Another post on dating strategies

Well, to be fair, the data for the UK is (implicitly) reported as an “average”, not a “median”, and it’s reasonable to assume, unless you have information to the contrary, that “average” = “mean”. Furthermore, it’s impossible to have a median of 12.7 sexual partners. To defend my fellow mathematicians, I think we all know what “median” means.

Doormat: I know a guy who has exactly 12.7 heterosexual partners. You see, one of his girlfriends was 70% heterosexual and 30% homosexual.

Yes, the point is valid if you are using mean rather than median. False reporting (including different definitions of sex) are almost certainly the biggest explanation. Prostitution can’t be, since prostitutes are part of the universe. Sex tourism could genuinely result in different means, but come on, how big a factor is that?

Richard Lewontin has a wonderful essay on the inability of surveys to capture the reality of sexual behavior, which has a good discussion of this issue.

It also mentions that on the big NORC sex survey that set the standard for these things in the 1990s,

Two matters were asked about that were considered so jarring to the American psyche that the information was elicited only on written forms: masturbation and total household income.

The file “Boxing Helena” could explain the .7th partner.

Prostitutes and other women who have a lot of sex may be part of the sample, but will they report accurately?

Once I was flirting with a beautiful woman, and this topic came up, and she said she lost track around 70. I think I succeeded at not looking too startled or disapproving, and complimented her on her memory. I’d never before met a woman who claimed to have gotten past 30 without loosing track.

I can imagine there might be many more women with high numbers, since a man would have to persuade many women while a woman need only be easy to persuade. And if those women give underestimates then it could all work out.

“I would say that men have more partners on average…”

Sounds plausible because the language is sloppy. Of course the mathematical averages must identical, because Npartners_women = avg_women*Nwomen, Npartners_men = avg_men*Nmen. By definition, Npartners_women = Npartners_men, and Nwomen ~= Nmen.

Nevertheless, the typical or so-called “average” woman probably has fewer partners than the typical or “average” man, which would be captured by the medians.

I read about this phenomenon somewhere a few years ago, and apparently men and women use different methods for determining how many people they’ve slept with. Men, when given one of these studies, are more likely to give a ballpark estimate, while women are more likely to count up specific partners. I guess this could have some effect on the Brits’ averages, but probably not on the Americans’ medians, since I can’t imagine counting vs. estimating making much of a difference if you’ve only had sex with 4-7 people.

Also, the difference between the Brits and the Americans could be due to using medians vs. means. For example, I’d guess that a fair number of American men claim to have had sex with more than 14 women, while I doubt that anyone was claiming to have had sex with fewer than zero women. So the Brits might still be prudes after all.

One more thing: I haven’t read these surveys, but the people conducting them would have to be incompetent not to give a relatively clear definition of “sex partner.” So my guess is that respondents are just lying–either the men are exaggerating, or women understating, or both.

“oh that time with Chet? That didn’t ‘count'”

the author of the article probably doesn’t know what the difference between a mean and a median is

I’m fairly certain that Gina Kolata, who has been a science reporter at the Times for at least the last 25 years knows the difference between a median and a mean. The article may or may not be any good, but there’s no reason to assume this level of ignorance on the part of someone who has reported on science news her entire career.

On the merits, I would generally agree with post 4 about the inaccuracies reporting and the different definitions of sex. That said, my intuition–that men would define sex more narrowly and phallocentricly , to mean only “vaginal penetration leading to male climax” and that women would have a broader definition–would lead to the opposite results, in which men would report fewer sex partners than women.

jacob:

Well, given that her idea of “science news” involves omitting entire interviews and ignoring caveats of study results, surely she knows what she’s doing. Is she ignorant, or is she just being catty? Surely the truth lies somewhere in the middle.

If we’re being picky, then please specify *arithmetic* mean.
2 more fudge factors for ye: (1) Perhaps women only count a threesome once (2) Alcohol/drug-initiated liaisons: perhaps women are less able to remember such encounters; perhaps men are more likely to fill in any memory blanks with some wish-fulfilling ex post facto consummation.

I covered this in Slate today. The numbers from the UK (which is from a 2001 Lancet study called “Sexual Behaviour in Britain: Partnerships, Practices, and HIV Risk Behaviours” if you want to look it up) are indeed means. And I think it is highly, highly unlikely that Gina Kolata doesn’t know the difference between a mean and a median. Rather, I expect she made a conscious choice to gloss over the distinction in this article, and I think that was the wrong choice.

This is silly. There’s a bunch of reasons to believe that both the mean and median would diverge in this way before you even consider intentional false self-reporting.

First, I simply don’t believe that the researchers will adequately represent prostitutes in their sample. And the demographic of prostitutes (streetwalkers, quid pro quo drug addicts) which are most likely to be undersampled are also those who are most likely to have an extreme outlying number of sex partners. Very, very few men will have the number of sex partners that a run of the mill street prostitute has.

Second, someone in the linked thread mentioned that the question asked was “since you’ve been 18 years old”, which would skew the average if the sex excluded across that boundary line isn’t equally distributed between the sexes. And there’s every reason to believe that it’s not. Age differences in sex tend to find the men older and the women younger.

Third, there is the large problem concerning the definition of sex. Before we even consider that men have the social incentive to be more inclusive and women less inclusive for these reporting purposes, there’s the simple matter of sexual violence. Given the high rates of sexual violence against women and the high rates of women who’ve suffered it, especially as compared to men, then it’s easy to see that a number of men will be inclined to view as sex an act that the female partner saw as non-consensual and probably will not report in the count of sex partners.

Third, in the comments of the linked thread there is a quote of a standard used for determining “sex” that is intended to include more than intercourse. That’s all well and good, but the wording then bases the definition upon a specifically enjoyment and sexual arousal. This is problematic, as the commenter argues, because women are more likely to engage in non-aroused sex than are men.

