Over at Certain Doubts, Gregory Wheeler reviews our lies about sex: In survey after survey within country after country men report having more heterosexual partners over their lifetime than women do, and as this article and this clarification point out, what people say in these surveys cannot be a reflection of what they do.
Bob Knaus, I don't think those numbers reflected academics, it is my understanding they are taken from the population of the country as a whole (correct me if I'm wrong). Those numbers don't seem low at all to me and I know some pretty promiscuous people. I would say if you've reached triple digits you are probably a very very very very very very very very far outlier.
Another potential loophole - men and women could be using different definitions of what counts as a sexual partner.David Buss considers these and other potential explanations for the sex difference in reported numbers in the revised edition of The Evolution of Desire but concludes that "the real reason" of the discrepancies is the misrepresentation of prostitutes in sex surveys (p. 259).
Doug S, yes, a study I ignored while I was finding the one that looked so good claimed that the women in their sample figured that it only counted when they could remember the man's name.
Another potential loophole - men and women could be using different definitions of what counts as a sexual partner. If, for example, men considered act X sufficient for counting someone as a sexual partner and women did not, then that could also cause differences in results. If two people can disagree on whether they were partners, then the High School Prom Theorem doesn't hold even if everyone answers honestly and the sampling rate is 100%.
My conclusion: academics are not getting enough sex.
The number of lifetime partners (for both genders) reported in Robin's original post seems low to me based on personal experience and what close friends have told me. If those numbers are accepted as reasonable by the commenters on this board... then I can only conclude that they need to get out a bit more.
In my experience, there are a few women who are in fact statistical outliers, and the ones I know are not prostitutes. You should call them escorts, by the way. The one friend who finally did progress from trophy wife to stripper to escort (in her 40s!) had passed the triple-digit number of sex partners a decade earlier. She's a welfare mom now, which provides some kind of moral, though I can't think exactly what.
"I'll accept that lying prostitutes may explain the difference."
It isn't just prostitutes that lie. This particular study suggested that prostitutes in Colorado Springs tended to have about a 5 year career, and during that time they would tend to be underrepresented in polls for various reasons. For example, if you poll by residence, the prostitutes tended to live in places that wouldn't get polled.
With an incidence of less than 3 in ten thousand, an average 5 year career would increase that to only about 3 in a thousand. Not that many studies include ten thousand women.
And the prostitutes on crack brought up the average considerably, and they were much rarer, and also have a much shorter life expectancy. So they would be severely underrepresented in polls.
This reminds me of an episode of I believe it was Coupling (a British sit-com similar to, but better than Friends in my opinion) where one of the characters brings this up and says that when you ask a man how many people he's slept with you always have to subtract 2 or 3 from their number and when you ask a woman how many people she's slept with you always have to add 2 or 3 from their number. I would suspect that in surveys and studies that ask that sort of question you'll find much of the difference can be explained by the fact that men like to appear like they're studs and women like to appear as if they're not sluts because there exists an odd double standard in many cultures (America especially). I would be more interested to see the result of surveys like this in non-western cultures, like some remote native tribe in South America. Perhaps there are some places where that cultural double standard isn't as strong, is non-existent, or even the reverse. That might lead to less of a reason to lie when asked how much you get around.
Funny thing though, I myself (a man) often lower the number if asked by most people, however I have met girls who are probably lying when they say they've slept with absurdly large numbers of men (25+ !!). Guess not everyone follows the typical pattern.
The other thing to note is that you might not be able to leave out the self-identified liars, because they might be a skewed sample. As in, assume the Lady X hypothesis (or prostitution) and then suppose that Lady X is more likely to lie than anyone else in the survey. This will skew the numbers. If she self-identifies as lying and you throw her out, this will probably skew the numbers even more.
Now suppose, as is plausible, that guys with few partners are more likely to lie than guys with many, and women with many are more likely to lie than women with few; then throwing out the self-identified liars could actually make the discrepancy worse, not better (depending on by how much they lie).
I'll accept that lying prostitutes may explain the difference. I guess I didn't realize how common it was; maybe I'm missing out on a good thing and should look into it? :)
Thomas: "I'm a little doubtful. It was published in october 2000. If it actually solved the problem, would people have debated it online more or less continually for 7 years longer?"
