February 25, 2005

Estrogen Week: Poor Larry Summers, Victim of Political Correctness

Yes, yes: Poor Mr. Summers. Why, all he did in his talk about women in science and engineering was point out what everyone knows anyway:

There may also be elements, by the way, of differing, there is some, particularly in some attributes, that bear on engineering, there is reasonably strong evidence of taste differences between little girls and little boys that are not easy to attribute to socialization. . . . So, I think, while I would prefer to believe otherwise, I guess my experience with my two and a half year old twin daughters who were not given dolls and who were given trucks, and found themselves saying to each other, look, daddy truck is carrying the baby truck, tells me something. And I think it's just something that you probably have to recognize.

There are two other hypotheses that are all over. One is socialization. Somehow little girls are all socialized towards nursing and little boys are socialized towards building bridges. No doubt there is some truth in that. I would be hesitant about assigning too much weight to that hypothesis for two reasons.

First, most of what we've learned from empirical psychology in the last fifteen years has been that people naturally attribute things to socialization that are in fact not attributable to socialization. We've been astounded by the results of separated twins studies. The confident assertions that autism was a reflection of parental characteristics that were absolutely supported and that people knew from years of observational evidence have now been proven to be wrong. And so, the human mind has a tendency to grab to the socialization hypothesis when you can see it, and it often turns out not to be true.

The second empirical problem is that girls are persisting longer and longer. When there were no girls majoring in chemistry, when there were no girls majoring in biology, it was much easier to blame parental socialization. Then, as we are increasingly finding today, the problem is what's happening when people are twenty, or when people are twenty-five, in terms of their patterns, with which they drop out. Again, to the extent it can be addressed, it's a terrific thing to address.

(Incidentally, that's a whopping logical fallacy in the second-to-last paragraph above. Consider:

a. Some people suggest differences in scientific ability between men and women can partly be attributed to socialization.
b. Some other variations in human behavior once attributed to socialization have since been proven not to be attributable to socialization; therefore
c. The set of gender differences we are currently discussing cannot be attributed to socialization.

Uh, no. And he's president of Harvard.)

Yes, yes . . . poor Larry Summers:

Today this study from the American Institute of Physics argues that the disparity in the numbers between men and women working as university faculty in physics actually begins before college and is not a result as of a lack of innate ability etc.
What's all this fuss about? We know women's brains are different from men's. They can't even read maps as well:
I remember studies like this being on the front page of Time in the 1980s. Women, men, math, maps, different areas of the brain. Same stuff. But look at the percentages. You still got millions of men and women who don't fit into the categories. We are all a lot more variable than these studies suggest. (Actually, the studies themselves are usually rigorous, but the popular science reporting isn't. Remember, this is the same MSM that reports on the war in Iraq and the 2004 election. Scientists are not happy when their findings are generalized out of all recognition.)
I'm telling you, it's those rotten PC police, stifling dissent just like they always do:
These days, to be politically correct is to be a cowed conformist, too afraid to speak the truth which -- surprise! -- invariably turns out to be an old conservative idea.

We have come full circle so that uttering a stereotype is classified as daring. Soon, the brave new thing will be to call your 56-year-old secretary a "girl" and the forward thinkers will say that white men can't jump. Summers's own breath of fresh air was to challenge the notion that men and women can be equal at science -- an idea so politically correct that not a single woman has held a math chair at Harvard in 370 years.

Barbara Grosz, chair of the Harvard task force on women in science and engineering, recaps the argument with some exasperation: "The criticism of Summers's talk was not that the ideas he expressed were politically incorrect, but that they were just plain incorrect." How come, she wonders, when Summers talks he's being open-minded and provocative, but when his challengers offer a spirited rebuttal they're accused of trampling on academic freedom?

When will these P.C. monsters learn to just trust the science? WHERE'S THE SCIENCE?

When they do study sheer cognitive prowess, many researchers have been impressed with how similarly young boys and girls master new tasks.

"We adults may think very different things about boys and girls, and treat them accordingly, but when we measure their capacities, they're remarkably alike," said Elizabeth Spelke, a professor of psychology at Harvard. She and her colleagues study basic spatial, quantitative and numerical abilities in children ranging from 5 months through 7 years.

"In that age span, you see a considerable number of the pieces of our mature capacities for spatial and numerical reasoning coming together," Dr. Spelke said. "But while we always test for gender differences in our studies, we never find them."

Poor, poor Larry Summers.

(Credit where it's due: Boston Globe and NY Times excerpts via feministe [here and here], whose coverage of this incident has been the most comprehensive I've found.)

Posted by Ilyka at February 25, 2005 10:53 PM in estrogen week
Comments

c. The set of gender differences we are currently discussing cannot be attributed to socialization.

I don't think that's what he said.

"And so, the human mind has a tendency to grab to the socialization hypothesis when you can see it, and it often turns out not to be true."

