Tuesday, February 03, 2009

What types of people cheat on their spouses?

++Addition++An error in the original posting has been updated in response to the astuteness of commenter DK, who pointed out that the figures for male median income were far too high to be plausible. They have been corrected. Also, please see Blode's comment challenging the assertion that thin women, as measured by GSS interviewers, are likely the most attractive.


Reading one of Agnostic's recent posts on the subject of female sexual activity by age led me to wonder what kinds of people cheat. Fast-paced thrillers designed to make the reader feel like an insider (I'm thinking Grisham or Crichton here) tend to portray the high-powered elite types in their forties and fifties and the aspiring members of the next generation in their twenties and thirties who are ready to replace them always running wild, with passionate but shallow mistresses (established elites) or one-night tangles (ambitious up-and-comers). My impression has been that those presentations are American Beauty-esque, portraying the upper-middle and solidly middle classes engaging in behaviors that more accurately characterize the working- and underclasses, especially the urban underclass.

The GSS suggests cheaters are pretty evenly distributed across social category markers. One shortcoming in using GSS data is that it only asks those who are married or have previously been married whether or not they have ever cheated on a spouse. Since those who marry are on average more intelligent and affluent than those who do not, much of the underclass is left out. What should we assume about the 70% of unmarried new black mothers with regards to fidelity? Commitment levels and expectations are a world apart here relative to the burgher world, so it's an apples and oranges comparison, although my experience* tells me faithfulness is more fleeting among the underclass.

In any case, interviewers are asked to report on the respondent's weight. The percentages for each weight class who report to have had sex with someone other than a spouse while married:

Men (N = 659)
Slim -- 29.2%
Average -- 20.1%
Stocky -- 25.7%
Fat -- 30.4%

Women (N = 843)
Slim -- 2.6%
Average -- 12.9%
Stocky -- 15.7%
Fat -- 11.6%

The higher rates of infidelity across the board for men is a result of the male tendency to overstate sexual activity and/or the female tendency to understate it. But there is little reason to assume that the level of exaggeration shifts by weight categories across genders. The GSS does not ask a question about physical attractiveness, but I would guess slim women tend to be the hottest (and also probably the least masculine--so try to snag a slender girl for marriage!), while average weight is probably the most attractive male category. Three-fourths of male respondents are gauged as average, while nearly two out of three women are. So among both genders, fatties are more likely to roam (probably in part because they tend to be married to other girthy folks who leave them wanting).

The average (mean) years of education for those who have cheated and those who have not, by gender:

Men (N = 6394)
Cheaters -- 13.49
Faithful -- 13.62

Women (N = 8551)
Cheaters -- 13.25
Faithful -- 13.27

There is little difference among cheaters and the faithful in terms of educational attainment.

The average (median) annual individual income, adjusted over time for inflation:

Men (N = 4243)
Cheaters -- $24,782
Faithful -- $24,830

Faithful men are slightly more educated than cheating men are. They make the same amount of money.

I am not sure how to gauge the same for women without artificially skewing the average based on family situation. Since faithful women are presumably more likely to remain married than cheaters are, they are more likely to have a breadwinner at home and consequently have less need to earn an income independently.

Average IQ, as estimated by converting Wordsum test scores, presuming a white male mean and white female mean to each indicate an IQ of 100 (with a standard deviation of 15) for the genders, respectively:

Men (N = 3152)
Cheaters -- 101.4
Faithful -- 99.7

Women (N = 4347)
Cheaters -- 101.7
Faithful -- 99.3

Cheaters tend to be a little more intelligent than the faithful are.

Across three of the four variables, the differences among cheaters and the faithful are modest. Life outcomes in terms of income, intelligence, and education do not seem to substantially predict levels of infidelity, although it is worth noting that those who are faithful appear, in aggregate, to make a little more (income, educational attainment) out of what they're given (intelligence) than the unfaithful do.

What, then, predicts fidelity? The percentage of cheaters for each response to the question of confidence in the existence of God:

Men (N = 2835)
Do not believe -- 30.2%
No way to find out -- 28.3%
Some higher power -- 26.9%
Believe sometimes -- 24.3%
Believe with doubts -- 21.1%
Know God exists -- 20.2%

Women (N = 3747)
Do not believe -- 16.7%
No way to find out -- 18.2%
Some higher power -- 25.0%
Believe sometimes -- 11.0%
Believe with doubts -- 13.6%
Know God exists -- 11.5%

Stay away from girls who are into mystical stuff like Kaballah, Yoga, and Buddhism.

The percentage of cheaters among self-described political liberals, moderates, and conservatives:

Men (N = 5549)
Liberal -- 25.5%
Moderate -- 21.5%
Conservative -- 18.7%

Women (N = 7247)
Liberal -- 16.5%
Moderate -- 13.1%
Conservative -- 10.5%

Get a girl you're interested in to tell you all her thoughts on God and the President. Other things equal, find a conservative girl who loves Jesus (and probably her mom and daddy, too). The same advice applies to women seeking husbands. The firmer his theism is, the less likely he is to run around on you.