Finally, when we move into the realm of intentional inaccuracies of self-reporting, the implicit assumption among those who argue the explanation hinges upon this is that the distribution of inaccurate self-reporting is even. And it needn’t be, and more to the point, almost certainly isn’t.

The distribution of sexual partners for each sex is very likely not in either case evenly distributed. If there are outlying women with large numbers of sex partners, thus explaining a divergent median, then it’s also reasonable that it’s those very women who are most likely to undereport, thus falsely skewing the average without requiring the majority of women to be underreporting. Furthermore, there’s probably a contingent of outlying men with zero sex partners who are also highly motivated to falsely self-report.

Taken together, all these factors lead me to expect that most respondents will be honest and yet we’ll see these results where even the average is divergent and the median even more so. And that is to say, the results are mostly correct.

Over at Salon, Tracy Clark-Flory contacted Gale about the mean/median issue. He replied:

I’ve gotten several messages making the same point. If you look at Gina’s article you will see that I never attacked the statement about medians. I tried to carefully avoid saying anything directly about the median statement in the article because, as you realize, it could be correct even with accurate data. What I did was to get a copy of the CDC report and used the data in its tables. It groups people into four groups and gives the percentage of men and women in each group:

0-1 partner: Men, 16.6. Women, 25.0.
2-6 partners: Men, 33.8. Women, 44.3.
7-14 partners: Men, 20.7. Women, 21.3.
15 or more partners: Men, 28.9. Women, 9.4.

From these figures you can estimate the total partners claimed by each sex. I got between 40 percent and 75 percent more male than female partners depending on how you guess the average on each interval. Thus, the raw data is inconsistent (so it doesn’t matter whether you take averages or medians or any other statistic). I hope this clarifies.

It is impossible for guys and girls to have a different percentage….
Unless they are prostitutes like the guy stated above and she didnt report correctly. But somewhere someone is lying about how many people they slept with…..

At a recent reunion from my law school class I got to speaking with a woman I had literally not exchanged 1 word with while in school, even though it was a relatively small class and she was pretty attractive. Within 5 minutes she asked me to come home with her.

We ended up sharing a cab with another friend of mine, and they got to talking about their respective “number”. My male friend was quite sure his number was bigger, and was chuckling about it. I was pretty sure he was higher, since he’s lived a number of years in SE Asia, and he has scores with and without prostitutes. When finally called on to give his score, he says, “including prostitutes – 550”. She just smirked, and says “3200”, and then pointed at me, and says “you’ll be 3201”.

Given how quickly she propositioned me, I have little doubt in the accuracy of her number. I also have a pretty good sense that she’s one of the women responsible for tipping the mean number to even…

31: ‘“You often see bright men with dumb women, but you hardly ever see a smart woman with a dumb man.” This suggests that smart women are in short supply, which strikes me as not that unlikely.’

Er, no it doesn’t. But finding a man making such a basic error in the very act of criticising womens’ intelligence is fun, so thank you anyway.

“Even if you’re going to be a reporter, you still need to pay attention in algebra class.”

Why? This is statistics, isn’t it?

Will says that the typical woman probably has fewer partners than the typical man, but I think Will is not imagining the small-of-stature, shy, non-athletic male. A similar female could find a partner (my number is 555-…) if she were not picky. A low-end male is often a pathetic creature, when it comes to mating opportunity. This is true among, bovines, equines, baboons, hunter-gatherers and college boys.

As to the disparity between US and UK males’ success rates, there is a simple explanation. In many parts of the US, population density is very low. Women are scarce. If there is one woman within a day’s drive who is a willing partner, you’re lucky, boys. So especially for males in the vast grasslands and the western mountains, even one partner is something of an accomplishment. You don’t know what it’s like…

What I did was to get a copy of the CDC report and used the data in its tables. It groups people into four groups and gives the percentage of men and women in each group:

From these figures you can estimate the total partners claimed by each sex. I got between 40 percent and 75 percent more male than female partners depending on how you guess the average on each interval.

There you have it. GIGO.

Dilemma resolved.

And I’d also like to call bullshit on this statement in comment #9: since a man would have to persuade many women while a woman need only be easy to persuade

This is exactly the kind of bullshit that flawed survey results are used to justify.

“justify cod-scientific conclusions about male/female gender roles and sad excuses for evolutionary biological ‘reasoning’ about the differences”

Exactly which scientific paper, in which scientific publication, uses these survey results to “reason” about biological differences?

Or maybe it doesn’t whether this entire premise is true, and what’s important is that it serves at the pleasure of “a wider point”?

Bi, either I’m being stupid, or your sentence doesn’t scan. Nevertheless, I meant these kinds of figures, the like of which is fairly frequently bandied about, rather than this particular set of figures. Sorry if that was not clear.

And I’d also like to call bullshit on this statement in comment #9: since a man would have to persuade many women while a woman need only be easy to persuade

OK, you’ve called. Want to raise?

What is your objection?

s/doesn’t whether/doesn’t matter whether/

“I meant these kinds of figures, the like of which is fairly frequently bandied about, rather than this particular set of figures”

… Oh, _duh_. As always, truth needs no facts.

“I see a lot of weaving and dodging and just plain shitting on occam’s razor in this thread because this article upends people’s cozy expectations.”

Where “Occam’s Razor” entails taking one comment #9 and calling it “a lot of weaving and dodging”. And that comment isn’t even from a scientist writing in a scientific publication.

But I don’t see anything suggesting that the survey was made specifically to allow people to conclude silly things about gender roles. Nothing at all. If people want to conclude weird things from the survey results, is it the survey’s fault? Or maybe Occam’s Razor demands that we see an ulterior agenda behind the survey itself even if none’s apparent.