Life's Better Ideas

Occasional links to, and comments on, ideas that I think will make this a better world, and remarks about things that need fixing, too.

Location: Denver, Colorado, United States

Tuesday, March 22, 2005

Gender Bias

There's been a fair amount of commentary lately about the differences in male/female accomplishments lately, particularly with the Larry Summers flap at Harvard and the Susan Estrich/Michael Kinsley dustup at the LA Times. Today I read an article by Amy Sullivan at Washington Monthly that thoughtfully explains some differences that occur in grade school that might help explain why there are fewer female columnists, for example.

What researchers found should track closely with memories of your second-grade self. Those are the years in which children learn how to participate in group discussions. The teacher explains that to maintain an orderly conversation and allow everyone to speak, students should raise their hands when they have something to say and wait to be called on. Simple enough. But what happens next? Usually, the teacher poses a question to kick off the discussion, and several children raise their hands with answers. As the conversation continues, one or more of the boys, either overly enthusiastic about his point or merely impatient, calls out his comment without waiting to be chosen by the teacher. She might stop him and remind everyone of the rule—raise your hand and wait to be called first—but often she just lets him go ahead. It is less disruptive, after all, than letting him jump up and down, waving his hand, and yelling, “Oooh, ooh, me!” But if a girl bursts out with a thought, the teacher's response changes. The Sadkers report that teachers almost always chastise girls who violate the rules. After all, teachers rely on girls—their “good students”—to remain quiet and maintain order in the classroom while teachers focus on the boys, keeping them in line by drawing them into discussions. And so, as the conversation races around them, girls sit, waiting to be called on, first holding up their tired arms with the other, then lowering them, and finally not bothering to raise their hands at all.

I guess that's the way I recall it, too. It's my perception that most grade school teachers are women. I wonder if the tendency of most teachers at these grade levels is peace-at-any-price and an unwillingness to, or difficulty in, maintaining order. I suppose a teacher could lay down some ground rules like, for example, alternating girls and boys for discussion. That way everybody gets a chance to comment.

Update: Ruth R. Wisse has some strong words about the Larry Summers flap. HT: PowerLine.


Blogger goyishekop said...

We can all dredge up any number of memories
to justify just-so stories that "explain"
differences in performance between women
and men (or explain anything else, for that matter).

Indeed, any number of just-so stories have been constructed to do so. Hey -- I get drunk at parties and make them up myself.

In this particular area (math and sex differences), there's a wealth of data. To the best of my knowledge, only one sociological factor seems to be supported by careful data analysis: *On average* women devote less time to their careers.

This difference produces differences in *average* performance in other fields, and Summers speculates that it could also apply to Harvard.

I don't mean that there are no other possible such explanations -- only that people have made many such hypotheses and that's the one that's held up so far.

What's triggered the Summers firestorm is something else: a higher variance in ability among men will produce grossly unequal representation among samples taken from either tail. That's not an opinion about causes of difference in variance, it's a statement about mathematical distributions.

The data support such a difference in variance. The tiny fraction of men who are really, really good at math, and the tiny fraction who are really, really bad is a bigger fraction of that half of humanity
than in women.

Put another way, women are slightly more humped up in the middle, men slightly more evenly distributed across the spectrum.

Even tiny differences will produce a big
difference in samples chosen from the tails.

Early on, I pointed this out to a friend,
and wondered why Summers had focused on
mean performance, which people are forever arguing about, and not on variances, which seem not to be in dispute.

Imagine how surprised I was to find, later, that Summers had been talking about variances.

Indeed, Summers declines to opine on whether there's a difference in average ability.
It's simply not relevant to his point.

Bottom line: any explanation
to account for the differences in representation at the high-end
has to produce a difference
in variance: a flatter curve for males,
with longer tails at both ends.

This difference in variance is seen for
many traits, not just math and science,
and not even just intellectual.

Any sociological theory to explain away
the difference in spread that confines itself to math/science will smell like special pleading.

7:21 AM  
Blogger David Aitken said...

Thanks, Jeff, for a valuable explanation.

10:16 PM  
Anonymous Anonymous said...

I've seen one teacher come up with a strictly fair way for every kid to get their say, which is to proceed in a circle around the room. Each kid with a hand up gets their chance, and knows that there's nothing more to do than raise one's hand to get their say.

4:53 PM  

Post a Comment

<< Home