Thursday, December 16, 2010

Women who get around while unmarried get around once married

In continuing to document who is marriage material and who is not, well, life's about trade offs, right? The following table shows the number of men a woman has had sex with since the age of 18 and whether or not she has ever cheated on her husband while married. The table starts at 2, since someone who has been married and is either a virgin or only had sex with one man hasn't cheated (as is always the case, a few respondents or data transcribers muffed up somewhere though, responding that they were married, never had sex, but had still somehow managed to have sex with someone other than a spouse while married):

Men bedded

Not surprisingly, women with high sex drives who got around a lot before they married are more likely to continue getting around after taking their vows. The same applies in non-marital relationships. If your girl has a lot of sexual history (and likes to talk about it), don't go in desiring any kind of serious or long-term relationship. You're in pump and dump territory.

This is another one of those posts where pointing to the blog's tagline is sufficient.

GSS variables used: SEX(2), EVSTRAY(2), NUMMEN(2)(3)(4)(5)(6-10)(11-20)(21-350)


Anonymous said...

Same with men, right? Those that have gotten around a great deal continue to get around after marrying, I'd bet.

Once you've tasted of many springs, your thirst cannot be quenched by one.

I don't think there's anything surprising here. Human nature.

Jokah Macpherson said...

It's interesting how even a single additional partner has a noticeable effect at the low levels.

Anonymous said...

where did you get the statistics could you post the link?


Audacious Epigone said...


The General Social Survey (GSS). Access is free. I have to warn you though, it can be addictive!

Anonymous said...

"Same with men, right?"

Nope. It takes a lot for married guys to wind up. A wife can bang the pizza boy in 2 minutes. The husband has to work on the nanny for months...

Jokah Macpherson said...

"A wife can bang the pizza boy in 2 minutes."

Haha, maybe not the pizza boy, I'd hope she'd shoot a little higher than that. In porn the "pizza boy" is a grown man with gym muscles and tattoos but in reality he's either an awkward long-haired teenager or a non-English speaker.

Anonymous said...

that's a lot of cheating

blackjack said...

All right, I didn't look at the study but you're the one interpreting it so I must ask this question:

How does this prove that women with more sexual partners before marriage are more likely to cheat? As far as I can see (based on your interpretation of the data) this doesn't say anything about how many sexual partners the woman had before she was married. E.g. I'm as likely to interpret this as saying that virgins who got married ended up cheating on their husbands with 28 guys, or women who slept with 50 guys married their husbands (51) and then never cheated again.

Anonymous said...

Note the knee jerk assumption that men are the same as women. I have known men who played around before marriage, and most were faithful husbands. They told me it stopped being fun after while to be sleeping with many women, so they actually sought a marriage with the intention of being faithful to one wife.

Anonymous age 68

Audacious Epigone said...


It's not definitive, but certainly suggestive. If a woman has had a lot of partners and has cheated, perhaps she had almost none before getting married and then went nuts, or conversely a woman who had 30 partners before marriage may have only cheated once after becoming married.

The broader point is the same either way--women who are not selective in who they bed are less likely to be faithful to their husbands than more selective women are.

Sir S said...

This analysis appears to contain heavy residual confounding. The variable you're analyzing is total number of lifetime partners since age 18, right? It is no the number of partners before marriage. Obviously women who have cheated on their husband are more likely to have more lifetime partners.

If the relationship isn't true then a woman with e.g. 6 partners before marriage who isn't going to cheat will always have 7; but a woman who had 6 partners and is going to cheat must have at least 8.

This is a classic experimental design error. Your predictor variable is potentially causally related to your outcome variable, rather than the other way around. You can get around it by calculating the total number of partners minus the number in last 5 years, with some age adjustments. But your results as presented here don't prove your point at all.

Audacious Epigone said...

Sir S,

Good point, it's a quasi-tautology. If we subtract one partner from the cheaters (presuming an average of one cheat, which is probably understated, although not necessarily always so, since one can cheat without actually having sexual intercourse) the strong relationship still holds.

Sir S said...

No, the relationship doesn't hold, or at least you can't say it does without using data you don't have. Consider your table's first entry: 2 men bedded, cheated 10%. We have to deduct TWO from this category to find out how many men the women had before they married (ONE for the husband, ONE for the cheatee). This means women who had 0 partners before marriage cheat 10% of the time.

Then your next row has to be divided between the 0 and 1-partner cell. Outside of their husband, the women have had 2 partners: one of these is a definite cheatee (by definition). The other is either a second cheatee or a pre-marital partner. For the women who cheated twice they have 0 partners before marriage, while the others had 1. We can't decide how to assign these numbers, can we? But we know some proportion of them must be 0 partner women.

Similarly, the 4-partner women get redistributed amongst the 0 - 2 categories, and so on.

If the null hypothesis is correct, then all these women cheat at roughly the same rate; when we reassign the figures, some of the "15%" in the second row get redistributed to the first row, bumping up its rate; some of the "18%" from the third row do; and so on. And given the weak trend here (starting from 10% at 0), even a small redistribution screws your figures. And that's without adjusting for the never-married category, divorces, age of marriage and current age. If older women are more likely to cheat, this will massively confound your results.

If you want to "prove" cheap stereotypes, you need to do a little better than this. You could start with the correct data, and then maybe once you've got past the basic experimental design you could try and handle the problem of confounders.