* Anecdotally, three times I've been appoached by the girls of guys I'm acquainted with (though not friends with). All three are from working class backgrounds, and two of the three are dysfunctional at that.



Ron in L.A. said...

I think you're testing people for their truth telling or desire to please or impress a test taker, rather than actual infidelity. The uglies probably see the dalliance as an affirmation of their desirability, and the people who profess certainty about God are the most pretentious liars.

Just a theory.

georgesdelatour said...

You don't mention the most obvious factor.

How sexually active is the marriage? If the couple are having lots of hot sex every night and are both fully satisfied, the marriage is unlikely to have a problem. But what if the sexual aspect of the marriage is not so hot? Especially if one partner turns out to have a lot less libido than the other? The main reason a partner cheats is, presumably, because they want more (or better) sex than they're currently getting. If you're satisfied, why bother cheating?

I suspect the most enduring marriages are, quite simply, the most sexually active marriages. Is there any evidence for or against this?

Anonymous said...

I would have expected, controlling for religious beliefs, the gap in behavior between conservatives and non-conservatives would disappear.

I was surprised to find the POLVIEWS gradient was steeper if you look at men who "know god exist" (N=324). More surprising is that conservative atheists appear to more likely to cheat than liberal atheists (this is true even when you define atheists very broadly, i.e. "believe sometimes").

Jokah Macpherson said...

AE, do you plan on following this advice yourself with regards to belief in God? I gather from reading your blog that your own religious beliefs are not clearly defined, and I wonder if differences in religious fervor would offset any statistical advantages for the marriage that might otherwise exist.

Georgesdelatour, I found your question intriguing so I went ahead and checked for myself using SEXFREQ. Your theory mostly holds but I noticed two points of interest. First of all, cheating began increasing again when sex exceeded once a week - I suppose you can't control for "sex with spouse" vs "sex with mistress." Secondly, those who weren't having sex had the lowest cheating rate of all, presumably for the same reason.

Donna B. said...

If I could figure out the GSS (I'm not that smart) I'd check it myself for age and duration of marriage. Age, especially.

As for fatties, that's really interesting. I've been fat all my adult life and have never had a problem getting a partner, fat, slim, or in-between.

Even now, in my late 50s, I occasionally get hit on. I rather like it, though I've never acted on it.

rob said...

AP, don't you need to compare cheaters to noncheaters on those various dimmensions? If say, 1/2 of married people are fat, and 30% of cheaters are fat, fat people aren't more likely to cheat. Interesting post.

Jason Malloy said...

"More surprising is that conservative atheists appear to more likely to cheat than liberal atheists"

No, conservative atheists appear to be the least moral in general:

"It’s true that religion is the essential reason conservatives give more, and religious liberals are as generous as religious conservatives. Among the stingiest of the stingy are secular conservatives."

J said...

I hope you were audacious with those three ladies.

Reader said...

Stay away from girls who are into mystical stuff like Kaballah, Yoga, and Buddhism.

Unless you're just looking for an easy lay, of course.

georgesdelatour said...

You also have to factor in opportunity. Lead singers of rock bands probably have lots more opportunity to cheat than farmers do. It's always easier to resist temptation if you're never exposed to it.

In terms of the religious aspect, how would this apply to the world's second most popular religion - Islam? Under certain versions of Islam, a man can have up to four wives; his wives are never allowed to refuse him sex; he can get divorced simply by saying "I divorce you" three times; and a "marriage" can last as little as one hour. So most of what non-Muslims would consider extramarital cheating can actually be incorporated into regular Islamic marriage. On the other hand, adultery can get both parties stoned to death. This seems to be less because the adultery hurts the feelings of marriage partners. It's more because it's a trespass on clan property, which, if not dealt with severely, could lead to Montague-Capulet type clan wars.

Audacious Epigone said...


That's always a possibility when looking at self-reported answers that don't require verification by the interviewers. But what is there to do about it? Short of disregarding most social data, it's just something to keep in mind. If there isn't a reason to think the level of truthfulness systematically varies by responses (you're suggested it does, here, of course), it's not really a problem.


Re: the frequency of having sex, I wonder how that breaks down by gender. My assumption is that to the extent that it is an influencer, it is more so in men. The marriage would have to be really bad from the man's perspective for the woman to want it more than he does. I'm out of town this week, but your suggestion sounds interesting and worth taking a look at.

Re: Islam, that would be an issue of question interpretation, I suppose. But the GSS is a US survey, so it shouldn't materially influence the results.


I think of Dave Matthews pleading, "You don't need to prove a thing to me, just give me faith, make me believe..."

