From The New York Times:

Take the question of promiscuity. Everyone has always assumed — and early research had shown — that women desired fewer sexual partners over a lifetime than men. But in 2003, two behavioral psychologists, Michele G. Alexander and Terri D. Fisher, published the results of a study that used a “bogus pipeline” — a fake lie detector. When asked about actual sexual partners, rather than just theoretical desires, the participants who were not attached to the fake lie detector displayed typical gender differences. Men reported having had more sexual partners than women. But when participants believed that lies about their sexual history would be revealed by the fake lie detector, gender differences in reported sexual partners vanished. In fact, women reported slightly more sexual partners (a mean of 4.4) than did men (a mean of 4.0).

So how sketchy is the research on human sexual behavior, anyway?

New Comment
78 comments, sorted by Click to highlight new comments since: Today at 6:50 PM
Some comments are truncated due to high volume. (⌘F to expand all)Change truncation settings

Everyone has always assumed — and early research had shown — that women desired fewer sexual partners over a lifetime than men.

Of course, the study reports how many sexual partners participants said they had had, not how many they wanted. Since it was filtered for an all-heterosexual population, the fact that men and women have about the same number of sexual partners (duh) doesn't tell us anything about whether one sex wants more than the other.

Link to the original study:

As far as to the sample they studied:

An initial sample of 248 male and female undergraduates at a regional campus of a Midwestern university participated as partial fulfillment of a research requirement in their Introductory Psychology course. To keep the sample somewhat homogenous, we used only data from unmarried, heterosexual, 18- to 25-year-old participants.

PDF [].
Thanks. So it's a rather small and unrepresentative sample for men and women in general, but the study does supply a clue about whether people answer scientific study questions accurately. Of course, we still don't know about their actual sex lives, and we don't know how many decided that it was worth bluffing the "lie detector". Edited to add: I'd expect college students to be more likely to have heard that even real lie detectors aren't accurate. I could be wrong about that....
According to the study: Basically the thought of being embarrsed by getting caught lying feels worse then the thought of admitting a high number of sexual partners. You don't need a lie detector that's 100% accurate to trigger that effect.
Sure, but "more likely" here means a base rate of maybe 1% instead of a base rate of maybe 0.1%. Either way, the fact that they even mentioned lie detection is likely to scare people into telling the truth whether or not they believe (as opposed to alieve) that lie detectors work.
Good point about alief vs. belief about lie detectors.
My guess would be an order of magnitude larger than that (but still much less than 50%).
They ask you at the end if you thought the experiment was real or if you thought the researchers fibbed. Then they exclude the ones who weren't fooled.

Related past LW discussion: 2011; 2009.

On a lighter side, this study reinforces (by a small quantity, due to all the caveat outlined in the comments) my idea that women are as promiscuous as men, but they are culturally forced to lie about that: not really big news.

On a more interesting side, the "fake lie detector" is another one of the techniques that are used to circumvent lies that occur even in anonymous surveys: the first that I heard of, anyway, was employed in a survey regarding illegal owning/hunting/farming of something in some parts of Africa (yes, I've lost almost all the details: can someone point me to the original article?). It consisted of telling people that for some answers, you need not to answer truthfully, instead you needed to throw secretly a dice and report the answer that came up. Apparently this, instead of randomizing the answer, gave the 'farmer' an excuse to tell the truth (yes, I really need to dig up the source).

ETA: see Alicorn's comment for the exact reference.

Are you thinking of this?

