because kids will get it one way or another.  What we find with Black students is that when you don t create the conditions that they feel, where they re pushed to excel and they feel comfortable and safe, and they like school, they will find, and get self-esteem that way, they will drift into social relationships as the primary foundation for self-esteem, and that s dangerous.  So, students will get it one way or another, and you have to be attentive to that.  Christina Hoff Sommers:  Amy, do you have any final on that?  Amy Wax:  Yes, I ve looked at the self-esteem literature, I m radically under-whelmed by it, mostly because self-esteem is so ill-defined, and it doesn t really correlate with the stuff we care about.  I think the literature that s more promising, in my view, is literature on locus of control, which is, do you believe that you have control over your life, your destiny, your own achievement, as opposed to, do you believe that others control, or it s just a matter of luck, or, discrimination, or, the way you re treated by outsiders?  And students who believe that they can self-improve and can control their own destiny, they re the ones who succeed, I mean, that s a vast oversimplification of the literature, but it is very interesting literature.  Christina Hoff Sommers:  Well, thank you very much, very enlightening and exciting session.  [Applause].  We move now to Charles Murray, summing up, and the panelists from earlier today, anybody can come up and have their say, we ll have a regular open mic.  Oh, let me introduce my colleague properly.  Charles Murray is the W.H. Brady Scholar of Culture and Freedom at AEI.  He s the author of many brilliant books, too numerous to mention, Losing Ground, The Bell Curve, my favorite is, Human Accomplishment, The Pursuit of Excellence in the Arts and Science, 1800 BC to 1950.  I highly recommend it.  I ve been at AEI for more than ten years and I have heard Charles speak on many occasions.  I always look forward to it.  He s always interesting, exciting, I learn things, and today he has the job of summing it up, all up, telling us what it means, and why it matters.  Thank you.  Charles Murray:  Thanks for that topic, what it all means and why it matters, that was what I was told I was supposed to talk about.  First, I just want to say that it is very nice to have spent this day listening to people talk about this issue with data, to argue about it in terms of what we know and what we don t know, and I think it is a model for the kind of ways this topic should be discussed.  I congratulate all the panelists.  That s preface to saying that I have so little data I don t even need PowerPoint.  I am going to try to talk about what it all means and why it matters, but I cannot do that from 30,000 feet with an unprejudiced, unbiased view.  I have my own opinions.  They will not be shared by some of the panelists.  I m going to limit my remarks to under 20 minutes, which will give us some more time for the other panelists who feel moved to respond, to do so, and I think that perhaps Professors Barnett and Spelke and Aronson might be the first to be so moved, and I will call on them first.  Does it all matter, and why does it all matter, and how much does it matter?  Well, the good news is, I think it matters a lot less now than it did 50 years ago.  Fifty years ago, coincidentally, coincides when my 18-year old older sister was a freshman at Iowa State, sitting in an engineering course, and having the professor look directly at her and say in front of the entire class that he didn t think women belonged in engineering.  That doesn t happen anymore.  To the extent, well, I won t say it doesn t happen anywhere anymore, things are a lot better.  And, in fact, is there anybody in this room who either has daughters or who knows of daughters of friends who are not encouraged because of their science and math ability?  I actually that, if anything, it s an overabundance of adults urging on their daughters and female friends into math and science, which caused my own daughter to major in Italian Renaissance history and lit, I think she took it as a statement of independence from the kinds of pressure she was under.  Things are better.  We can argue about how much remains to be done, but has the environment been improving?  Yes it has, and that s good news.   Nonetheless, it all does matter a lot, and I think it s going to matter a lot more in the years to come for reasons that hinge on my own expectation of how this scientific story is going to unfold.  Now, here, if you think I m wrong in my expectation, you don t have to worry about anything else I say.  I will tell you what that expectation is.  I think that we are within a relatively few years of knowing for a scientific fact that there are different propensities and cognitive profiles between men and women, which mean that men will always be over-represented in math and science, relative to women, for reasons that have nothing to do with social arrangements or political institutions, and have to do with biological entities.   That doesn t mean that it s going to be the entire explanation.  I m with a bunch of other people in saying, we don t know what the mix is going to be, but the debate right now is not between those who say it s all nurture or it s all nature.  The debate is between those who say it must all be nurture, it cannot be nature, and to admit any admixture of biology as an explanation for these differences is inadmissible.  I don t think that that can be sustained.  A couple of reasons for saying that, the first is that until now we ve been restricted to the phenotype, whether it s Elizabeth Spelke or whether it s Josh Aronson or whether it s Charles Murray, we deal with experiments, we deal with survey data, and the rest where we can look at the visible outcroppings of something that s going on inside the mind, and we try to infer and deduce what s in there.  