Gender and Science:

A reader kindly pointed me to a couple of interesting items on a matter related to something we touched on a few weeks ago: whether there are material biological cognitive factors that lead men and women to be disproportionately represented in certain fields. (Note that the question isn't whether these are the only factors, or where any particular woman or man falls within the distribution of certain cognitive skills.) The posts on this blog were about law, where matters may be quite different from science. Still, the debate about gender and science is interesting in its own right, and likely overlaps in some measure (though of course not entirely, given that much of the scientific debate is about mathematical aptitude, something that lawyers of both genders are infamous for not having) with similar debates about law.

The reader's two sources were:

1. has a debate between Harvard psychology professors Steven Pinker and Elizabeth Spelke.

2. Nature has an article by Stanford neurobiology Prof. Ben Barres, taking the view (as best I can tell) that there are no such material biological differences, and that the entirety of the disparity between men and women in the sciences comes from societal factors. (This is not completely clear, since his introduction frames the issue as being between the "Larry Summers Hypothesis" "that women are not advancing because of innate inability rather than because of bias or other factors," and the rival hypothesis that "women are not advancing because of discrimination." But since Larry Summers' point was that part of the reason for the disparate representation of women is biological cognitive differences, I take Barres' rejection of the "Larry Summers Hypothesis," and specific criticism of Summers, to mean that he's saying that biological differences are no part or perhaps next to no part of the matter.)

I'll have a bit to say shortly about details in the Barres source, but for now I just wanted to post the links.

The Original TS (mail):
First, the idea that there are no biological factors whatsoever that impact the relative success of men and women in the sciences is a pretty dubious hypothesis on its face. Men and women are different and many of those differnces are innate. Perhaps none of those differences have any effect when it comes to science but that is only one possible outcome out of many. It's somewhat like flipping a coin a thousand times. You might get exactly 500 heads, but it is substantially more likely that you'll get either between 490 and 499 heads or between 501 and 510 heads. In other words, that none of the innate differences between men and women have any impact on their ability to do science is only one possible outcome out of many.

Second, there are plenty of documented differences that could easily affect the relative success of men and women in the sciences. Some are statistical, like the outlier effect -- The distribution of intelligence for men has a higher standard deviation. Men as a group are, therefore, both smarter and stupider than women as a group.

This is not to say that there are not social differences that need to be addressed. There most definitely are. But to reject the idea a priori that there may be innate differences between men and women that affect their relative success at the highest levels of scientific research is, well, bad science.
7.25.2006 2:11pm
GMUSL Rising 3L (mail):
GNXP justifiably and completely trashes Barres' mishmash of half-truths here.

The smoking gun is the comment about how Barres (a female-to-male transsexual) has noted that "his" spatial abilities are noticably better as a result of the hormonal changes. So clearly, a life-long exposure to higher testosterone cannot have any effect.
7.25.2006 2:20pm
Kingsley Browne (mail):
Barres lost all credibility with me in his description of the "convincing" data in the MIT Discrimination Study. That study was about as "data free" as a study can be, and it was presided over by Nancy Hopkins, the person whose complaint of discrimination precipitated the study in the first place. (Isn't there something about no woman being a judge in her own cause?) For a demolition of the MIT study, see Judith Kleinfeld's "MIT Tarnishes Its Reputation with Gender Junk Science" (

Barres seems quite ready to accept Nancy Hopkins' assertion that she was discriminated against, but then questions the motives of the many women who deny that discrimination is a problem in science, suggesting that they may just be pulling the ladder up behind themselves, "perversely believing that if other women are less successful, then one's own success seems even greater."

As for his assertion that no significant differences in mathematical ability exist between the sexes, he looks at average scores in broad samples and does not deal with the issue of the greater male variability that results in a disproportion of men at the very top, as was discussed in some earlier threads.

I have an article in the Cardozo Women's Law Journal on the Summers episode, including discussion of some of the relevant data. It is available at:
7.25.2006 2:43pm
John Armstrong (mail):
The question is, "Are there any gender influences on mathematical (or scientific) aptitude?" The influence could be as little as a higher standard deviation, as in the case of running speed between black and white athletes.

The problem with this whole discussion is that one side is of the opinion that the question shouldn't even be asked. The crowd that called for Summers' head assert that to even entertain the notion that there might be non-sociological factors is the first step down the slippery slope to Nazi gas chambers. Please note that I am not magnifying their arguments -- this exact argument has been put to me.
7.25.2006 2:43pm
statfan (mail):
The distribution of intelligence for men has a higher standard deviation.