Thank you Alicorn, it was exactly that (although I learned of its existence from a different source than Lesswrong).
Keep in mind that this study only reflects upon individuals born between 1978-1985. Based on the recent increase in entertainment promoting promiscuous behavior (ie. the American Pie series, EuroTrip, ), I expect that current attitudes (of 18-25 y/o's) would differ, even from those in 2003.
Well, the study has its limits, and this could be another one of them, so as I said it's far from conclusive. However, in this particular case, I wonder if the kind of entertainment that you indicate really nudges young adults sexuality or limits to expose a widely known but denied truth (in Italy we say "Punchinello's secret", also when someone declares that he has discovered something that everybody else knew very well in advance, it's said that he has "discovered hot water"). I'm thinking for example of pornographic movies, that exist since the beginning of cinematography, and it's not that the women were particularly forced to star in them... Of course not even this is strong evidence, though it's still something. In my opinion, the biggest piece of evidence in favor of women's promiscuity are sperm wars.
Also, I think there are random factors affecting which art gets made. What gets popular is a subset of what happens to get made, not a clear indicator of the minds in the audience.
What do you mean by "equally promiscuous"? Do you mean that women and men have the same average number of sex partners? Than this follows from basic arithmetic and I don't see why this study is relevant. Do you mean that they have the same average desired number of partners? Than this study has nothing to say on the subject.
Yes, of course. Why everybody keep saying this? It's true only if the group surveyed is closed under the "sexual partnership" relation, which is hardly the case in any study. An extreme example: a population of 10 men and 10 women, in which all the women have sex with just one man. Then in any group not closed under sexual partnerhsip the average for women is always 1 and for men is always 0, correctly indicating that women are more promiscuous than men. The average is trivially equal only if the group is extended to the whole population, but in that case the average is not a good indicator of promiscuity (see the above example). The study is indeed interesting because there are bound to be asymmetries, and it shows that they are skewed towards women.
Unless the group surveyed is deliberately gerrymandered for that, I doubt that would make for a very large difference. If the difference is 10%, as in this study, I have no trouble believing that, especially considering this [] (though there probably are other sources of noise); but if someone finds nothing obviously wrong with the studies without a fake lie detector where the difference is a factor of 2, and when the arithmetic argument is pointed out to them they say that maybe the group surveyed wasn't closed under sexual partnership, I would call that clutching at straws.
And yet the previous studies found otherwise, and were interpreted as confirming that men are more promiscuous than women.
Here it is. []
Perhaps KZN rangers could undergo polygraph tests [] ?

Surely the mean number of sexual partners must be equal for both genders, unless one is numerically larger?

Edit: Or is homosexual sex is responsible for the whole difference, I suppose.

No, the study is run on heterosexual undergraduates between 18-25 years of age.

Women are more likely to have sex with older guys than guys having sex with older women.

What's omitted is that previous research has shown ridiculous differences, including counting strictly heterosexual partnerships. The scientists in question generally know the number is wrong, and have some rather entertaining comments about it, but reporters frequently -don't-, and treat it like it's accurate.

[-][anonymous]10y 10

but reporters frequently -don't-, and treat it like it's accurate

That kind of stuff happens with everything. I no longer consider reporters any more reliable than word-of-mouth.

This is pretty much the default state of everybody who has read one or two media reports of events they've been directly involved in.
One would think so, but this hasn't been my experience with people.
This is interesting and suggests that I and the people I talk to about such things are unusual. Do you mean people don't extrapolate from the inaccuracy of such reports, or to they find them to be not too inaccurate?
The people I have in mind say something along the lines of "But the newspaper reports are the only information we have, therefore we have no choice but to believe them".
Mmm, yeah, but even when word-of-mouth is the only information we have, we “have no choice but to believe them” in that sense (assuming they mean something sensible by that phrase in that context).
That's why you would want to use the median rather than the arithmetic mean when making that comparison.
There may also be people who are still willing to lie when tied up to a lie detector, or people who realised it was fake,

I wonder to what extent this was caused by people overcompensating for the lie detector.

How is being dishonest in the other direction going to help you pass a lie detector? I suspect I missed the point of this comment.
Given the way memory works, people probably remember only the number they want to believe and a vague idea of which way they're biasing it. So when presented with a "lie detector" they try to compensate for their bias and wind up overcompensating.

