Thursday, July 31, 2008

Show Me the Test Scores

I am really getting sick of all the discussion about Janet Hyde's new data showing a lack of difference between the mean scores of males and females on math tests. Cool. But. I really don't give a flying fuck if the variance of the distribution of the males' scores is larger than that for women. And it pisses me off when people try to use that somewhat unreliable* statistic to explain why women are underrepresented in science and engineering.

The quote from Hyde's paper that some people seem to be having such an orgasm over:

If a particular specialty required mathematical skills at the 99th percentile, and the gender ratio is 2.0, we would expect 67% men in the occupation and 33% women.

Well whoop-dee-doo. However, I can't think of a particular specialty that actually does require that level of performance ON A STANDARDIZED TEST THAT YOU TOOK IN HIGH SCHOOL. Let's get real here. Anyway, the ratio is 1.2 at the most. And,

Yet today, for example, Ph.D. programs in engineering average only about 15% women.

They must be pulling from the EXTREME right tail.

Just to get anectdotal, I scored high on those tests, but not that high. Of course I did go on to get a degree in one of those fields that is supposed to require high levels of math abilities after spending years doing nothing more mathematical that learning polyrhythms. Somehow, after a few years of actually using teh maths every day, yes, I decimated the GRE quantitative section, which one would expect to be a reasonable analog for the types of tests that high school students take these days. Is that me trying to hint at an explanation for the increased variance in boys? No. Because I don't think it matters. You don't have to score on the extreme right tail in order to join the scientific community - it's not MENSA.

I would like to suggest that anyone who wants to make an argument suggesting that the reason that I was one of only four women in an entering class of fourteen, or that I work side-by-side with eight men and two women under a male PI is that, in order to walk the hallowed halls of the Academy, you have to have a score above the 99th percentile on a multiple choice test should be required to disclose the entire history of their standardized test results, and if any are found to have scores below the 99th percentile, they should be summarily kicked to the curb. Losers.

*For more substantive posts on this topic, see Janet's post on Adventures in Ethics and Science, and Jake Young's post at Pure Pedantry. The winner for the best headline goes to Scott Lemieux at Lawyers, Guns and Money.
h/t: Bitch PhD
Just don't read the comments unless you want to fume like me.
Edited to add: Also check out Academic's take on this.


Rebecca said...

I think some of this is the public's misconceptions about scientists. Scientists are not necessarily the smartest of the smart. I'd say they're all very smart, but the common denominator is more like they're the most persistent of the persistent.

And is there some specialty that requires 99th percentile math skills? I must have missed that one. It's pretty stupid to think that a score on a standardized test is anything more than one's ability to take the test at the time the person took it.

acmegirl said...

Right on, Rebecca. Some of the smartest people I know are really not very good at standardized tests. But ask them to give you a chalk talk on their research and they will blow you away.

Persistence does seem to be a key factor as well. You have to be willing to fail repeatedly to succeed in science.

Academic said...

Another about the highest-tail is that it seems to assume that people cannot be in the highest-tail in multiple areas. We do what we decide is worth doing.

Jenny F. Scientist said...

I had to take a subject GRE to go to grad school- the test was in my degree field, mind you- and I missed over 1/3 of the questions: all the ones on quantum chemistry. Speaking of standardized tests. Have I ever used the Heisenberg equation again? I have not. Tests: lame.

Nicky said...

It comes down to this: people are desperate to blame the lack of women in the sciences on anything other than something societal/institutional. Remember, if it's biological, they're off the hook as far as trying to fix the problem.

I'm in one of those fields that does require high math skills, and where women are unbelievably under-represented (my incoming PhD class was 3 women out of 40, and we were the most gender-balanced class in years) but you'll never convince me it has anything to do with tests of high school algebra. The year I served as a student rep on the PhD admissions committee more than showed me that it's all about faculty attitudes towards women. The ways male vs. female applicants were discussed in committee was disgusting. It didn't help that I was the only woman (out of ~20) on the committee.