This is surely explainable by purely social factors. Imagine that women fill in IQ tests by putting down the answer that they think is likeliest to be correct. When they don't know, they leave the question blank. They do this because they were raised to be risk-averse. Imagine further that men answer completely at random in all cases (this is an exaggeration -- but surely someone less risk-averse would guess more often). Now, if you look at the outcomes of IQ tests, you'll get the result you cite (higher standard deviation on IQ tests) -- but not for the reason you cite (higher intelligence).
7.25.2006 2:44pm
anonyomousss (mail):
john - i think you have the question wrong. the real question is whether there's a biological difference in something called "scientific aptitude" that's large enough to explain all, or most, or almost most, of the enormous gender disparity in science. if there is such an effect but it's very small, nobody is going to care. the problem is that people will almost invariably from "suggesting" that there "might" be "an effect" to concluding, or implying, that (a) there is an effect and (b) it's so large that discrimination and other social causes are irrelevant, or such a small factor that it's not important. just look at the commenters on this and other posts on this issue on this very blog, approximately none of whom are familiar with the details of the research on this issue or the feminist critiques of it, but many of whom will happily assert just that. and that's what summers said, or implied so strongly that he might as well have said it.

and the idea that social factors aren't important is crazy. men dominate every field and subfield of science, just as they dominate every other aspect of our society, even though success in the various fields and subfields of science involves radically different skill sets.
7.25.2006 2:57pm
The Original TS (mail):
Nice try, statfan. But the deviation is also noticeable at the other end of the spectrum. Males are more likely to be mentally retarded than females, too.

I suppose you could come up with some social explanation for this as well, but Occam's razor will almost certainly decapitate your argument.
7.25.2006 3:14pm
John Armstrong (mail):
anonymousss: I'm not saying social influences are unimportant, or even that they're not the dominant factors. Still, when I advance the opinion that biology might have anything to do with the situation, I have been all but labeled a Nazi.

My point isn't to debate biology vs. sociology. I'm just pointing out that the anti-biology side is shot through with such rabid partisans that it is impossible to have a rational discussion. To them, even raising the mildest form of the question is tantamount to advocating the wholesale slaughter and enslavement of women.

As an aside, I've generally found that the academic group most likely to take this radical stance is men in the humanities. Women in the humanities disagree but discuss, while the response from both sexes in technical fields is generally, "I don't really know enough about the state of research to have more than a gut feeling, but sure it's worth asking."
7.25.2006 3:24pm
John Armstrong (mail):
The Original TS: Thanks for pointing that out. The effect the deviation has on both tails of the distribution has stuck me as one of the more egalitarian aspects of this model. On average the groups are the same, but one has more representation at the high and the low end than the other.

It seems plausible to me (see the last paragraph in my previous comment) that from the biological perspective men and women have equal mathematical ability on average, but the standard deviation for men is slightly higher. This means that if you only look at the top end of professional mathematics, you'll find more men. It also means that if you only look at the depths of innumeracy you'll also find more men, which might explain the gender imbalance in law and politics :D
7.25.2006 3:30pm
jimbino (mail):
Why is it that otherwise right-thinking folks buy into the religious fundamentalist's speaking of "gender"-distinctions when what we are talking about here are "sex"-distinctions? There are lots of genders of lots of things; there are two sexes in humans, and I'm not embarrassed to say so.
7.25.2006 4:31pm
Isn't it also the case that Summers has not said that he actually believes in the hypothesis, but rather only that the hypothesis is, as a purely academic matter, a legitimate one that is a valid line of academic inquiry?

I may be wrong, but if not, it seems that Summers gets short shrift when the distinction is not pointed out.
7.25.2006 4:47pm
orson23 (mail):
Precisely jimbino.

Honest discussion of controversial topics must begin with honest labeling - clearly something absent when an author begins lableing a well-known psychometric observation the "Larry Summers hypothesis."

These are well-known sex-linked differences. Simply generalizing from one persons experience of exposure to one, albeit major, steriodal hormone is equally silly. Nature, just as in topic like man-made global warming, is prostituting itself. This phenomenon seems to be a function of the socialization of J-schools (eg, the leading repoting texbook says reporters SHOULD take sides) and the way Big Science has turned prestige publications into "profit centers," rationalizing the hiring of elite J-school grads instead of scientists to edit them, thereby entrenching the politicization of science. But I digress.
7.25.2006 4:50pm
orson23 (mail):
numitor - you are correct.

Summers made his observation in order to further research into the varied and various actual cuases for observed social differences between the sexes. His purpose was to open avenues of discussion, debate, and research. By contrast, the PC-nazis purpose is to cast out heretics who dispute that all differences are social constructs. Now, tell me who is open-minded here? - and who is being ofensively dogmatic and anti-science? (Only one answer allowed: this is a categorical question.)
7.25.2006 4:57pm
The Original TS (mail):
Why is it that otherwise right-thinking folks buy into the religious fundamentalist's speaking of "gender"-distinctions when what we are talking about here are "sex"-distinctions?