And we have gotten pretty good at doing that, but the black box has remained a black box, and we can only infer and deduce by working with the phenotype.  What we have learned at least in my bailiwick, which is survey research subjected to multi-varied analysis in which you have a bunch of independent variables that you use to control for various things, what we have learned is a little dispiriting for those who want to say it s only nurture.  First place, it soon became apparent, after multiple regression analysis reached maturity in the 1970s, that whereas such things as parental income and level of education and other socioeconomic variables do explain some things, they leave a whole lot unexplained.  And, as the models have gotten more sophisticated, it still remained unexplained.  And, furthermore, as you then took on a variety of other very sophisticated research techniques, enabled you to ask, well, could this just be residual sexism and racism, which pervades, and therefore explains the remaining differences, those patterns have not looked promising.   Furthermore, as we learned about the environment, we have learned that what should be the important things in the environment aren t nearly as important as we think they are.  The usual suspects for explaining, let s say, education attainment would be such things as schools, how much do your parents make, what kind of neighborhood you live in, these are, if you re comparing siblings, these are what are known as the shared environment.  And, in recent years, as people have disentangled, mostly using sibling studies, the genetic contribution and the environmental contribution, it turns out that what is called the non-shared environment dominates.  Now, there are some things in the non-shared environment, like peer groups, which are easy to understand, but, it turns out that a lot of that non-shared environment is what Professor Baron-Cohen was talking about, events in the womb.  And that kind of thing is simply not susceptible to systematic social or political interventions.  So you have a situation where we know there s a large genetic component to such things as IQ and so forth, but now we also know that an awful lot of the environmental influence is not the kind of thing we know how to manipulate.   Then you add my second reason for thinking that coming down the pike is news that a lot of us don t want to hear, and we heard some of that today with Richard Haier and with Professor Baron-Cohen.  It s as if we had a jar of marbles and we couldn t see inside the jar, all we can do is reach in and pull them out.  Every time we pull out a marble from the increasing understanding of the neuroscience and the biology of qualities such as ability in math and science, it s a marble.  It s a purple marble, if you think of the purple marbles as the ones which are kind of pointing toward a biological explanation.  The number of white marbles, which would point toward environmental explanations, are much fewer in number.  Today we kind of arranged it so we have an equal number of white and purple marbles represented in the panels, but I suggest to you that if you follow the literature, which is coming out virtually weekly, adding incrementally to our knowledge, none of these is an absolutely dispositive piece if information which will change our minds forever, they are lots of cumulative little pieces.  So when I hear that testosterone levels in the amniotic fluid can have the kinds of relationships with outcomes that have been documented, that s the kind of stuff which is a whole lot harder to argue with than literature on stereotype threat.  I am suggesting that we are only a few years away from the point at which to say men and women have absolutely equal biological propensities and abilities regarding science and math will be as hard to sustain as the proposition that the Earth rides on the back of a turtle.  That it is going to be such that this, that the conversation cannot be held among educated people, in the same sense that after a certain point the heliocentric solar system was the way everybody knew the world worked, even if the Catholic church didn t like it.  Now, if I m wrong, fine.  If I m right, and we are within a few years of that kind of knowledge, why is it important?  Well, the bad news here is, it s not going to be just regarding men and women.  We are going to find, as we unravel the entire genetic story, that, on average, the statistical distributions of important characteristics are different between economists and English professors, as well as between mongoloids and cockazoids, as well as between poor and rich, as well as between Dutch and Italians.  There will be lots and lots of ways in which group differences exist.  They exist because of biology, and you can t get around it.  The good news is that there is no real reason to be afraid of that knowledge, no real reason.  We will always be talking about differences and means, we will always be talking about differences in which there are so many individual exceptions they don t need to worry about them.  We will also be talking about differences that cut in lots of different ways, and everybody will be able to weigh the ways in which they have an advantage so that they are the most important.  And Dutch will continue to be quietly glad that they are not Italian, women will continue to be quietly glad they are not men, and vice versa.  Everybody will be able, in a reasonable world, to say, I m glad I m a member of Group X because the ways in which Group X does just great, happen to be the most important ways.  But, it won t work out that way.  It should.  It won t.  And it won t, and here I have to introduce the ugly head of politics into a session which has been blessedly free of them, well, Josh s discussion of George Bush and public speaking was, got a little in the margin there, but, here s a statement that I think is not an ideological statement, but a statement of fact.  