The brunt of Schaffer's argument seems to be that since men and women have roughly similar outcomes for any given mental task, therefore we must be using roughly similar processes to achieve those outcomes. But there's no reason to believe that's true.

She needs to better define the question she's addressing. Is it whether or not sex can be a predictor of one's capability to perform a certain task (I'd agree that it isn't) or the larger question of whether or not men and women think differently, even if we are equally competent at the end of the day?

--s

Actually I think she's saying that the data don't support any meaningful claims of mental differences between the sexes. This is interesting because there are people pushing strong claims with shoddy analysis.

But what's a "mental difference?" To take an extreme hypothetical, if both men and women spoke 16,000 words per day at 30 words per minute, but the men spoke only of methods for playing football and the women spoke only about feelings regarding food, surely that would be some kind of "mental difference," wouldn't it?

It's like saying there's no difference between cars and bicycles, because some people get where they're going by car and other people get where they're going by bicycle, but both get where they're going. And in a very limited sense that's true.

--s

Ok, "mental difference" was a squishy little phrase to let me avoid a slightly harder discussion. Here's what I was getting at:

First, clearly there are physical and physiological differences. Denying that would be foolish.

Second, the measured neurological differences are small in terms of standard variations for both groups AND there is no clear line from the measurable differences to a specific outcome.

Finally, the measurable behavioural differences can not be well categorized thus far. That is, if your example were true, we would not know whether the difference was based on innate or learned factors given the data currently available.

blueshift:

Ok, "mental difference" was a squishy little phrase to let me avoid a slightly harder discussion. Here's what I was getting at:

First, clearly there are physical and physiological differences. Denying that would be foolish.

Second, the measured neurological differences are small in terms of standard variations for both groups AND there is no clear line from the measurable differences to a specific outcome.

Finally, the measurable behavioural differences can not be well categorized thus far. That is, if your example were true, we would not know whether the difference was based on innate or learned factors given the data currently available.

I'm not sure what you mean by measured neurological differences. Both humans and chimps can be trained to press a button when a red light comes on - that's one qualitative data point. But that doesn't mean there's no difference between humans and chimps.

So, the question is, is there ANY qualitative test such that men and women can be identified by their scores? I don't know, but the SAT certainly wouldn't qualify since if such a difference were to appear, they'd adjust the test to remove it. That is, it's a test explicitly designed to treat men and women equally.

Regarding the college student example, that's also pretty artificial - the college environment is going to influence the result, as college students tend to have similar daily experiences. So it's reasonable that they'd speak a similar amount. Once again it's an environment that's been explicitly designed to equalize men and women.

With innate versus learned behavior - I'm not sure that makes any difference, really. We're complex creatures - learning is part of our being, and at least some of the structure of our learning comes directly from the structure of our brains (e.g., grammar.) So if men and women are teaching boys and girls to behave like men and women - that doesn't in and of itself prove that the teaching and learning behavior aren't both being organized by something structural within our brains.

--s

I think pretty much all of your points are true. And I don't want to speak for someone else, but I think that Schaffer would probably agree with you. She does not seem to be arguing that there are no differences or that any difference we do find should necessarily be attributed to "nurture" (i.e. learned gender differences) rather than "nature" (i.e. innate differences). She seems to mainly be making the argument that many of the differences claimed so far are based on flimsy evidence and also that they do not seem to be statistically significant in comparison to the diversities within the population.

Speaking for myself, I'm fairly appalled by some of the shoddy science that gets published in this arena. No matter what they say, it is simply impossible to control for learned behaviors when conducting population studies on anyone old enough to take any kind of aptitude test. And for the most part, the research I've read doesn't particularly try to do this very well anyway.

As to why it matters, I think it depends on where your coming from. If there truly is a significant difference in the distribution of mathematical aptitudes between men and women such that there are many more men in the "high aptitude" tail of the distribution then you can take that at face value and say okay, the probability of any particular man being better at math then any particular woman is higher...though it doesn't mean that he necessarily is. If that's all your concerned with, that's fine. The issue is that it leads to sweeping conclusions that don't necessarily follow. Differences in mathematical ability at the high aptitude end do not necessarily explain why fewer women are chairs of physics departments than women. It could be a contributing factor, but as a physicist I can tell you it's definitely not the only (and I doubt it's even a significant) one. Unfortunately, such findings are often used this way.