I don't think this line of reasoning works for determining the quality of a finding. There have been pretty convincing evidence for evolution for many years now, yet people debate it fiercely online anyway. I guess one can refine it a bit by discounting the value of debate if there are emotional or other reasons to assume people are unlikely to want to change their minds or reach a consensus. But sex partner issues are likely in that category, so we cannot apply the debate ending heuristic here anyway.
But it would be interesting to analyse this more in detail. Just how much information about the quality of a finding can we infer from the existence or nonexistence of a debate several years afterwards?
estimate 1 Swedish study estimates 50 clients/prostitute/year
estimate 2 1985 west german study estimates licensed prostitutes at 13 clients/week.
estimate 3 2006 Croatia study estimates 5-6 per day.
The best study I found American analysis of previous data. About 2% of american women said they'd ever sold sex. Estimate of 23 per 100,000 full-time equivalent prostitutes in the US population during 1970-1988, assuming Colorado Springs is typical of the USA. Studies at less typical sites give higher numbers.
"98 adult prostitute women reported a mean of 347 male sexual partners in the last 6 months (median = 103; interquartile range = 11–228; range = 1–5,401). The four prostitutes with the most partners (2,700–5,401) in this period reported very heavy cocaine/crack use. Their high level of reported activity is consistent with field observations of crack-addicted prostitutes (17). We doubled the 6-month mean to obtain an estimated mean of 694 male partners in the last 12 months for these women.
"This doubled figure is consistent with an estimate derived from prostitutes' reported number of male partners in the last 5 years. For the latter estimate, we assumed that the rates of entry into and exit from prostitution were equal for the Colorado Springs cohort of prostitutes, which implies that these women were, on average, halfway through their prostitution careers (18). Because prostitute women in Colorado Springs have a mean career length of 5 years (15), prostitute women in this sample most likely worked as prostitutes for only 2.5 of the last 5 years, on average. Therefore, currently active adult prostitutes' (n = 98) mean of 2,171 reported male partners in the last 5 years corresponds to an estimated mean of 868 male partners per year (2,171 partners per 2.5 years) for working adult prostitutes. We opted for the more conservative estimate based on doubling the 6-month mean."
The authors claim that prostitutes at these rates entirely account for the difference in reported numbers of sex partners between american men and american women. I didn't look at it carefully enough to make sure there were no flaws, but it looks plausible.
I'm a little doubtful. It was published in october 2000. If it actually solved the problem, would people have debated it online more or less continually for 7 years longer?
Robin Hanson postulates a reporting bias. Recent research seems instead to support a sampling bias. It appears that men report substantially more sexual partners than women do, not because members of either sex consciously or unconsciously lie about the number of partners they've had, but because female prostitutes are underrepresented in the surveys which register those reports. This is what Devon D. Brewer and his collaborators claim in "Prostitution and the Sex Discrepancy in Reported Number of Sexual Partners" (Proc Natl Acad Sci U S A 2000 October 24; 97(22): 12385–12388). Here's the abstract:
One of the most reliable and perplexing findings from surveys of sexual behavior is that men report substantially more sexual partners than women do. We use data from national sex surveys and studies of prostitutes and their clients in the United States to examine sampling bias as an explanation for this disparity. We find that prostitute women are underrepresented in the national surveys. Once their undersampling and very high numbers of sexual partners are factored in, the discrepancy disappears. Prostitution's role in the discrepancy is not readily apparent because men are reluctant to acknowledge that their reported partners include prostitutes.
Ok, clearly a proposal like "Lady X" doesn't change the averages (once more: 99 males with 2 partners one being Lady X, 1 male with Lady X as partner; 99 females each with 1 partner, Lady X with 100, resulting male median 2, average=1.99 where female median=1 but average=1.99). However, it does suggest that only Lady Xs need to lie to generate a discrepancy in averages, whereas a lot of males would have to do so to generate an equivalent effect. It seems plausible off-hand that Lady X has a strong motivation to under-count her partners, and that some males will have a strong motivation to over-count theirs. Has somebody proposed a test which distinguishes between (or even finds the overlap between) these motivations as causes? Or one which rules them out?(Of course the subsample size of Lady Xs would be much smaller than the overall sample size.)