Around the 1960 s, for a variety of reasons, within academia in general, and I have to say on the left in general, it became inadmissible that there d be biological differences among groups, whether men or women, or Blacks and Whites, or poor or rich, or whatever.  And it remains that way today.  So when there is scientific evidence beyond dispute that some element of the differences in outcomes between these groups is owed to biology, some element, people are not going to easily be able to say, oh well, some of it s due to biology, so what, and go on their merry way.  A major prop will have been ripped out from under a very large ideological base.  For example, consider the classic triad of race, class and gender.  You have lots of literature courses being taught in universities around America that analyze this stuff in terms of race, class and gender.  Why are you doing that?  Because you re going to be able to point out all the cultural and political and social ways in which gender is a social construction.  Now, gender will still partly be a social construction, but insofar as important attributes of gender will turn out to be biological, it s a lot harder to continue the kind of energy, intellectual energy, that has gone in to a great deal of the intellectual conventional wisdom within academia.  Furthermore, I am scared of over-reaction.  I am not a psychologist, and I don t know if I m using the phrase cognitive dissidence precisely right.  I will stand corrected if any of the panelists want to do it, but one aspect of cognitive dissidence occurs when what you think is increasingly divorced from what you say, or that your mind is changing and you keep on saying the old things you used to say, but you aren t thinking quite the same way.  Well, there s a tension buildup there that gets released, and in the area in which I have done my own most controversial work, IQ, it is especially dangerous because, in my view, there is no group of people in the world who think IQ is more important than the same academicians who say IQ doesn t exist.  That when they get up in the morning and they shave or put on makeup they are worried about how smart they are relative to their colleagues.  And, I have found, there s no way to say this not bluntly, okay, when I ve talked about IQ, when Dick Ernstein and I talked about IQ in the bell curve, we never used the word inferior versus superior, because, to us, that was an inappropriate way of describing differences in IQ.  There are differences, lower and higher, there are lots of things that are lower and higher and to say inferior puts a whole umbrella of connotations on it that we thought were inappropriate.  My impression is that a lot of intellectuals really do think that people who have lower IQs are inferior to people with higher IQs.  I m afraid of that underlying disposition being let out of the bag.  Now, I m asking the question, what happens when such a central tenant of an overarching political view is ripped out from under people?  And I ask it as a question because I don t know the answer.  I ve tried to give you a few of the reasons that I m scared of that answer, but it ll play out as it plays out.  May I close by suggesting that there is a solution to this that can unite us all, that I think could apply to every person in this room, of whatever political opinions, and that is, whatever else we believe, we all believe that whatever group differences do or do not exist is extremely important to take people as they are, as they come to us as individuals.  So if you are a woman, or if you are a Black woman, if you want to multiply the, that s irrelevant to whether, if I m hiring a violinist you know how to play the violin.  If I have a daughter, it is irrelevant how many Nobel Prizes are won by women, what is important is that my daughter have every opportunity to develop the talent she has, and if that happens to be in science and math, that she is given a full opportunity to do that.  I am saying that we need to return left and right alike to an open, active, passionate embrace of individualism traditionally defined, and that we can continue to argue about whether such and such a social program is a good idea or a bad idea, we can continue to argue about how close we are toward the ideal we could achieve, but we will not argue about is the importance of treating people as individuals.  That traditional American ideal has gotten lost in the wash in the last few decades as we have focused on outcomes for groups.  There is no reason, theoretically, that it cannot return as a universal American ideal.  To do so does not require us to change our political stance, it requires us to start saying out loud what I think many of us will once we get used to the idea, enjoy saying once again, we deal with people not as members of groups, but as individuals.  That is about as American as it gets.  Thank you.  [Applause].  Now we can have, yes, I ll start with Professor Barnett.  Rosalind Chait Barnett:  Hi, thank you.  I have two points.  One is, actually it s not just you that did this, but other people as well, when you talk about the brain, you talk about it as a static entity without any awareness of the fact that if anything we have a constant interaction between the brain and the environment that whatever the differences are in brains, if there are differences, that we want to talk about that, the amount of stimulation, the amount of interaction, the amount of experience the young child, boy or girl, gets, shapes that brain, which, in turn, is open to new kinds of experiences.  It s entirely interactive.  It s not brain, or, nurturer, or in each that ways, it s a combination of both, and I think the way it s being talked about it just doesn t jive either with the current literature of anyone s understanding of how people develop.  