More importantly, attributing differences in adult populations to innate differences rather than learned behavior gives us an excuse to not examine and change very real differences in how girls and boys are taught and treated. There is overwhelming evidence that girls are assumed to be less competent in certain fields, so even if there is some contribution from innate difference, shouldn't we want to give everyone equal opportunity to develop their skills?

But, coming back to the science question, I think there is a confusion that when someone refutes any particular claim of "proof" about innate gender differences, they are denying any such claims exist. I don't think a careful scientist could do this. The important concept here is that there is an absence of evidence to suggest such differences exist. I doubt if many would agree that there is strong evidence of an absence of any innate gender differences. The latter is extremely hard to prove scientifically.

I should clarify one point. As to the question of the tails of the distribution of mathematical aptitude, the way I phrased my statement was not strictly correct. If both a man and a woman are above average in math, the probability that the man is better than the woman (assuming the research suggesting fatter tails in the male distribution is correct) is higher. Similarly, for two people below average, the probability that the man is worse than the woman is also higher.

I think you are making Shaffer's point.

There is no solid data to say women have innate cognitive differences from men. She's not saying that there is none, but to date, we can't say anything meaningful. However, there are authors pushing there own shoddy readings of the data out there. So Shaffer is not exactly making a point or answering a question. Instead she is pointing out that others who seem to be pushing an agenda while claiming to be following objective truths.

Well put slatesurfer, clearer and faster than I was :)

SlateSurfer:

More importantly, attributing differences in adult populations to innate differences rather than learned behavior gives us an excuse to not examine and change very real differences in how girls and boys are taught and treated. There is overwhelming evidence that girls are assumed to be less competent in certain fields, so even if there is some contribution from innate difference, shouldn't we want to give everyone equal opportunity to develop their skills?

Good points. I was going to say that this approach would require that one teach girls precisely the same as one taught boys, even if that wasn't suitable. But I realized that isn't actually correct - whatever characteristics we're ascribing to female brains rather than male, everyone agrees I think that the differences in individuals exist, the question is whether or not you can identify them based on sex. So, given that, a good teaching method would evaluate the individuals according to their aptitudes / preferred learning styles, rather than making assumptions based on sex.

To the other point, about physics chairs, if the point of this argument is that there's no inherent sex-based bias that gives men greater aptitude in physics (or whatever) then by definition if the distribution of physics chairs isn't 50/50 male/female then it must be the case that "the system" is imposing a bias. But it's possible that both are true.

That is, we know that 2 year old boys have different cognitive abilities than 2 year old girls. Therefore, I don't think we can simply assume that 20 year old boys MUST have cognitive abilities precisely equal to 20 year old girls. Maybe they do, maybe they don't - but it's a valid question. And therefore, we can't conclude that absent any bias in the system that the ratio would be 50/50 - maybe it'd be 60/40, or 90/10, who knows?

--s

blueshift:

I think you are making Shaffer's point.

There is no solid data to say women have innate cognitive differences from men. She's not saying that there is none, but to date, we can't say anything meaningful. However, there are authors pushing there own shoddy readings of the data out there. So Shaffer is not exactly making a point or answering a question. Instead she is pointing out that others who seem to be pushing an agenda while claiming to be following objective truths.

I'm not sure of what "innate cognitive differences" means, but if we just mean that men and women might process data differently from each other, I think there is indeed plenty of evidence that this is true. Schaffer's proofs demonstrate that men and women are equivalent in outcome, but this doesn't prove that there's no difference in the processes by which we arrive at that outcome.

Also, as I pointed out elsewhere, her counter-proofs are on qualitative assessments of outcome that were intentionally designed to minimize any sex-based differences. So, yes - where men and women are the same, they're the same. But that doesn't prove there are no situations in which we think differently.

I haven't read these two books, so I don't know if they make any other claims. If all these are books are saying is that women have larger vocabularies and speak faster, then yes, I agree that we can dismiss them out of hand. But somehow I doubt that's all they're saying.

