August 13, 2007

the importance of math education

There's an article in the New York Times pointing out that it is logically impossible for men and women to have different average numbers of sex partners. Which is a problem, since men and women report significantly different lifetime averages (e.g. 4 sex partners for women, 7 for men in the US, or in Britain 12.7 for men and 6.5 for women). Anyone who is decent at math and remembers the definition of the mean will be able to tell you that in a world with about the same number of men and women, it is impossible for men to have more hetero sex partners than women. Indeed, a mathematician has actually proved it. So, explanations?

1. The researchers say it might be that men have significant numbers of sexual experiences with women who aren't counted in the survey, like prostitutes or women in other countries. That would mean that there's got to be some country somewhere where women have many more sex partners than men, and thus far no one's found it. Also, I question their ability to screen prostitutes out of the survey data.

2. Men and/or women are lying their heads off in what they think are the socially approved directions. I think this is quite likely.

3. Men and women have very different distribution patterns for their data. The mean number of partners must be the same for men and women, but that doesn't mean the median number of partners must be the same. As an example, imagine that there are only 5 men and 5 women in the world, and a total of 13 sexual partnerships. The men have had 0, 1, 3, 4, and 5 sexual partners; the women have had 1, 1, 1, 5, and 5 sexual partners. The mean for both groups is going to be 13/5 = 2.6 partners, but the median number of partners for men is 3, while the median for women is 1. That's a huge percentage difference just because women had a few people with high numbers of partnerships, while most women still had fewer sex partners than most men. Is this likely on a large scale? I don't really think so - it's a lot harder to skew data like that when you're not making it up and there are thousands of people in your data set. But I don't really know, and apparently the social scientists who could find out for us aren't. My best guess still goes to #2.

This is an example of bad math education perpetuating a particular social ill, namely the idea that men get around more than women. To his everlasting credit, the mathematician who inspired the article points out exactly this: that taking this data as accurate works to “reinforce the stereotypes of promiscuous males and chaste females" and may skew self-reporting and behavior. If math education were better, and people understood mean and median and interpreting statistical graphs, and if the news media could therefore publish graphs of the statistics in question, there'd be a lot more ways for people to check stereotypes against reality.

When will those darn math teachers do something about it?


Lenoxus said...

Perhaps the lateness of this point makes it break some kind of blog etiquette, but, here goes... what about homosexuality? Wouldn't the whole thing also work out if more men have sex with men than women with women? Nonetheless, I'm still going with your second possibility as well.

Laurel said...

The studies actually looked at just the number of heterosexual partners. So we're back to option 2, unless a significant number of gay men are claiming their male partners as heterosexual liaisons - in which case we're still back to option 2, since they'd be lying their heads off in a socially approved direction.