Secondly, your first point about discrimination being less than it was, it s definitely less overt than it was.  I don t know if you saw my show about the Medical Research Council data in Sweden.  There s certainly a covert discrimination that s palpable and measurable, and it affects women dramatically such that the opportunities to advance are curtailed, and they re not judged on their own competence, clearly.  And those data show it in a very dramatic way.  Charles Murray:  Josh, do you have anything you want to say, because this is sort of a summing up, and I m perfectly happy, I m sorry, I wasn t, I didn t deliberately ignore you.  This is a summing up, so this doesn t have to be questions of me, this can be a presentation of an alternative vision.  Elizabeth Spelke :  I want to heartily agree with two of the things that you said, and, then say some things that I hope might make us feel better about the possible implications of those things.  The first is that the mind is not a blank slate, we are not entirely products of culture, that we are endowed with genetically determined systems that then develop in predictable ways in the environment to give us the abilities that we have.  I don t think that this is something that we need to be afraid of, or worried about, because of two things.  First of all, you seem to be assuming that if the mind is a product both of biology, genetic biology and culture, then the way the world is now is the way the world will turn out to have been biologically ordained to be.  Right?  It will follow that the reason that there are so many more men in science than women in science is that our genes have determined it to be true.  But I think if we look across cultures, if we look across historical times, we ll see that despite the fact that it s always been true, that humans are creatures both of biology and of culture, there s enormous flexibility and enormous variability over time in how those biological capacities express themselves.  The fact that our intelligence is biologically based does not imply that gender differences in intelligence are biologically based in a way that will produce the gender patterns that we see.  That remains to be seen.  The second thing, you spoke as if the evidence for biological determinism meant we have to throw up our hands and give up, that if we learn that there is some gene that s responsible for the development of some quality or other, it follows that we can t influence the development of that quality.  In fact, I think if you look at medical research, you see exactly the opposite, that the more we learn about the biological mechanisms of disease, the more empowered we become to develop new treatments for disease when we learn that there s a genetic predisposition for certain women to get breast cancer.  We don t take this as terrible news, we have to throw up our hands, we ll never be able to cure breast cancer.  Instead, we use it as a way of targeting, okay, what can this tell us about the mechanisms of the development, about strategies for prevention, and so forth.  My guess is that the same will be true of research that s now investigating the genetic and biological foundations of intelligence.  It s going to empower us to achieve social goals as we choose to achieve them better and more effectively, as we better understand both our biological nature and the way it plays out in culture.  And then, finally, the last point that you made, treating everyone as individuals, I could not agree more, and that is the reason that I would recommend that you and everybody here read, seriously, the much maligned NAS report on bias and barriers.  What that report attempts to show, in example after example, is that despite the best efforts of many people to treat all of us as individuals, we don t fully succeed, we end up establishing different criteria for people, depending on the groups that they re in, and this can be done even by people like E.G. Boring who professed to evaluate each person only individual by individual, independently of their group membership.  We discover, in case after case, that this ideal has not been completely reached.  And it seems to me entirely in the spirit of American enterprise that when we discover evidence like this we want to talk about ways to overcome it.  We want to talk about ways to get beyond the barriers, to get beyond bias conscious, or unconscious, so that we can actually achieve the ideal that you laid out, that, I take to have been the goal of the NAS report.  Christina Hoff Sommers:  Can I just say one thing, and then you can answer, if you would.  You made the comparison to breast cancer, which, if find out there are biological inclination, if we don t take, as you said, well, we have to cure it, but, see, I don t think more women going into psychology, sociology, teaching and medicine, and more men into engineering and chemistry is in breast cancer.  So, we don t have a pathology in need of a cure.  I feel that it s simply an unassumed assumption that, by, and I m not going to keep mentioning this report, but, by some studies, it s just an unassumed APR assumption that difference means discrimination.  I think there are some differences that are benign and interesting and just an expression of people s individual choices.  And, yes, they happen to be from groups, so they, you know, when you add them up, you re going to get some transcendences, but, again, not a pathology in need of a cure.  Elizabeth Spelke:  I agree, and I didn t mean the breast cancer analogy in that way, I meant it as an example of how understanding more about the biological development of a phenomenon does not render us more powerless in the face of it, but, the opposite.  I agree with what you re saying about individual differences in interests, and, I don t think that we would want to take as a goal, and the NAS report did not take as a goal, that we want a quota system where we have exactly equal numbers of every group.  