Not nearly as sketchy as the reporting on it. I've seen too many newspaper articles treating claims that strictly heterosexual men have more sexual partners on average than strictly heterosexual women on average seriously, to take -any- reporting on the matter seriously. (Hint for those puzzling about what's wrong with this: The two numbers have to match. ETA: As Luke_A_Somers points out, the numbers don't have to be identical, and within particular subsets of the population it's possible for large disparities.)

One other case where the numbers may not match: differences in age.

For example imagine a population where everybody lives for 60 years, men always marry at 40 where they have sex once, and women marry at 20 where they also have sex once (assume they always have twins). Women over 20 will have had on average one partner, and men over 20 will have had 0.5.

So, the numbers don't need to match if you have more pairings where the man is older than the woman than the other way around (which seems to be the case); I expect that to account for some of the difference, though I expect exaggeration and flexible interpretations of what is meant by "sex" to account for a bigger chunk.

In most such studies, women report drastically lower sex partnerships than men (2 to 12 partners, for one study I recall from memory and may be dramatically misremembering so don't take my numbers too seriously). This study is actually interesting in that the numbers of the lie-detected class of respondents are pretty close.

The numbers don't need to match, even if everyone was counted and reported accurately.

To take some extreme examples:

The population is Alex, Betty, Carrie, Daphne. Alex hooks up with each of the others. Mens' average: 3. Womens' average: 1.

Now take Alice, Bob, and Cindy. Alice hooks up with Bob and then leaves the population (emigrates or dies) Womens' average: 0. Mens' average: 1

Granted. But such distortions aren't exactly prevalent in society, which has a near 1:1 gender ratio, and relatively insignificant departure rates (they're going to be less than the error margin in any honestly-conducted study).
Older men are consistently attracted to younger women. Consider implications.
Interesting. That would work in the opposite direction from the two effects I was thinking of, and it could well be stronger than either of those effects. Anyway, it strengthens the point that the numbers don't need to match.
In the Cracked article I first saw this in (can't say if OrphanWilde got it from elsewhere) they admitted that there could be a gender ratio imbalance (they also pointed out that homosexuals exist) but noted that the evidence was being interpreted strictly as evidence for promiscuous males. (And ... whatever the opposite of promiscuous is ... females.)
Chaste. (Traditionally, chastity includes both celibacy outside of marriage and fidelity within.)
That objections disappears if you assume an equal gender ratio in the population at large, which is usually the case.
There is a slight unevenness in population, and it tends to favor the females (to the tune of 0.5% in the US), so effect 1, though exaggerated above, does exist. The second objection is not even marginal - indeed, it is very pertinent, especially when one considers the long tail of female sexuality. Most women will have few partners, and then you get the 1% who have over 100 times as many. In particular, streetwalkers: they are the tippy tip end of that tail, and they face greater dangers than other women - largely from the effects of criminalization, but partly from the factors that put them in such a circumstances in the first place (since streetwalking is one of the least attractive sex work options, let alone work options, and has a very low barrier to entry, an elevated fraction of them will be in dire straits to begin with). Similarly, highly promiscuous amateurs have much higher STD rates than usual. Since transmission from male to female is much higher rate than female to male, the women will bear the brunt of that too. So no, it doesn't actually disappear.
No, it's still a legitimate objection to my statement that the numbers -have- to match. You can still have differing departure rates.
Always try to say yes rather than no :)
Even in this case, you could have a minority of men who have sex with 100 women very quickly, then die, surrounded by other men who have sex with either 0 or one woman but live for a long time. Women in this scenario all live a long time. Then the average for men will be lower than the average for women.
Actually, this seems to fit some subset of criminals: make some bold crimes, have a lot of money, be attractive (courageous and rich), have as much sex as possible, and die young (during the next crime, or killed by competition).
It could be that people count different things as sex. Intercourse always counts, but any other act that can cause orgasm (and quite a few that can't) are counted as sex by some people and not by some others. Can't see why men would have a wider view of what counts than women, though; I'd expect the reverse.
Isn't the usual explanation that "having had a lot of partners" is seen as a positive trait for a man, but a negative one for a woman? (or at least, it's a more desirable/less undesirable trait for a male).
Yeah, but you don't need to postulate different definitions of sex then; they're fudging the numbers, and nothing tells you whether they use this parameter to do it. I was proposing a model where people are completely honest and still give different answers. I read in Cosmo (probably the least reliable source possible) that women often don't count one-night-stands. No idea how to test.
Testing is pretty easy. You split your group into two. One group get's asked: How many sexual partners did you have in your life? The other group get's asked: "How many sexual partners did you have in your life, if you don't count one-night-stands?"