Ah, yes. Nouns have gender. People have sex.

You are, of course, correct. But I've given up on that particular linguistic battle.
7.25.2006 5:17pm
John Armstrong (mail):
jimbino: you are, of course, correct. My choice of terminology within this comment thread was to mesh with the common use of terms in the ongoing debate, which is phrased (erroneously) in terms of "gender" distinctions.
7.25.2006 5:48pm
tomjedrz (mail):
Interesting .. I seem to recall a very long series of posts and comments relating to usage and word choice a few months ago. If I recall, the argumant boiled down to, essentially, that a word means what we agrees that it means, and meaning can change as usage changes. Consequently, gender can properly be applied to people, because enough speakers and writers are doing so that the meaning is clear.

In any case, there is another reason to use the term gender. The alternative frequently causes the documents to be flagged by the various protective filters on computers and networks.
7.25.2006 6:25pm
logicnazi (mail) (www):
The word gender is widely used to simply mean sex. It might be nice if the words clearly meant different things but clear usage facts tell us that you can use gender to mean sex. Besides, in some circumstances it is very usefull, e.g., in contexts where there could be confusion over wether you mean intercourse or type of sexual equitment.

Frankly, I would be happy to see the word gender reduced to a pure synonym of sex (when it isn't talking about nouns but context makes this clear). Yes it might be nice to have a technical word to refer to the socially constructed aspects of sex but the word gender has been abused too much to do this without deep confusion.

For instance take the claim that There are more than two genders. frequently used as a slogan by LGBT and other 'alternative' sexuality advocates. While I agree with the sentiment (don't torment people who don't fit into out social notions of male and female sexuality) the claim just doesn't make any sense.

If gender just is our social constructed sexual expectations the very lack of acceptance for LGBT folks is proof that there are only two genders. Perhaps we should have more than two genders but as it is now we clearly have only two (though being gay is approaching becoming a third). Besides, surely what these advocates want isn't just 5 stereotypes of sexuality that society recognizes. Really what they want is something like the dissolution of gender or at least toleration of those who don't have gender.

The notion of socially constructed sexual roles is a subtle one and shouldn't be referred to by a familiar short term like gender as it just leads to confusion. Besides our intuitions about the word 'gender' are all wrong for this role. Obviously there are people who fall into no socially recognized sexual role (not male or female) but it sounds really weird to say, "They have no gender" and wrongly implies they are somehow asexual.
7.28.2006 5:30am
logicnazi (mail) (www):
I have to say I was pretty disappointed with Speckle's arguments. Her justification to assume that the differences are societal are primarily based on course selection and performance in school. Yet this is just begging the question. The (some) innate difference explanation predicts exactly the same result because the greater conscientiousness of women compensates for any differences in interest/ability.

Moreover, this isn't simply a situation of two equally plausible competing justifications. If the SAT-M and similar tests that are flawed in such a way as to suppress women's scores despite actual equality AND women were more concientious we would expect greater enrollment and perfomance in math classes. Since we don't see that we must discard the hypothesis that women are more concientious but now we can no longer explain the greater college graduation and enrollment rates of women in general. Or to put the point differently if class performance and selection were really the gold standard we should conclude that women are in general smarter than men and the advantage is just less severe in mathematics but the supporters of the nurture position don't seem prepared to take this view.

Additionaly it seems unmotivated to take class performance/class choice as the gold standard yet throw out similar measures like the Putnam exam which is much closer (though still too 'tricksy' for my taste) to real mathematical reasoning. Pinker's theory explains the difference in female performance here (effort makes less difference in the Putnam) the nuture theory has a much harder time. It seems totally unmotivated to object to the Putnam on the same grounds the SAT-M was rejected. Besides, it is equally possible that class based measures are biased the other way.

Also, as Pinker points out, simply showing a lack of difference on some measures doesn't do anything to dispute a finding of difference on other measures.

More generally the nuture view should be given less credibility because it lacks any theory of how the effect works. Back in the 70s when the nurture view clearly spelled out that by treating your female children like your male children you could reverse the effect the nurture view was just as viable a theory but when the obvious nurture effects were disputed it has become an ad hoc theory.

The nature theory on the other hand has specific mechanisms claimed to produce the effects. By showing a lack of effect by pre-natal hormones or genes on the X chromosome one could cast serious doubt on this theory. Without a similar mechanism for the nuture theory that isn't jury-rigged to be untestable we should give it less credance.

If this is the best the extreme nurture people can do I've moved from simply finding their view unlikely but still plausible to implausible in view of the evidence.
7.28.2006 10:25am