Sounds to me like what he's saying is that in the modern PC environment, people have a tendency to jump to the conclusion that any visible differences are attributable to socialization, when that may not be the case. He never said the differences under discussion cannot be caused by socialization, just that they may not be. Which, I think that was the whole point of his speech: not "women belong in the kitchen, barefoot and pregnant, because they suck at math," but rather "why don't we try to figure out if the observable differences in outcome are caused by natural differences in brain structure?"

I can understand why women might be insulted by this suggestion, but that doesn't mean he intended it the way many have interpreted it.

Posted by: Mike C at February 25, 2005 11:50 PM

Oh come on, Ilyka. The only thing Summers did wrong was to apologize so abjectly. I read his comments and found nothing exceptionable. Much of the reporting on them, however, gave me the strong impression that the writers either didn't bother to read what he actually wrote, or are themselves so math-challenged they can't understand statistical distributions or the concept of variance. (How else to explain the absurd contention that "Summers said girls can't do math!", "Summers said females can't be top scientists!"; or the offering of missing-the-point anecdotes such as "But I know girls who are really good at physics and math!"?)

Summers's autism example was just that - an example of prevailing wisdom that was in time proved to be erroneous. I certainly didn't read it as an attempt at proof-by-analogy. The fact that "cognitive differences are all the product of socialization" is the prevailing wisdom does not in itsef mean that it is wrong - but Summers isn't making that argument. By the same token, however, the fact that the hypothesis "some cognitive differences may be the product of biology" reflects traditional assumptions is not evidence one way or the other as to whether it is a true statement, despite what Ellen Goodman seems to think.

Summers's detractors believe (or want us to believe) that there isn't a shred of evidence to support a hypothesis that high mathematical ability may have a biological component. And that just ain't so. So why can't we talk about it?

Posted by: Moira Breen at February 26, 2005 01:52 AM

He never said the differences under discussion cannot be caused by socialization, just that they may not be.

I'll grant that "cannot" is too strong in context, but he's using some pretty weak evidence to say the socialization argument should not be "assigned too much weight:" a story about his twin daughters playing house with trucks (anecdotal), a tale of a trip to a kibbutz (which I didn't quote both because it was lengthy and because it wasn't clear, due to his wording, what the conclusion was, though I suspect he was saying that despite starting out without gender restrictions for tasks, the kibbutzes ended up with women working in nurseries and men fixing tractors--again, though, anecdotal), and two examples of subjects unrelated to gender that turned out to go against the prevailing wisdom of the time--twins and autistics.

Not the most effective support there: two anecdotes and, well, apples and oranges.

As to this:

Summers's detractors believe (or want us to believe) that there isn't a shred of evidence to support a hypothesis that high mathematical ability may have a biological component. And that just ain't so.

I'd like to know why, then, he didn't use that evidence, instead of dragging in twins, autistics, and kibbutzes.

In contrast, we do have a study, linked in the post by Lies and Statistics, that appears to refute the notion that "biological differences" are chiefly responsible for the gender gap:

Ivie added that she thought innate differences to be an unlikely cause of the disparity. She noted that although women earned only 18 percent of the Ph.D.ís in the United States in 2003, this is a large increase from the 2.4 percent of total Ph.D.ís earned by women in the 1970s.

Such a quick change in innate ability would be highly improbable, she said.

So yeah, hey: Why can't we talk about that, instead of lamenting Summers' treatment at the hands of the P.C. police?
Posted by: ilyka at February 26, 2005 02:26 AM

Why didn't he exhaustively cover all the evidence? I dunno, maybe because he wasn't giving a detailed scholarly talk on the subject, but was addressing the refusal to even consider the possibility that socialization and/or discrimination don't explain everything? Ilyka, are you really going to go out on a limb here and assert that no study has ever found a link between testosterone and visuo-spatial intelligence? That nobody has tested and verified comparable sex differences in other mammals? That androgenized girls don't tend toward "male" interests and to have higher math scores?

As for the study you link, Summers did not claim that other factors (discrimination, etc.) could not be at play in the disparate representation of the sexes in certain fields. Nor does that study disprove innate differences as a factor. I'd also ask you to re-read the linked article carefully - there is little or nothing in there, including the very first line, that counts for evidence against the existence of biologically based differences in the distribution of mathematical ability. And the statement in this paragraph:

"The findings of the study also run counter to Summersí contention that the primary reason for the relatively small number of women in academic positions in the sciences is due to conscious choice, as a result of the conflict between the long workweek of tenure-seeking professors and the desire for child-rearing."

is not supported by anything else in the article. Maybe it is in the study itself, but the article doesn't elucidate.

I do think you left out the more interesting bits from the NYT article you linked in your post, that would have been more supportive of an "anti-innate" view. The part you did quote was about basic math ability in children; the paragraphs following, however, do describe the significant sex differences in SAT scores that appear later. And I think the authors are putting a bit of effort into fuzzing up the issue of averages vs. the issue of variance. It's the tails of the distribution that are at issue here.

What is interesting is the part about the scores from Japan and Iceland - unfortunately, it's not clear if they're talking about averages or the scores at the tails. Another good point raised was that girls think they have to have 790s to pursue any kind of engineering or science career, whereas boys with relatively mediocre scores trundle along just fine.