Three. Third group gets "counting one-night-stands".

Different definitions of sex sound like a plausible way of fudging the numbers (I don't see any other ways of fudging the numbers that aren't outright lying), and fudging the numbers seems like a good motivation to have different definitions of sex. The lie detector in the OP makes outright lying seem a bit more plausible though, if it's true.
Obviously that's the main implication, and it's very likely a significant cause of the difference. It's not necessarily the only one - and if we're trying to measure that difference, we'll need to take into account the other causes of differences.
In a How I Met Your Mother episode, Znefunyy pbhagrq jung Yvyl qvq jvgu ure rk-oblsevraq nf frk naq Yvyl qvqa'g. (Znefunyy punatrq uvf zvaq ol gur raq bs gur rcvfbqr.) (Yeah, fictional evidence, I know...)
I am totally willing to believe that men and women count different things as sex, but they could both be using the same standard - "times a partner makes me orgasm" - and it could still result in a greater mean number of acts for men than for women even in a fixed heterosexual population (unfortunately)... granted I have no idea if people do this.
Society sees a man who has sex with a lot of women as being good with woman. Society sees a woman who has sex with a lot of men as being a slut. A woman doesn't want to see herself as a slut and might therefore use a more conservative way of counting.
Like what?
Nearly every kind of BDSM play under the sun, and talking dirty. Most people can't orgasm from those alone.
Wait... People count that as sex? I knew phrases like “cyber sex”, “phone sex”, etc., but I thought people thought of them as cases where [adjective] [noun]s are not actually [noun]s, as with “dwarf planet”, “fake money”, “soy meat”, etc., and I'm surprised if there are people who would count that as sex when asked how many people they've had sex with (except possibly as self-sophistry in attempt to fool the lie detector).
Lots [] of [] people [] count [] "helping a partner come by using words" (typically by one or both partners masturbating at the same time) as sex; it actually seems to be the default position among sex bloggers. That sounds solid to me; if Alice is doing something to pleasure Bob, it doesn't particularly matter how indirectly. Fewer people (in the list above, Cliff Pervocracy and Greta Christina) count as sex the dirty talk alone, without the possibility of orgasm. I can't think of a situation where that would feel like sex to me, but it appears to work for them. Naive attempts to dissolve the question of "what is sex?" tend to fail because of the assumption that describing the acts is enough, when the psychology is a large part of what the question is getting at. For example, I think that two people masturbating to each other's sight is more like two people masturbating each other than it is like a circle jerk.
Ontopic: most people count it as cheating, IIRC, but then some people count almost anything as cheating. Offtopic: does anyone know if that actually works? I know I'm generally aware of when I'm deliberately manipulating the truth; indeed, it's actually easier (in my experience and according to conversations with a friend) to loose track of when you're telling them something you actually made up, as if it were the truth (I assume this would help with most forms of lie-detection.) OTOH, I know nothing about lie detection of any kind.
You mean “most people count [dirty talk] as cheating [when you're in a relationship with someone else]”? Or “most people count [self-sophistry in attempt to fool lie detectors] as cheating [when taking surveys]”? If the former, yes, but then most people would count French kissing as cheating, too, whereas virtually no-one would count is as sex. If the latter... I don't think such people give a damn whether they are cheating, or if they did they'd also answer honestly when there's no fake lie detector. AFAIK lie detectors measure the effect of emotional responses (heart rate, sweating, etc.), so to a zeroth approximation you'll get caught iff you alieve you'll get caught. So, it'd depend on how the respondents alieve lie detectors work. IME (though these anecdotes date back to my childhood, so my recollection may be unreliable), when Muggle adults see a toy that purports to be a lie detector (e.g. an analog thermometer like a mood ring []), they start to say statements whose truth value they couldn't possibly