Log in

No account? Create an account
Sex differences in math/science - Arvind Narayanan's journal [entries|archive|friends|userinfo]

Sex differences in math/science [Dec. 30th, 2008|11:09 am]
Arvind Narayanan
[Tags|, ]

I've been posting a lot on other forums lately and I'm going to repost things here occasionally, because I enjoy your comments.

This is what I feel on the issue of intrinsic sex differences in math/science ability:
What I fail to understand is why this is such a huge issue. Whatever sex or gender differences there may or may not be, it is clear that they are not sufficient to explain the vastly skewed sex ratio we find today, and that bias plays a big role in explaining it. So why not work together to eliminate the bias instead of getting bogged down in divisive debates?
As I see it, if and when participation levels in math and science reach, 55-45, or even 60-40, then it will be time to talk about whether that last 5-10% is due to bias or intrinsic differences.

Oh, by the way, I guess I've toned down my views on the sex ratio in the VC industry, mainly as a result of the comments there :-)

So what do you think?

[User Picture]From: paper_crystals
2008-12-30 05:34 pm (UTC)
I look at it similarly:

If there is a genetic or otherwise gender bias that somehow we have no control over this doesn't mean that all women should abandon science en masse nor that they should be discouraged from going into the field. Bias doesn't mean that some women are good at science and the only way to find out which ones are is to encourage all of them.

If there isn't a genetic or otherwise bias that we have no control over than we should still encourage women to get into maths and science because we are loosing valuable talent by not doing so.

So really either way the best conclusion is to try to eliminate the bias regardless of the cause of the bias.
(Reply) (Thread)
[User Picture]From: ethernight
2008-12-30 09:57 pm (UTC)
It seems unlikely to me that there are many reasonable people suggesting we should discourage women from getting into math or science. I think it is usually brought up as a counter-argument to the idea that these gender gaps are due to oppression and sexism.

I personally think it is much more complex to think that the gaps are either just caused by sexism, or just caused by gender predisposition. However, I do think it is just as harmful to teach women that they are being oppressed if they are not, as any actual oppression.
(Reply) (Parent) (Thread)
From: kupamanduka
2008-12-30 05:49 pm (UTC)
Are you talking about graduate studentships and higher positions in STEM disciplines? If so what about the "standard deviation" argument? I quote from ex-Harvard president Summers' controversial speech :
if one is talking about physicists at a top twenty-five research university, one is not talking about people who are two standard deviations above the mean...But it's talking about people who are three and a half, four standard deviations above the mean in the one in 5,000, one in 10,000 class. Even small differences in the standard deviation will translate into very large differences in the available pool substantially out. [via]
At least partly due to a similar reason, there are many more math and science retards among males than females ( lazy to produce statistics ). Does anyone care to be bothered? If you find such differences in statistics that cover the *average* person on the street then we might have reason to suspect that there may be bias.
(Reply) (Thread)
[User Picture]From: arvindn
2008-12-31 01:55 am (UTC)
of course we're talking about the standard deviation argument, that's the only argument from the other side that has any merit.

you're claiming there's no sexism at all in math and science? awesome. perhaps fixious has more energy to argue with you than i do :-)

Edited at 2008-12-31 01:57 am (UTC)
(Reply) (Parent) (Thread)
From: kupamanduka
2008-12-31 07:20 pm (UTC)
I did NOT make ANY statement about whether sexism exists or not. From your reasoning it seems that you are making a MATHEMATICAL flaw :
As I see it, if and when participation levels in math and science reach, 55-45, or even 60-40, then it will be time to talk about whether that last 5-10% is due to bias or intrinsic differences.
So you seem to believe that a participation difference of greater than 60-40 *at any level of study* implies bias.

On the other hand, I don't think such a consideration would have arisen if you had used standard deviations, because some statement along the following lines should be easy to prove :

If two normally distributed random variables X_1 and X_2 have the same mean m and standard deviations d_1 and d_2, where d_1 > d_2, then for every epsilon > 0 there exists N such that for all m > N, p(X_2 > m)/p(X_1 > m) < epsilon.

vinodv once told me that you used the standard deviation argument to refute a conspiracy theory I used to have about thrice as many men as women committing suicide in developed countries. So what is the aspect of your argument that applies to academics and not suicides?