--s

AbbaZaba2000:

To the other point, about physics chairs, if the point of this argument is that there's no inherent sex-based bias that gives men greater aptitude in physics (or whatever) then by definition if the distribution of physics chairs isn't 50/50 male/female then it must be the case that "the system" is imposing a bias. But it's possible that both are true.

That is, we know that 2 year old boys have different cognitive abilities than 2 year old girls. Therefore, I don't think we can simply assume that 20 year old boys MUST have cognitive abilities precisely equal to 20 year old girls. Maybe they do, maybe they don't - but it's a valid question. And therefore, we can't conclude that absent any bias in the system that the ratio would be 50/50 - maybe it'd be 60/40, or 90/10, who knows?

--s

Hmmm, I'm not entirely sure I understand what you're getting at. Yes, I agree with you that there is no way of knowing if the discrepancy in numbers is entirely due to institutional bias or if inherent ability plays a role. That institutional bias exists is extremely clear to me (and has been demonstrated in various studies). But I'm not denying the possibility of innate differences as well. Instead, I'm objecting to people like Larry Summers saying b/c maybe there's some evidence that innate differences exist, then we can pretty much dismiss the possibility that institutional biases are contributing to the skewed ratio of men to women in these positions and excuse ourselves from having to fix them.

In fairness to Summers, he says a bit more than that. He also argued that women aren't dept chairs and CEOs b/c they are more interested in careers that allow them flexibility to spend time and take care of their families. That's probably true (though I've actually never seen data on this). But we don't know whether the reason it's true is that they've been conditioned to consider themselves as the primary caregivers or that they innately gravitate to that role. What's undeniably true is that the relatively short period of pregnancy + breast-feeding which can (but not always does) suppress productivity can derail an entire career, particularly in academia. Even if her husband wants to be the SAHD after the biological needs only the mom can fulfill are done, it often no longer makes sense b/c her career stagnated while his took off. I don't want to get into a debate about working conditions, but I'm just pointing out an example where something you might think is a strong demonstration of differences between the sexes is so hopelessly entangled in societal norms and institutional policies that it's almost impossible to really know what's contributing to what.

The Summers incident also demonstrates another point that Schafer makes. Even when we can demonstrate an innate gender difference, extrapolating to vague societal trends is very risky. It's true that having a high mathematical aptitude is important for success in physics. But no one would argue it's the only thing that's important. In fact, to be successful in academic science (incl. physics), you must be able to write clear scientific papers, put together visual poster presentations, give engaging scientific talks, and in many areas form and maintain collaborations with other colleagues. In most areas of physics aside from certain theoretical subfields, my experience has been that once you have demonstrated that you have at least the mathematical ability to pass the courses these skills are *more* important that mathematical aptitude. Notably, these are also abilities more commonly attributed to women. My point isn't to say that in reality there ought to be more women than men in physics. I honestly don't know what the "natural" ratio would be. My only point is that it makes no sense to identify one single trait and make a broad conclusion.

I've never seen any data showing demonstrable cognitive differences between 2 year old boys and girls. Can you provide the reference?

(I'll just reference SlateSurfer's comments rather than citing the whole quote)

I agree with what you're saying. I think the real question here becomes "what's the point?" - on both sides of the argument, what point are they trying to make in these books, and what point is Schaffer refuting.

It's true, as Schaffer points out, that given all this information you certainly can't point to an individual and say "because you're a girl, therefore you'll be good at this and bad at that." But is anyone actually claiming that you can? Historically of course people have indeed said that, still do, but I think it's pretty clear that this is foolish.

But I think Schaffer is missing a key point. Take her note about how only 44% of women have an "E" brain, which she uses to dismiss this concept entirely. Yes, that means given a particular woman, we can't presume she has an E brain. However, given a large enough group of women versus a similar sized group of men, we CAN presume that the women's group has more E brains than the men's group does. Also, if we get together a group of E brains and a group of S brains, that the E group will contain more women and the S group will contain more men. I think that's potentially useful information.