I'm agnostic, hoping for theism (of the kind that will be gentle toward me for being so uncommitted!). I


Well, there you go! If you're a heavyset woman in your fifties who is hit on, there must be some attributes you have that are well above the average in attractiveness, like a feminine face with relatively tight skin.


All the comparisons are of those who have cheated versus those who have not. So in the case of fat men, 30% who have been or are married have cheated on their spouses, while for thin women, fewer than 3% have. Or am I missing your question?


Heh, no, just a gentle refusal along the lines of "flattered but spoken for". I'm only an average looking guy, so it's not as though it were anything special about me, and loose women don't appeal to me. Besides, only one of them is hot.


Interesting. I wonder if that is because non-religious conservatives tend to feel self-described most accurately as libertarians, but because of standard political spectrum questions like the GSS, have to place themselves somewhere on the liberal-moderate-conservative spectrum and go with conservative because they feel that's the least inaccurate.

rob said...

Ah, I totally misunderstood what the numbers were. I thought they were % of cheaters in those categories. I should read more closely and think a bit before I comment. Like see if percents add to 100.

Audacious Epigone said...


The first set comes pretty close to 100%, but that's just coincidence. Anyway, it just means my presentation needs to be clearer!

Douglas Knight said...

But what is there to do about it?

Ron didn't say "they're lying, so disregard all sex data," but proposed an alternative theory of who would benefit from which lies. It probably has testable consequences, but it sounds overfitted. I think it would be useful to find someone who has not read the post to guess which way the lies would go.

Anonymous said...

Audacious Epigone wrote:

"Get a girl you're interested in to tell you all her thoughts on God and the President. Other things equal, find a conservative girl who loves Jesus (and probably her mom and daddy, too). The same advice applies to women seeking husbands. The firmer his theism is, the less likely he is to run around on you."

Sound advice.

DK said...

The average (median) annual individual income, adjusted over time for inflation:

Men (N = 4243)
Cheaters -- $72,944
Faithful -- $76,282

LOL. You can stop right here. This is clearly not a sample representative of the USA. Median household income here is 48K. So where do these men earning >70K individually come from???

Blode032222 said...

I'd hesitate before assuming that the below average females are the most desirable physically. Anorexia nervosa has a prevalence of perhaps 1.1 to 2.8% in females aged 16-65. This is noted to be higher among the younger age groups. And the Below Average category for females on the GSS encompasses only 7.2% of the population ... a fair chunk of them are going to be distressingly skinny to anyone like me.

Granted, not everyone is like me, but then again, I've also noticed that the first thing Americans assume about female beauty is that other Americans think only ultra-thin women are attractive. Most everyone seems to exempt themselves ... and supermodels, famous though they may be, just don't get as feverish a response a curvy glamor types (who, granted, are thinner than the curvy glamor types of the 1950s).

Anyway, alls I'm sayin is, the fit women in the tenth and twentieth percentiles are still in the GSS Average category....

Audacious Epigone said...


Wow, I missed the obvious there. I just ran the numbers again, which is a tedious process when multiple variables are used and the median is sought for, and found that the median is essentially same for both men who have cheated and men who have not (among those who are or have been married), at just under $25k, which is much more reasonable, given that some men are not working or are only working part time. I'm not sure where I got the $75k numbers. Thanks for alerting me to it. Sorry for the error.


Good point. The average category really is too broad to expect much insight from.

Anonymous said...

"Stay away from girls who are into mystical stuff like Kaballah, Yoga, and Buddhism."

has a link, but the link only goes to the comment page for me. I was expecting the link to go to an article on why mystics are more likely to cheat.

Anonymous said...

The original post reads:
Women (N = 3747)
Do not believe -- 16.7%
No way to find out -- 18.2%
Some higher power -- 25.0%
Believe sometimes -- 11.0%
Believe with doubts -- 13.6%
Know God exists -- 11.5%

Stay away from girls who are into mystical stuff like Kaballah, Yoga, and Buddhism.

The highest number is for "some higher power" which might fit mystics, or then again it might not fit mystics who have highly specific beliefs. So I guess the author means that believing in "some higher power" automatically makes a girl interested in mystical Qabalah, mystical Buddhism, etc.

(Funny, I'm writing from Asia, and all the Buddhists I meet are quite conservative, and mostly quite uninterested in mysticism.)

Is there a body of evidence that I'm missing that connects the various occult subcultures (tantric yoga, neopaganism, etc.) to the numbers in the post?

Blode032222 said...

Oh, by the way, the Vanderbilt statistics I cited for prevalence of AN may have overstated the case by an order of magnitude. (At least they're not as bad as the statistics Naomi Wolf cited, e.g., half of American female students suffers from bulimia or AN.) Feminine Beauty is an interesting source, and the page I linked is even safe for work!

Actually that site made me feel a lot better ... starving yourself to death has to be pretty awful, whatever the root cause. Looks like it's worst in Italy, of all places!