But no one is counting what you do by yourself!
Bob Knaus, I don't think those numbers reflected academics, it is my understanding they are taken from the population of the country as a whole (correct me if I'm wrong). Those numbers don't seem low at all to me and I know some pretty promiscuous people. I would say if you've reached triple digits you are probably a very very very very very very very very far outlier.
misrepresentation => underrepresentation
Doug S wrote:
Another potential loophole - men and women could be using different definitions of what counts as a sexual partner.David Buss considers these and other potential explanations for the sex difference in reported numbers in the revised edition of The Evolution of Desire but concludes that "the real reason" of the discrepancies is the misrepresentation of prostitutes in sex surveys (p. 259).
Doug S, yes, a study I ignored while I was finding the one that looked so good claimed that the women in their sample figured that it only counted when they could remember the man's name.
Another potential loophole - men and women could be using different definitions of what counts as a sexual partner. If, for example, men considered act X sufficient for counting someone as a sexual partner and women did not, then that could also cause differences in results. If two people can disagree on whether they were partners, then the High School Prom Theorem doesn't hold even if everyone answers honestly and the sampling rate is 100%.
My conclusion: academics are not getting enough sex.
The number of lifetime partners (for both genders) reported in Robin's original post seems low to me based on personal experience and what close friends have told me. If those numbers are accepted as reasonable by the commenters on this board... then I can only conclude that they need to get out a bit more.
In my experience, there are a few women who are in fact statistical outliers, and the ones I know are not prostitutes. You should call them escorts, by the way. The one friend who finally did progress from trophy wife to stripper to escort (in her 40s!) had passed the triple-digit number of sex partners a decade earlier. She's a welfare mom now, which provides some kind of moral, though I can't think exactly what.
"I'll accept that lying prostitutes may explain the difference."
It isn't just prostitutes that lie. This particular study suggested that prostitutes in Colorado Springs tended to have about a 5 year career, and during that time they would tend to be underrepresented in polls for various reasons. For example, if you poll by residence, the prostitutes tended to live in places that wouldn't get polled.
With an incidence of less than 3 in ten thousand, an average 5 year career would increase that to only about 3 in a thousand. Not that many studies include ten thousand women.
And the prostitutes on crack brought up the average considerably, and they were much rarer, and also have a much shorter life expectancy. So they would be severely underrepresented in polls.
This reminds me of an episode of I believe it was Coupling (a British sit-com similar to, but better than Friends in my opinion) where one of the characters brings this up and says that when you ask a man how many people he's slept with you always have to subtract 2 or 3 from their number and when you ask a woman how many people she's slept with you always have to add 2 or 3 from their number. I would suspect that in surveys and studies that ask that sort of question you'll find much of the difference can be explained by the fact that men like to appear like they're studs and women like to appear as if they're not sluts because there exists an odd double standard in many cultures (America especially). I would be more interested to see the result of surveys like this in non-western cultures, like some remote native tribe in South America. Perhaps there are some places where that cultural double standard isn't as strong, is non-existent, or even the reverse. That might lead to less of a reason to lie when asked how much you get around.
Funny thing though, I myself (a man) often lower the number if asked by most people, however I have met girls who are probably lying when they say they've slept with absurdly large numbers of men (25+ !!). Guess not everyone follows the typical pattern.
The other thing to note is that you might not be able to leave out the self-identified liars, because they might be a skewed sample. As in, assume the Lady X hypothesis (or prostitution) and then suppose that Lady X is more likely to lie than anyone else in the survey. This will skew the numbers. If she self-identifies as lying and you throw her out, this will probably skew the numbers even more.
Now suppose, as is plausible, that guys with few partners are more likely to lie than guys with many, and women with many are more likely to lie than women with few; then throwing out the self-identified liars could actually make the discrepancy worse, not better (depending on by how much they lie).
I'll accept that lying prostitutes may explain the difference. I guess I didn't realize how common it was; maybe I'm missing out on a good thing and should look into it? :)
Thomas: "I'm a little doubtful. It was published in october 2000. If it actually solved the problem, would people have debated it online more or less continually for 7 years longer?"