--s

AbbaZaba2000:

It's true, as Schaffer points out, that given all this information you certainly can't point to an individual and say "because you're a girl, therefore you'll be good at this and bad at that." But is anyone actually claiming that you can? Historically of course people have indeed said that, still do, but I think it's pretty clear that this is foolish.

But I think Schaffer is missing a key point. Take her note about how only 44% of women have an "E" brain, which she uses to dismiss this concept entirely. Yes, that means given a particular woman, we can't presume she has an E brain. However, given a large enough group of women versus a similar sized group of men, we CAN presume that the women's group has more E brains than the men's group does. Also, if we get together a group of E brains and a group of S brains, that the E group will contain more women and the S group will contain more men. I think that's potentially useful information.

Well, in all fairness we aren't told what fraction of men have E and S brains. But let's assume that a higher fraction of women do have E brains. That still doesn't answer the question of whether it's because they've learned to be empathetic or because they are innately more so. The books Schafer is debunking claim that the differences are innate. Schafer makes the much less definitive claim that the evidence they cite is shaky at best and completely misinterpreted or has been revoked at worst.

And there are real consequences to getting this correct. It drives very different social policy and educational practices if you assume that women are more empathetic on the whole than men.

But for me, the main problem is that it is simply wrong and unscientific to claim that the existing data prove the premise Pinker and Brizendine are pushing. And quite frankly this country has enough of bad science already. It seems like we're devolving to a society where it's enough to believe something is true and shape any facts to match that belief no matter how much we have to butcher reality to do so. I think that we ought to always strive toward thinking critically about ideas and information with which we're presented. And that is the exercise Schafer pursues here.

SlateSurfer:

Well, in all fairness we aren't told what fraction of men have E and S brains. But let's assume that a higher fraction of women do have E brains. That still doesn't answer the question of whether it's because they've learned to be empathetic or because they are innately more so. The books Schafer is debunking claim that the differences are innate. Schafer makes the much less definitive claim that the evidence they cite is shaky at best and completely misinterpreted or has been revoked at worst.

It's true that it doesn't answer the innate versus learned question, but Schaffer seems to be accepting the Baron-Cohen's numbers, but rejecting their relevance.

However, once again she's mixing up two different standards/measures. I think she's doing the equivalent of looking at DMV driving test scores between race car drivers and non-race car drivers, which I expect would be equivalent, and concluding that since you can't differentiate the groups based on this DMV test result data therefore there must not be any cognitive differences between the two groups.

When she speaks of "malleability," I think she's correct to a degree, but only to the extent that men and women on average are equally competent. Most people are sufficiently empathetic, and sufficiently systematic. However, the fact that most people can learn to drive competently doesn't mean that most people could learn to be winning race car drivers.

SlateSurfer:

And there are real consequences to getting this correct. It drives very different social policy and educational practices if you assume that women are more empathetic on the whole than men.

Not necessarily. As I said before in terms of social policy and education practice it shouldn't make any difference - you should base the social policy / educational practices on whether the individual actually has an S brain or an E brain, not infer the existence of such based on sex. The premise is that you can tell, and that it matters - so, there you go. Detect it and respond accordingly on an individual level.

The issue of course, is the question of equal outcomes. If 90% of physics students are male, that means some combination of two things: males are inherently better at physics, and society supports the advancement of males in physics and suppresses the advancement of females.

To my mind, there's no reason that both of these can't be true. Schaffer seems to be suggesting, though, that only the latter can be true and it's not possible that these different outcomes reflect any innate difference.

SlateSurfer:

But for me, the main problem is that it is simply wrong and unscientific to claim that the existing data prove the premise Pinker and Brizendine are pushing. And quite frankly this country has enough of bad science already. It seems like we're devolving to a society where it's enough to believe something is true and shape any facts to match that belief no matter how much we have to butcher reality to do so. I think that we ought to always strive toward thinking critically about ideas and information with which we're presented. And that is the exercise Schafer pursues here.

I wholeheartedly agree that this country has enough of bad science, and it may very well be that these books are indeed bad science, I don't know I haven't read them. However I'm not convinced that Schaffer's rebuttal represents "good science," either. I think she too is starting with a premise and rejecting out of hand any evidence that would tend to contradict her premise. --s