We do want a situation where barriers to entry into a group unrelated to ability and tied to preconceptions about the proper practitioners are, as you said.  Charles Murray:  Do any of the other panelists want to do some summing up?  Amy Wax:  Yes, I am going to be a little bit of a skunk at the party here, because I ve already taken on that role, but, I do agree that just because there may be gender differences that are fairly intransigent and we really don t know, I think at this point we have to be agnostic about it.  Although I think there is some evidence for it, that doesn t mean we can t work in a progressive way to try and reduce the conventional barriers to women s success.  I mean, I don t see any inconsistency there at all, but I think that sometimes the best is the enemy of the good.  I think there is this overwhelming temptation to indulge in these utopian assertions of faith, sort of dogmatic assertions, about how there are no cognitive differences, there are no temperamental differences, and I think it s a little bit of sugarcoating to say, well, we can all accept the differences.  I think what happened to Larry Summers stands as a cautionary tale, a rather frightening one to me, of how certain assertions violate the orthodoxy and will not be tolerated.  So I think there are a lot of mixed signals out there about what one can and cannot say.  To Charles Murray, I would say, just very briefly, I think you re sugarcoating, a little bit, on the implications of IQ differences by group.  The fact is that intelligence really does matter a lot in our competitive and technological society, and good things come to people who are smart.  If there are differences in intelligence, then, I think politically we are going to have to cope and come to terms with different outcomes for individuals who vary by intelligence.  Charles Murray:  Let me try to be very explicit, because I don t want anybody to misinterpret.  I said we need not fear the coming scientific knowledge, but I think we will.  I think we will, in fact.  I think there is a kind of passion out there, I will move away from gender differences to a specific example, which is etched very clearly in my mind, many of you who have not read The Bell Curve I bet, think that the main thesis of the book was that Blacks are genetically inferior to Whites with regard to intelligence.  Do you know what our actual sentence was about the sources of the difference in test scores between Blacks and Whites?  As I recall, it was, so, we think that either those who think it is all nurture or all nature have failed to make their case, and that it is some mix of both.  As to what the mix might be, we are resolutely agnostic, we do not think the data yet permit an estimate.  That is the sentence that got us in so much trouble.  We said, simply, biology is probably some unknown portion of this.  The degree of paranoia about biological sources of group differences cannot be overestimated, as Larry Summers found out, and, as I found out.  But, when it comes to the business of large group differences, whether they are male/female propensities for science and math, or whether they are racial differences in test scores, one can both agree that these are important, as Amy Wax just indicated in terms of all sorts of social and economic outcomes in life, without them impinging on our assessment of the human worth and dignity of people around us.  I have a thought experiment that I will now try out on all of you.  I will not ask you to raise your hands with your answer, but here comes the thought experiment.  I want you to seriously think about it.  First, I want you to imagine someone who you know is not nearly as smart as you are.  Let s say they are 30 IQ points lower than you are.  Do you, despite your best of intentions, tend to feel a little condescending toward that person, feel a little sorry for them?  That s the first half of the thought experiment.  The second half of the thought experiment, think of somebody who is a lot smarter in a pure IQ sense than you are, 30 points higher.  Do you feel inferior to that person?  Now, if you do not feel inferior to the person who is smarter than you by 30 points, I was at Harvard and I knew lots of people 30 points higher than I was in IQ, I did not feel inferior to them.  They could do things I couldn t do, intellectually.  So what!  A lot of them were really, kind of goofy in other senses.  Well, if you, in that thought experiment, did not feel inferior to the person 30 points higher than you, what makes you think the person 30 points lower than you thinks that you re so great?  What makes you think that they feel inferior to you?  And the fact is, they don t, in most of the cases I m aware of, that they are dealing with life as it comes to them, they have a pretty realistic sense of the role that this raw thing called intelligence is, and we ve got to stop worrying about these differences as having anything to do, whether we re talking about people we will cherish and like and respect or anything else.  So, Amy, I will disagree with you to that extent.  I will agree with all the social and political implications of differences in various kinds of abilities, but, on a human level, I think that we, intellectuals, are way over-absorbed with the importance of IQ.  Josh? Joshua Aronson:  Let s suppose that we do, Richard Haier develops the technique and we can get an IQ readout after a brain scan.  What policy implication would that lead us to?  Charles Murray:  Well, you re talking to a guy who just published an article saying, we ought to get rid of the SAT, because, it turns out that you can predict college performance just as well using achievement tests and high school grade point averages.  You can if you also throw in the SAT.  