I don't think this line of reasoning works for determining the quality of a finding. There have been pretty convincing evidence for evolution for many years now, yet people debate it fiercely online anyway. I guess one can refine it a bit by discounting the value of debate if there are emotional or other reasons to assume people are unlikely to want to change their minds or reach a consensus. But sex partner issues are likely in that category, so we cannot apply the debate ending heuristic here anyway.
But it would be interesting to analyse this more in detail. Just how much information about the quality of a finding can we infer from the existence or nonexistence of a debate several years afterwards?
estimate 1 Swedish study estimates 50 clients/prostitute/year
estimate 2 1985 west german study estimates licensed prostitutes at 13 clients/week.
estimate 3 2006 Croatia study estimates 5-6 per day.
The best study I found American analysis of previous data. About 2% of american women said they'd ever sold sex. Estimate of 23 per 100,000 full-time equivalent prostitutes in the US population during 1970-1988, assuming Colorado Springs is typical of the USA. Studies at less typical sites give higher numbers.
"98 adult prostitute women reported a mean of 347 male sexual partners in the last 6 months (median = 103; interquartile range = 11–228; range = 1–5,401). The four prostitutes with the most partners (2,700–5,401) in this period reported very heavy cocaine/crack use. Their high level of reported activity is consistent with field observations of crack-addicted prostitutes (17). We doubled the 6-month mean to obtain an estimated mean of 694 male partners in the last 12 months for these women.
"This doubled figure is consistent with an estimate derived from prostitutes' reported number of male partners in the last 5 years. For the latter estimate, we assumed that the rates of entry into and exit from prostitution were equal for the Colorado Springs cohort of prostitutes, which implies that these women were, on average, halfway through their prostitution careers (18). Because prostitute women in Colorado Springs have a mean career length of 5 years (15), prostitute women in this sample most likely worked as prostitutes for only 2.5 of the last 5 years, on average. Therefore, currently active adult prostitutes' (n = 98) mean of 2,171 reported male partners in the last 5 years corresponds to an estimated mean of 868 male partners per year (2,171 partners per 2.5 years) for working adult prostitutes. We opted for the more conservative estimate based on doubling the 6-month mean."
The authors claim that prostitutes at these rates entirely account for the difference in reported numbers of sex partners between american men and american women. I didn't look at it carefully enough to make sure there were no flaws, but it looks plausible.
I'm a little doubtful. It was published in october 2000. If it actually solved the problem, would people have debated it online more or less continually for 7 years longer?
Robin Hanson postulates a reporting bias. Recent research seems instead to support a sampling bias. It appears that men report substantially more sexual partners than women do, not because members of either sex consciously or unconsciously lie about the number of partners they've had, but because female prostitutes are underrepresented in the surveys which register those reports. This is what Devon D. Brewer and his collaborators claim in "Prostitution and the Sex Discrepancy in Reported Number of Sexual Partners" (Proc Natl Acad Sci U S A 2000 October 24; 97(22): 12385–12388). Here's the abstract:
One of the most reliable and perplexing findings from surveys of sexual behavior is that men report substantially more sexual partners than women do. We use data from national sex surveys and studies of prostitutes and their clients in the United States to examine sampling bias as an explanation for this disparity. We find that prostitute women are underrepresented in the national surveys. Once their undersampling and very high numbers of sexual partners are factored in, the discrepancy disappears. Prostitution's role in the discrepancy is not readily apparent because men are reluctant to acknowledge that their reported partners include prostitutes.
(Of course the subsample size of Lady Xs would be much smaller than the overall sample size.)
Which means you could get the same effect even if all Lady Xs were honest, if they're rare enough and the sample is small enough.
Ok, clearly a proposal like "Lady X" doesn't change the averages (once more: 99 males with 2 partners one being Lady X, 1 male with Lady X as partner; 99 females each with 1 partner, Lady X with 100, resulting male median 2, average=1.99 where female median=1 but average=1.99). However, it does suggest that only Lady Xs need to lie to generate a discrepancy in averages, whereas a lot of males would have to do so to generate an equivalent effect. It seems plausible off-hand that Lady X has a strong motivation to under-count her partners, and that some males will have a strong motivation to over-count theirs. Has somebody proposed a test which distinguishes between (or even finds the overlap between) these motivations as causes? Or one which rules them out?(Of course the subsample size of Lady Xs would be much smaller than the overall sample size.)