know themselves (e.g. “Juventus will defeat Manchester Utd tomorrow night”) and purport that they will bet based on the outcome. Which means that either 1) they are idiots, or 2) they realize that the widget cannot possibly detect lies, and sarcastically reduce the claim ad absurdum. If it's 1), such people may think that lie detectors somehow respond to the actual words you say rather than your state of mind, and so that they wouldn't detect a statement that is denotationally true but connotationally false. You mean “lose track of what the truth is”, or “lose track of what you told them”? I'd guess the former (or that there's a missing “than” somewhere); the latter would contradict my experience, but then AFAIK I'm worse at (and reluctant to) lying than most other people.
Whoops! Overlooked the double meaning of "cheating". I meant it regarding the sex stuff, not the lie stuff. Not that's an idea I hadn't heard before. I've always seen it framed as "people are nervous when they lie." I'm pretty sure I've made that same joke, as essentially a pun on the two meanings of "truth". Still ...
The matter could be complicated by female prostitutes with male customers, so that the amount of sex that each gender wants because of sexual desire is quite different. On the other hand, I don't even have an intuition about what proportion of sex is with prostitutes.
Indeed, psychologists generated this hypothesis and tried to make a survey not biased against prostitutes. Their survey found no imbalance. But it cannot both be true that this study and the lie detector study both explain the discrepancy. It could be that both effects are true, but there are unaccounted biases in the other direction (such as age). The fact that both studies claim the same size effect, the size of the error that they are trying to explain, makes me highly suspicious of the claimed magnitudes. The psychologists got the answers they wanted. How sketchy is this research? Par for the psychology course.
"Because of sexual desire" would perhaps be more accurate. The internal drive for acquiring resources is rather ubiquitous and powerful. It plays a significant part in certain long term mating strategies as well as prostitution and is also not all that different than the drives for status affiliations that have such a strong influence on other sexual and social behaviors. The cases where a distinction based on lack of internal drive comes clearly into play are with such things as sex slaves and rape.
I'm correcting the comment-- you're got a point.
If you're counting unique partnerships rather than acts, around 30-50%. Study here []; analysis of some of the flaws by a retired whore here []. There is, unfortunately, a lot of educated-guesswork involved, criminalization being what it is. NSFW, obviously. Well, those specific links are probably okay, but the site as a whole isn't unless you have a very understanding boss.
Considering that most "unique partnerships" with prostitutes are likely to be, well, shorter than average ...
Other extreme outliers could also skew the data; cultural stereotypes say that it's easier for women to be an extremely promiscuous outlier regardless of whether they're doing so by being a prostitute.
Well, assuming equal populations, anyway.
Luke pointed out some other potential divergence points, such as departure rates (death/emigration). There's also the potential issue that the subset queried may be divergent from the population as a whole; as a non-descriptive example, male college students may have a lot of sex with female non-college students, and female college students strictly have sex with male college students; if you're querying college students, you could get largely divergent numbers. But if your numbers are largely divergent, they're probably wrong.
[-][anonymous]10y 0


"America can do some stupid things sometimes, but we would never elect a Stalin, a Pol Pot, or a Kim Jong-Il"

Odd here you are implicitly arguing against Monarchy, by saying Hitler is the worst thing that can happen (you really should think about that for a second or two more). But at least we won't get a Stalin, Pol Pot or Kin Jong-Il! Oh wait those aren't monarchs. Can you say with a straight face that Louis XVI or Nicholas II where anywhere near as bad as those?

Oh and speaking of how democracy never picks leaders terrible for the majority...... (read more)

[This comment is no longer endorsed by its author]Reply

New to LessWrong?