9

In this Tedx talk about sexy mathematics, Dr Clio Cresswell says that there is an inconsistency between womens' and mens' reports of the number of sexual partners they have had.

Now, men report, on average, having had sex with two to four times as many women as women to men. And this does not make sense.

Creswell argues that this results from the difference between the way they count their partners: Men estimate; women enumerate.

She quotes one particularly interesting piece of evidence that this is true: at about 6:06 in the talk she claims that the key piece of evidence is that 80% of mens' estimates of the number of partners is divisible by 5.

Is her hypothesis (that the key explanation of the different reported numbers results from women counting and men estimating) consistent with other research?

Oddthinking
  • 140,378
  • 46
  • 548
  • 638
matt_black
  • 56,186
  • 16
  • 175
  • 373
  • When I have seen sexual partner data in the past beyond a few it normally has the categories divisible by 5. – Loren Pechtel Dec 23 '16 at 22:19
  • Are there really that many men (or women, outside of sex workers) who can't with a little thought remember exactly how many partners they've had? – jamesqf Dec 24 '16 at 03:19
  • @jamesqf It depends what you mean by partner. I'm pretty sure if you define "partner" as someone with which you had a relationship of at least, say, 1 month, everybody with a little thought could count them. If however you simply mean someone you had sex with you have to consider one-night-stands, escorts etc. and people don't generally have important memories of most of these, besides you could have tens or hundreds of these encouters each year... – Bakuriu Dec 24 '16 at 09:47
  • 1
    It is interesting that for heterosexuals, men have more sex than women. So the question is, “who with?” – ctrl-alt-delor Dec 24 '16 at 17:53
  • @Bakuriu: Of course I'm including one-night stands. Certainly it's phyiscally possible for men to have hundreds of one-night stands and/or paid escorts in a year, just as it's possible to win the lottery. But in the real world (or at least the one I live in) how often is that going to happen to someone, unless they are celebrities and/or rich? Perhaps I just have a much better than average memory, but I have no problem remembering all of them - and yes, there were more than I have fingers to count them on :-) – jamesqf Dec 24 '16 at 18:15
  • 6
    I would think a larger factor would be that men are counting women who are not surveyed - sex workers, for example. – Kate Gregory Dec 24 '16 at 19:06
  • 1
    @KateGregory And, again, the nature of the data bins on most such surveys. They have a top bin, the stud that slept with 200 and the prostitute that slept with 10,000 end up in the same bin. – Loren Pechtel Dec 24 '16 at 20:12
  • @LorenPechtel In the particular study in my answer, there were no bins, people could say any number. – DavePhD Dec 24 '16 at 21:05
  • Fair estimating in itself wouldn't explain the difference at all. Men would be just as likely to underestimate as overestimate - with no net effect. – Neil W Dec 25 '16 at 13:14
  • I suspect that the partner count where people switch from counting to estimating is probably about the same for men and women, but that the transition occurs at a higher percentile in women than in men due to the differing distributions. So yes, a higher percentile of women would give a count, and for ranges above this count, the data for men would contain more estimation, and as as result, more of any bias associated with that estimation. – Phil Sweet Dec 25 '16 at 17:05
  • @NeilW estimating is rarely fair. I would not be surprised if a typical male rounded up to the nearest 5 or 10. – matt_black Dec 25 '16 at 19:09
  • 1
    Men are likely to lie by overcounting and women by undercounting. That's an stereotype that could be true for people enough to affect means. – Pere Dec 26 '16 at 23:26
  • @Richard with the minority of women who have lots of partners. – ReasonablySkeptical Dec 27 '16 at 14:03
  • @jamesqf I don't know how many, but there are some. Especially in the generation which hit sexual maturity before AIDS and herpes. – ReasonablySkeptical Dec 27 '16 at 14:05
  • @CPerkins: I think you grossly overestimate the number of opportunities available to the average man (that is, not rich, not celebrity, not overly attractive) in those days. As I said before, while it's possible, it's pretty darned unlikely. – jamesqf Dec 28 '16 at 02:55
  • 1
    @jamesqf It's possible that you're underestimating. I've never been rich or a celebrity, but I lost count a long time ago. One night stands just aren't/weren't significant enough to make note of. Maybe it's a generational thing. I'm in "the generation which hit sexual maturity before AIDS and herpes". – ReasonablySkeptical Dec 28 '16 at 14:23
  • @CPerkins: So am I in that generation, yet I have no trouble remembering each & every one-night stand, even though from the numbers cited by DavePhD the number is in the top 10%. I even remember a good number of encounters that I hoped would become one-night stands (if not more) but didn't. Maybe I just have a much better memory than average? – jamesqf Dec 29 '16 at 05:10
  • @jamesqf Maybe so. Congratulations on your memory. – ReasonablySkeptical Dec 29 '16 at 15:01

1 Answers1

10

There are studies that support what she is saying to some extent, but she is exaggerating.

For example in Estimating number of lifetime sexual partners: Men and women do it differently 1999 (alternative link) Just looking at the top 10% (above the 90th percentile for men and women as one dataset, meaning more than 8 partners) men and women, as plotted in Fig. 1, the men are reporting numbers like 10, 15, 20, 25, 30, 35, 40, 50, 60 and 70, while the women are doing this to a lesser extent.

However, looking at the full data set men have a mean number of 3.79 partners and women 2.54 partners.

And when the top 10% were excluded men averaged 1.51 partners and women 1.53 partners.

So, the part about 80% reporting a number divisible by 5 is inconsistent with this data, as too many report a number less than 5 for this to be true. The 80% figure is only possible if you are looking at a tail-extreme subset like Fig. 1 of the above article.

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
DavePhD
  • 103,432
  • 24
  • 436
  • 464