And, in fact, the real advantage in that is that there is one particular class of young person who would be advantaged by Richard s technique who I most want to not advantage, and that is the really smart kid who coasts through high school not having to work very hard because high school isn t very hard, then aces the SAT, and gets into a selective college, and they haven t done a lick of work and haven t tried.  I d much rather give that slot to somebody not as smart who has worked hard.  And Richard s, it s okay for Richard to give his test as long as we never tell the scores to the student or to the Admission s Committee, but we just use them for pure research purposes, then it s okay, Richard.  You have the microphone.  Richard Haier:  So many things to say.  Today is my first day of retirement from the university.  I must say, I knew it was time to retire when I found myself agreeing with you on so many things.  [Laughter].  Charles Murray:  Thank you?  Richard Haier:  Twenty years ago, when I started intelligence research, it was very clear that this was a career killer, nobody was doing intelligence research in the United States 20 years ago.  Arthur Jensen basically killed the field, inadvertently.  I am very happy to see now that intelligence research has really come mainstream.  I think the remarks you have made about neuroscience and biology are probably correct.  I would propose my own thought experiment.  I know I ve mentioned this to you before.  I m just wondering, by a show of hands, how many of you would buy a pill that would increase your IQ by 30 points?  What if it only cost $5.00?  [Laughter].  And, if you had a choice between a pill that would increase your charisma or the IQ pill, but you could only have one, how many would choose charisma?  Well, I find this fascinating, because I would give anything to have another 30 IQ points, especially when I come to a meeting like this.  But, I think, the last thing I want to say is that, I think you are absolutely correct in observing that the last basket of real resistance to the idea that biology matters will be in academia.  I m very fortunate to do an awful lot of public speaking, and the public is way ahead on this.  I get no grief at all when I talk about these things to the public.  I get a lot of grief when I talk to academic audiences, and, I m not sure what it s going to take to change that.  Charles Murray:  Should I call on people, or, do you want to call on people?  Christina Hoff Sommers:  Why don t you call on people.  Charles Murray:  Okay, start there.  Female Audience Member:  I agree, and I find the data very interesting on biological differences.  What I disagree with is that they justify extreme occupational segregation and exclusion of women from the workforce in some countries, or, restriction of women to low-paying jobs in this country and others.  And that s the problem with believing in biological differences, is the reductionist approach, and I think, following Amy Wax s point, I think the contribution of the differences that have been articulated in these panels are trivial, are trivial in explaining occupational interest, choice and performance, professional performance.  I think you will find in a lot of professions the array of people and the mix of people, orientation, the systems orientation, is mixed.  In fact, if you say that the people skills are really important, then women should be the CEOs of most corporations.  There s no explanation for the range of skills that people need, and how these biological differences really contribute hugely to restricting certain people to certain kind of positions in our society.  Charles Murray:  I guess I do want to introduce a little data, because I was listening earlier today, and had to restrain myself.  I did a book called, Human Accomplishment, which was identifying sort of the real outliers, you know, the people who had accomplished at extremely high levels, by historic standards.  There is a very large literature on the characteristics of those people, and, ability is one of them.  Being pretty smart is important, in fact, it s a baseline condition.  But one of the interesting things that s been found that distinguishes the person who has a high IQ but just has a pedestrian career, and the person who really, really achieves, is, extraordinarily high levels of work and time.  Furthermore, it s the old 70-hour thing we were talking about earlier today, and arguing about, and, furthermore, it is also, still remains true, I m sorry to say, that peak accomplishment occurs around 40.  This has been remarkably consistent across time and across fields.  Well, if you re explaining the extraordinary disparity between men and women at a lot of these very high level jobs, whether it s CEO or whether it s Nobel Prizes or whatever else, as long as women have babies in their 20s and 30s, you are always going to be taking a very large chunk of the female half of the population and sort of put in a huge boulder in the way of accomplishing at these extreme levels.  Now, this would be true, even if it were purely the physical act of giving birth, that takes you out of action for a while.  But one of the other things that I don t think there s a lot of dispute over anymore is that motherhood is a profoundly different experience for women than fatherhood is for men, that men can be loving fathers and work for 12-hours during the course of a day, and their kids state of mind or health or well-being has not crossed their minds once.  I think that I am on safe scientific grounds in saying that any study of the ability of mothers to do that would find that a lot fewer are able to do it, that the ability to be monomaniacal during these years in which you must accumulate enormous expertise and do it by about the age 40, that women are at an adherent, inerasable disadvantage now and forever more.  And, so, you re always going to see those disparities because being CEO is not just skills, it is this accumulation of expertise through effort that does require monomania, in most cases.  