None of this, however, make Summers's speculations heinous or beyond the pale. Yeah, I do think the people hammering him are "pc monsters", who've shown a marked inability to dispassionately debate a perfectly frickin' reasonable hypothesis. I really can't see how a bunch of women running around essentially saying "lordy me my fluffy little female brain just can't get a handle on this 'variance' thing" are doing much to demonstrate the scientific and mathematical abilities of women.

Posted by: Moira Breen at February 26, 2005 03:57 PM

Why didn't he exhaustively cover all the evidence?

Mmm, straw men for tea-time . . . my favorite!

I did not state that Summers should have attempted to "exhaustively cover all the evidence." I did state that he could have provided better support for not assigning "too much weight" to the socialization argument.

Like it or not, the burden of proof is on Summers:

Shifting the burden of proof is a specialized form of the argumentum ad ignorantiam. It consists of putting forward an assertion without justification, on the basis that the audience must disprove it if it is to be rejected.
So yes, I expect a bit better than he gave. I don't think that's unreasonable.

Ilyka, are you really going to go out on a limb here and assert that no study has ever found a link between testosterone and visuo-spatial intelligence?

No. Which is why I haven't.

As for the study you link, Summers did not claim that other factors (discrimination, etc.) could not be at play in the disparate representation of the sexes in certain fields.

Nor did I claim he did.

That's enough straw. Let's get to what I did say and what I did cite:

Nor does that study disprove innate differences as a factor.

No--just assigns them low probability. And again, the burden of proof is on Summers. His critics need not disprove all theories of innate ability to note with justification that his support for that argument was extremely poor.

there is little or nothing in there, including the very first line, that counts for evidence against the existence of biologically based differences in the distribution of mathematical ability.

As above.

The part you did quote was about basic math ability in children; the paragraphs following, however, do describe the significant sex differences in SAT scores that appear later.

Right. And we know some innate differences can "lag" (hello, puberty), so I'd grant that this in itself does not disprove an innate ability hypothesis either. We don't know whether that shift is a result of innate ability. We also don't know whether it's a result of socialization.

Again: If you're discussing the gender gap in the sciences, and you make an argument that could reasonably be perceived as a "cop-out" (as it requires no adaptation on your own part, and as the same argument has been used in the past in just such fashion), it's up to you to support it. One can take issue with Summers' abject performance in doing so without being a "P.C. monster."

I think the authors are putting a bit of effort into fuzzing up the issue of averages vs. the issue of variance. It's the tails of the distribution that are at issue here.

Yes--which Summers does make clear in his speech. As for whether the authors of the article are putting effort into deliberately fuzzing up the issue, frankly, I doubt it. Most science and statistical reporting is poor from ignorance, not from malice. But who knows?

None of this, however, make Summers's speculations heinous or beyond the pale.

I don't think they were either. I do think they were dumb.

Yeah, I do think the people hammering him are "pc monsters", who've shown a marked inability to dispassionately debate a perfectly frickin' reasonable hypothesis.

Frankly, I'm seeing far more "passion" from his suporters. Bill Ardolino reads one study (the maps/gray matter/white matter one linked in this post) and concludes that "someone owes Harvard President Larry Summers an apology." An apology for what, criticizing him?

You, meanwhile, are resorting to spittlelicious closers like this one:

I really can't see how a bunch of women running around essentially saying "lordy me my fluffy little female brain just can't get a handle on this 'variance' thing" are doing much to demonstrate the scientific and mathematical abilities of women.

Right, because that's really what they're saying. Honestly, Moira, ENOUGH WITH THE STRAW. It's not doing much to demonstrate the reasoning abilities of women.

Posted by: ilyka at February 26, 2005 10:27 PM

Sorry I bagged on this last week (illness + dead cable).

If the points I raised are straw men, can I then conclude that you do not deny that there is evidence for innate sex differences in mathematical ability, it's just that you think the arguments offered by Summers were of poor quality? Not unreasonable.

However, that conclusion seems to contradict some of the rest of your commentary. Why is a hypothesis supported by evidence - whether or not all of it was proffered at the time of speculation - a "cop out"? You've agreed that he wasn't obligated in this venue to cover the subject exhaustively. Sounds like a starting point for exploration, to me. And I must be misreading you, because surely you cannot be arguing that the truth or falsehood of innate cognitive differences is dependent on how or by whom that theory was used in the past.

One error above that stood out: uh, no, the data presented in that newspaper article do not "assign a low probability" to innate differences being part of the explanation for disparate representation by sex at the high end of math-heavy fields. You have to seriously over-interpret the results and ignore a host of cross-field as well as biological data to make that conclusion from that study.

As for my "spittlelicious" comment (great word, btw) - well, yes I do think that the many commenters (including at least one you linked) who argued "but I'm very good at reading maps", "Summers says women can't do science" have indeed made unintelligent statements.

Posted by: Moira Breen at March 4, 2005 01:40 AM