A letter to the editor to New York Times (unpublished)
In the recent article in
NYT on The
myth the math and sex, Gina Kolata
takes up the well known fact that men
and women report widely different numbers of sex partners although this is mathematically
impossible. This result is true all
over the world. In our nationally
representative survey conducted in
Inspired by Kolata’s article, we checked in our data whether the difference is the same with those who have few partners and those who have many. It is plausible that those who have had few partners remember the number of partners better than those who have had many. But also, those who have had few partners are in a sense “losers”; their potential for producing offspring is lower. Men have a higher pressure than women to have more sexual partners, according to evolutionary theory. The number of partners is directly related to their social position, for instance. On the other hand, men and women who have had many partners are the “winners” and they have perhaps less pressure to change the reported numbers from the real ones, at least when they are men. On the other hand they have more difficulty to remember the precise number and therefore they might round the number off in the preferred direction.
This is coming out from the data in an interesting way. People who have had fewer than 10 partners have only a negligible difference between the numbers of partners reported by men and women. There is no over- or underreporting. Those who report having more than 10 partners account for all the difference: men report having 31.5 partners whereas women report having had only 19 partners. So big a difference cannot be based only on errors in reporting One possible hypothesis is that men and women have a different definition of what having a sex partner means (e.g. as in the case of Bill Clinton), but this is not borne out by the data. The verbal definitions of sexual intercourse given by men and women in our study were identical.
In fact, the exactly equal numbers of sex partners are true for those who have between 10 and 20 partners. Only after 20 lifetime sexual partners (15 % of respondents, 30 % are women and 70 % men) do the numbers diverge. For those above 20 lifetime sexual partners, the difference is very clear: men report 43 partners and women report 30.
So the question remains, who is cheating, men or women? Or maybe the assumption that women with extremely numerous partners (prostitutes) do not participate in surveys is true. This latter hypothesis is doubtful, because male respondents do not report high numbers of sex with prostitutes.
Our hypothesis is that the men have no reason to over- report the numbers of their life-time sex partners, except for occasional rounding off, whereas women have a very strong interest in underreporting. They should not be seen as being promiscuous. Thus, the number for men having more than 10 partners, i.e. 31 partners, is probably true for women also.
We did a further check controlling for education, under the assumption that it is even more important for highly educated, i.e. more powerful and independent women to under-report (the opposite hypothesis is also possible: women in a power position have no need to under- report). The results are somewhat equivocal. The difference between men and women is biggest among those who are highly educated (at least undergraduates). Women with highest education (university degree) report fewest partners, but the women with most partners are those with lower college education, not those with lowest education. In the less than 10 partners group the situation is completely different: university educated women have the highest number of partners and there is no gender difference. Corresponding to our opposite hypothesis above, they have no need to under-report.
In any case, it seems clear that a relatively small minority of men and women are responsible for the difference in misreporting for sex partners and that the number of partners has an important role here. The hypothesis mentioned by Kolata in the end of her article is probably untrue. Sex partners are underreported, not overreported by the group of respondents as a whole.
J P Roos
University of Helsinki
NB: Le Monde reacted to the article also in its issue of Aug 20
and proposed the "best girl of the dance" hypothesis which is
essentially the same as the extremely active prostitutes hypothesis.
Back to beginning