Sexism may or may not exist, but I refuse to take it as obvious, particularly when every other professor seems to be an over-correcting liberal.
(Reply) (Parent) (Thread)
[User Picture]From: arvindn
2008-12-31 07:22 pm (UTC)
dude. all this is irrelevant.

i'm saying that sexism exists because i see lots of sexist jerks around me all the time. it has nothing to do with math.
(Reply) (Parent) (Thread)
From: kupamanduka
2008-12-31 07:39 pm (UTC)
Fine, may be you see something I don't see, and whatever I see is wrong, but that was NOT the point of my comment. The point of my comment was that the argument of your post doesn't hold. Care to look through the rather trivial math?
(Reply) (Parent) (Thread)
[User Picture]From: arvindn
2008-12-31 07:42 pm (UTC)
sorry man, i'm not going to. math is overrated.
(Reply) (Parent) (Thread)
From: kupamanduka
2008-12-31 07:52 pm (UTC)

Yes, math is overrated.

Wait, you are thinking that I am claiming sexism doesn't exist, or that I was dismissing your claims to have observed sexism. I know I myself have many sexist inclinations, and believe me I struggle hard to correct myself.

But what I am saying is that you are in all likelihood making a mathematical mistake in your post. I agree that math is overrated, but that doesn't mean you should peddle wrong mathematics, right? Otherwise what is the difference between you and vedic mathematics guys? I consider your intelligence to be somewhere between mine and Terence Tao's, so it will be really sad if you indulge in mathematical quackery.
(Reply) (Parent) (Thread)
[User Picture]From: arvindn
2008-12-31 08:07 pm (UTC)

Re: Yes, math is overrated.

alright fine. i'm going to answer that. but please consider the thread closed.

there are multiple data points that suggest that in the absence of sexism, the true ratio would be no worse than 60-40. kd-phd posted cs enrollment data from top 10 schools from 25 years ago, when the stereotype of computer nerds had presumably not yet been established. the results were close to parity.

second, xkcd is about as elite as it gets. there are arcane references to lisp, relativity theory, and reasonably advanced math all the time. and yet, since it is a culture that is especially enlightened in how it treats women (one of the characters has a hobby of shooting down sexists, and this a reflection of the culture), the readership is around 2/3:1/3. this is in spite of the fact that the readers are drawn from those with a math/science incliniation, which is an already skewed group.

i could go on.. there's data from other countries that don't have a cultural stereotype.. but i don't have time.
(Reply) (Parent) (Thread)
[User Picture]From: cubby_t_bear
2009-01-01 01:27 am (UTC)
It might simply be a case of intrinsic differences being amplified by network effects. But that's not really the question we want to ask, is it?

The question we want to ask is, what achievable personnel selection policy best encourages the development of science and the improvement of society? It is a common fallacy to believe that science is, or ought to be, a meritocracy. If it is a discovered that a particular gender ratio (or native speaker ratio) creates a better social dynamic for the sciences, resulting in faster advancement, we should aim for that demographic mix. I can plausibly imagine a situation where a heterogenous research group with varied backgrounds and perspectives and interactions does much better than a group of "smarter" people who all think exactly the same way.
(Reply) (Thread)
[User Picture]From: arvindn
2009-01-01 01:42 am (UTC)
"The question we want to ask is, what achievable personnel selection policy best encourages the development of science and the improvement of society?"

no, that's the question you want to ask. i don't care about that nearly as much.

it's not a "fallacy" that science ought to be a meritocracy, it's my opinion. i'm not arguing that it's better than yours, but you should recognize that political positions aren't right or wrong, but different viewpoints. you can't impose yours on me.

as a libertarian, individual opportunity is paramount to me, and advancement of society is a distant second.
(Reply) (Parent) (Thread)