Female Audience Member:  But you just changed your argument.  You are now talking about life choices and innate biological differences in ability.  Charles Murray:  Yes I am, no, I m saying that for women it, I will argue that the science will, in a few years also confirm, that mothers think a lot more about their kids everyday all the time than guys do.  I will state that with confidence as a scientific finding coming down the pike.  Christina Hoff Sommers:  Someone new in the back.  Someone we haven t heard from.  Female Audience Member:  I just have a comment.  I don t, I probably have fewer credentials than anybody, scientifically, in this room.  I just got my bachelor s degree in biochemistry and pharmacology from McGill University.  But, what I wanted to say was that my four years at McGill went without a hitch.  I didn t rumple any feathers.  Nobody turned their heads, you know, look at this woman, she is in science.  Nobody cared.  Most of my professors didn t look up from their podium long enough to realize that I was a woman, much less care.  All I wanted to say from that, was, maybe it s my young, na�ve self, but, because of women like you, you had to freeze to death for a whole semester, and, you know, the research of Dr. Aronson, and pretty much everybody here, that, I was able to go through so easily.  I mean, I did work my butt off for four years, but, relatively easily, but, at the same time, it s because of people like Dr. Wax that, when I would fail an exam, I would say, I wouldn t say, gosh, I wish I had a bigger brain, I wish my brain was more predisposed to science, it was, you know, I really should have studied a hell of a lot harder than I did.  It was my fault.  I m not trying to be a cheerleader for science here, not saying that women have to be cheerleaders, but, I m just saying that I feel like I can do anything and it s because of discussions like this that that happens.  That doesn t really have anything to do, and I don t have any data, but I just wanted to say that it s days like this that make me realize that my four years wasn t for nothing and I have the whole rest of my life to be a scientist, and I think that that s really great, and whatever your view is I think this discussion needs to go on for a long time.  That s really what I wanted to say.  [Applause].  Christina Hoff Sommers:  Yes, you ve had your hand up for so long.  Zakya Kafafi:  I have a couple of comments here.  Let me introduce myself.  I m Zakya Kafafi, I m presently at the Naval Research Lab, and I m moving very soon to the National Science Foundation.  I ll be the Director of the Division of Materials Research.  One comment I have is regarding fear of failure.  At the age of 18, I fell in love with my next door neighbor, that was in Cairo, Egypt.  And, as a result, I was in high school then, I was 17 years old.  I had very poor grades.  I was usually a very good student, but I made very poor grades because I was distracted with this man I was in love with.  Soon after that, I moved to the U.S. and faced with my very first failure in life, I was determined that I m not going to let marriage distract me from my studies.  In six years, I had two children and a PhD in Chemistry from Rice University.  So, this is just a comment here.  The other thing I want to make a point of, is, that, I think it s very important to make the society aware of discrimination, whatever and whenever it happens, and the impact of that awareness, the effect of the awareness on changes in the society, and one of the examples that comes to my mind, the MIT report by Nancy Hopkins.  Look at how many women presidents were there before the Nancy Hopkins report, and look now how many excellent women presidents we have in many Ivy league schools, and many other excellent schools.  Female Audience Member:  Here comes the engineer with numbers.  Notwithstanding all these wonderful programs that they have now for girls, and, I agree with you, there are very few school systems, especially White middle-class school systems, but also inner-city schools that are starting to appease to children at a young age, appeal to children at a young age, think about math, think about science, science is fun.  The number of women going into engineering and staying in engineering peaked in the year 2000.  We ve been heading south ever since then.  Is this a question of biology or choice or locking the clubhouse door and saying, you are not invited in?  In the year 2000 a dear friend of mine sent her daughter, third generation, to go to MIT.  This girl wanted to study nuclear physics.  She wanted to be a nuclear physicist.  Her father and grandfather are both members of the National Academy of Science.  She went to a single gender school from the time she was in pre-K through 12, and excelled in everything she did, really excelled.  She came home at the holiday break crying, and saying, I want to leave MIT.  Her mother said, why?  The answer was, there were four girls in the physics major, and, we are, life is miserable.  The father, of course, who was an MIT grad, got on the phone, called the Chair of Physics and said, what s going on, my daughter has never cried, she s wanted to be a nuclear physicist since she was a little girl?  The answer that was given by the Chair was, women do not do well in physics here at MIT, I suggest she change her major.  That happened in 2000.  I m happy to report that she moved over to electrical engineering and has done very well.  But, I hear these stories all the time, this was just the one that popped into my mind right now.  This isn t a small collection of stories that I have.  So, please understand that there s this other component, and it s called denied access.  Charles Murray:  Well, I know nothing about that specific case.  I do know that at MIT there are lots of men who have always excelled all their lives and they go to MIT and they want to go into nuclear physics, and, guess what?  