Asymmetrical Information

I do wish that Elizabeth Spelke had commented more on her third point, the differences in variances. Her argument on that point seems to be: 1) the SAT-M is bad; 2) men and women have the same college performance. Neither of these really speaks to the point. Her point needs to be not that the SAT-M is inaccurate, but that it produces a biased standard deviation in scores. If men actually do have a higher standard deviation, however you test it, that is enough to explain a very large amount of the variance. Similarly, having the same college performance is irrelevant unless it includes the same variance in performance. If the argument is that males' higher variation produces more mathematical geniuses despite the same average (and also more mathematical morons), you do not refute it by pointing out that the averages are the same. Maybe the study's focus on the top tier in maths controls for this already; I have not read all the research to which they refer. Working as a statistician has probably distorted my views, but I believe that a minor change in variance in ability explains quite a bit, and Steven Pinker makes the point well.

For the opposition, I am concerend with Steven Pinker's comments on the need for meta-studies. He is right that individual studies with relatively small sample sizes have little statistical power to detect differences in populations. On the other hand, even in statistics, power corrupts. By that, I mean that with a sufficient sample size, you can find a statistically significant correlation where it may not be meaningful. Once you start using a sample size of 100,000 or "Scotland," your demands for how large and how significant a connection is should probably increase. Small random changes or artifacts of the data can start looking real once you stack a million of them together. Then again, considering that we are looking for the top few thousand academics out of six billion humans, even a 0.001% difference in average or variance could have a large effect.

I've got to go with revealed preference even in matters concerning men and women.

If women want to be scientists, there is nothing stopping them. If a woman can't get hired to teach sciences at Harvard, she should start her own university. Assuming a wealth of young women wishing to go into the sciences, there will be a ready market.

Are there differences between men and women? Certainly. What are the details of the differences? Irrelevant. Every human being on the planet makes choices after taking into account their strengths and weaknesses. Only the choice really matters.

P.S I've heard many times that men have used their superior strength to dominate women through the ages. Bull. Those very early women were not weak. They were certainly capable of killing a man, individually, or banding together as necessary. They didn't. They chose not to. If women have found themselves in a culture that places them in a subservient position, they can blame their earliest female ancestors - not men. And given the freedom of the culture they find themselves in today, they can only blame themselves.

Bravo, Randy. I also don't understand why some people can't see that through the ages men and women have worked together to survive and thrive, that they don't always or even usually compete for "dominance". Each has a role, into which they are not forced by anything but nature. Modern technology has alleviated the harshness of some of the natural pressures on men and women, but has not changed their basic natures.

As Larry Summers pointed out, wherever a group is substantially underrepresented due to discrimination, anyone who focuses their hiring on the underrepresented group gains a major advantage. Markets are perfectly capable of punishing irrational discriminators. Problems only arise when discrimination IS rational, when one group mean is substantially worse than another's -- in this case, uncertainty works against strong members of the weaker group. This is a problem worth addressing, but I don't expect it to happen anytime soon (or ever?) because comprehension of the problem requires acknowledgement that the discrimination is rational.

Oops, important clarifications/corrections:

- "Problems only arise when discrimination IS rational"... when there's a vibrant market, that is. There's no shortage of historical examples of irrational discrimination held in place due to the lack of a true market (a plethora of examples exist just when considering races in sports).

- "when one group mean is substantially worse than another's"... replace "mean" with "distribution".

Only a man would think the term "cage match" is an evocative descriptor in this context.

Dog of Justice,
Do you believe that higher education (especially at the highest levels) is substantially a free market, or is it the kind of system in which irrational discrimination can be held in place? In response to Summers, some have noted the differences between universities and profit-maximizing firms, notably that universities are not profit-maximizing institutions. They also receive considerable funding from government and alumni, which can shield them from market forces, and even their customers tend to be subsidized by government (and parents) through loans and other means. There are rather significant entry costs, particularly if trying to establish a university with some status, prestige, whatever. MIT and University of Phoenix are something less than perfect substitutes.

I would expect universities to preferentially hire women already, especially into math and the hard sciences. They are aware of the gender gap, and academics are frequently the sort to want to change that. This would lead us to believe that the gender gap still exists even though universities are actively pursuing female candidates neglected by other universities. This could add great strength to non-discrimination arguments. Elizabeth Spelke cites research about rating CVs, with surprising results if we believe that universities are preferentially hiring women. Could subconscious bias overcome conscious bias on this scale?

My big problem is the overwhelming fallacy I see in Spelke's line of argumentation. Basically, we assume, by default, discrimination against women.

I don't think [morally] that's the way things should work. When you've got a charge of discrimination as an individual, you bring it to court and it gets resolved.

Spelke's argument is geared towards building social policy on a guily-until-proven-innocent basis for large groups of people. Regardless of whether sex differences are actually causing these disparities, both Spelke and Pinker seem to be working from a default position that is against the proper legal procedure in this country (innocent until proven guilty).

I would expect universities to preferentially hire women already, especially into math and the hard sciences. They are aware of the gender gap, and academics are frequently the sort to want to change that. This would lead us to believe that the gender gap still exists even though universities are actively pursuing female candidates neglected by other universities.

Yes, this happens all the time.

Elizabeth Spelke cites research about rating CVs, with surprising results if we believe that universities are preferentially hiring women. Could subconscious bias overcome conscious bias on this scale?

The problem is that the subconscious bias is probably rational -- as far as I can tell, those who act on such subconscious biases make statistically better hiring decisions than those who don't! So, if performance is rewarded sufficiently, this is the result I'd have to expect.

This is very unfortunate, since I do consider these "statistically better hiring decisions" morally wrong.

Pinker says, "We have been colleagues at MIT, where I helped attract her, and at Harvard, where she helped to attract me."

I realize Pinker is speaking, not writing, but such an awkward and funny-sounding sentence from someone who says he is an expert on language doesn't do much for his credibility.

Zubon,

"They are aware of the gender gap, and academics are frequently the sort to want to change that."

True, in general, but I wonder how applicable it is to the specific disciplines under discussion. It is often said that scientists (and by corollary, I think, engineers) are meritocratic. As such, given a choice between hiring a candidate with a particular gender and a particular publishing record, I believe that the decision would likely rest on the latter criterion.

"Fifth, many of these differences emerge in early childhood. It is said that there is a technical term for people who believe that little boys and little girls are born indistinguishable and are molded into their natures by parental socialization. The term is 'childless.'"

This describes my experience quite well. Prior to becoming a parent, I subscribed heavily to the "nurture" argument and that differences in interest in math/science/engineering are acute due to socialization. I fondly recall a conversation with our pediatrician about why how my daughter became "girly" (for lack of a better word), at which point he furrowed his brow and pointed out that she was, in fact, a girl.

For the record, my wife and I both have degrees in mathematics. I chose further study in computer science/engineering and she into applications of math in economics. (Similar to Pinker's comments, I must concede that my wife is a far better at mathematical calculation than I.)

Comments are Closed.