They have to transfer to electrical engineering because they are good enough for electrical engineering, and they aren t good enough for nuclear physics.  That, in fact, you are looking at somebody who can be real smart and real able and not succeed in nuclear physics at MIT for reasons having nothing to do with being female.  Neither one of us, neither one of us have the basis for knowing whether that was the case at MIT.  I m all in favor of equal opportunity and aggressive equal opportunity, but to reflexively say that the kind of story you just gave is evidence of discrimination and denied access just goes way beyond the data.  Christina Hoff Sommers:  In the back, yes.  Female Audience Member:  Just a quick anecdote.  My mother was a mathematician in the 1940 s, working in computer science, and her comment was, you know, they were just so glad to have anybody who could do the work that there was much less discrimination than there would be otherwise, although there was some.  Now, the overall point to that is just to make, there is a demand for people that can do math and science.  It is an increasing demand.  There is a shortage of such people in the American economy.  Any woman who feels discriminated against in getting a professorship at MIT, can always get a job at SYSCO, or someplace, there are zillions of these jobs available for anyone that has any inclination to do the work.  So, my concern, since we all know that a competitive economy tends to erode any sort of irrational discrimination among people who are competent, my concern is more that we get into one of these awful affirmative action reverse discrimination kind of situations because I would not want inflicted on the math and science regime, which is sort of a jewel of achievement in this country, the sense by White men or Asian men that they need not apply for a particular job or a particular position, or whatever, because that s going to be marked for a woman, and these fellowships are marked for women, and all kinds of this crazy stuff, which turns our entire idea of meritocracy on its head.  Christina Hoff Sommers:  I will say that s already happening.  The National Science Foundation has a program called Advance, and over the past five or six years, schools have been given immense sums of money, $3.5 million, University of Colorado, $3.7 to New Mexico State for institutional transformation.  If you read between the lines, it looks as though it is a combination of mandatory workshops on gender, consciousness raising, as well as, it looks like quotas, but they use other words.  It may not be, it may be something else, but I do hope there are some good journalists and science writers here who will go on these websites and look at these advance programs and see what s going on, because I just feel that it all moved ahead without the benefit of serious discussion by scholars [cross-talking].  Charles Murray:  We re at 4:00, why don t I take one more.  Let s take one more and close the show.  Dr. Jessie DeAro :  Hi, my name is Dr. Jessie DeAro, and I am with the Advanced Program at the National Science Foundation.  It isn t the description that you gave.  The advanced programs are all about looking at the issues from a different perspective.  It s looking at the culture and climate of science and engineering research, and whether or not that can be modified in a way to be more inclusive of all people in society.  It doesn t mean that we re going to look at lower quality people or less high-performing individuals to do the science and engineering research.  It means that science and engineering research benefits from diversity of input from different perspectives and different experiences, and that s recognized.  The future of science is in incremental progress at the interdisciplinary boundaries between the, instead of within the disciplines.  You need people that are looking at things from different perspectives involved.  And, so, if you look at the organization of universities and colleges, can you make changes to that so that women can be as successful in that organization as their male counterparts?  It has nothing to do with lowering the standards of quality at all.  It s about looking at the organization and whether there are boundaries, or, I m sorry, barriers, in the organization that are keeping people from fully participating.  Christina Hoff Sommers:  I d respectfully disagree, having looked at the descriptions, but, let s leave that to those who want to investigate, go on the websites, look at the Advance Program and the details of how it s playing out at various universities.  I think we re going to see a politization of the sciences such as we ve never seen before in the next few years, and we ll have the culture war all over again in a different place.  Charles Murray:  You re being more apocalyptic than I am.  Christina Hoff Sommers:  No, it just looks, because the NSF funding, government meddling in something that may have been best left to be, you know, fine, to do something about it locally, but it just seems like a massive government intervention in a sort of delicate ecology, throwing millions and millions of dollars at single institutions.  I hope I m wrong.  Charles Murray:  Could I Just say in closing, because we are at the closing time, what I said at the beginning, we have engaged, over the course of today, in one of the most dicey topics there is, and people have occasionally gotten passionate about it, as is perfectly appropriate, given that there is passion out there on this, but we have also dealt with it, I think, with remarkable success as an issue to be thought about and discussed and to which data should be brought to bare, and an issue in which people of goodwill can disagree.  I think that is a major accomplishment, and, in closing, thank you all.  [Applause].  [END OF DOWNLOAD]    </body></html>