Within versus Between Sample Variation, or Why an Equal Marriage Means More Sex

It's clickbait, New York Times style: "Does a More Equal Marriage Mean Less Sex?" Spurred by findings of a paper by Sabrino Kornrich, Julie Brines and Katarina Leupp published in the American Sociological Review, the article argues that, "too much similarity in egalitarian marriages leads to boredom and decreased sexual frequency". In general, "the less gender differentiation, the less sexual desire.” "On an emotional level, “kindred spirits” sounds lovely. But when it comes to sexual desire, biology seems to prefer difference." 

But do these findings really stand up to serious scrutiny?

The key findings from the Kornrich, Brines and Leupp paper are summarized in the picture on the left. Sexual frequency (measured in times per month) is increasing in the men's share of non-core housework (things like car repairs), and decreasing in the share of core housework (things like vacuuming and laundry). The average proportion of core household chores done by men was, according to men's reports, 0.25. With that sharing of housework, a couple is predicted to have sex about once a week, or just over four times per month.

The graph looks impressive, but just how big are the reported effects? The standard deviation of men's share core housework share was 0.19, and the estimated coefficient on that share was -0.416. Because the authors used a negative binomial regression, and sexual frequency isn't normally distributed it's not entirely straightforward to know how big these effects are. But roughly speaking, going from say the bottom quarter of the household chores distribution to the top quarter would change sexual frequency less than once per month [corrected]. 

This is what we're getting so hot and bothered out about? Really?

Sharing of household chores might explain some of the differences in sexual frequency across couples, but the effects are not, by any stretch of the imagination, large ones. I don't have access to Kornrich, Brines and Leupp's data, so I can't replicate their findings. But I do have access to the Canadian Community Health Survey, which has a certain amount of information on the frequency of sexual activity. 

The Canadian Community Health Survey asks respondents (who have ever been sexually active) whether or not they have had sex in the past year. The results by age and marital status are shown below:

I don't know if it's safe to conclude from this graph that the sexless marriage is a myth - after all we don't know who married people are having sex with. It may not be with each other. Plus people may not always give honest responses. But almost all married people under the age of 50 have sex at least sometimes, while about 30 percent of divorced/separated/widowed people in their 40s have not had sex in the past year.

While not entirely conclusive, this graph makes me strongly suspect that the kind of differences in sexual frequency reported by Kornrich et al are dwarved by the differences in frequency of sexual activity between married individuals and single or separated/divorced/widowed individuals.

The between sample variation completely swamps the within sample variation. 

The basic logical flaw in the New York Times article should now be clear. The NY Times is arguing that, conditional upon being married, sharing of housework results in less frequent sexual activity. Yet it seems not at all unreasonable to suppose that willingness to share housework might have some impact on the probability of couples staying married. If sharing housework makes it more likely that a couple will stay together, then it's entirely possible that those with more egalitarian values get more nookie, not less.

Put another way: for a marriage to stay together, someone has to be getting something out of it; if not sex, then sharing, or friendship, or companionship. Kornrich et al may be right that men with traditional attitudes towards marriage have more sex - if they can persuade someone to marry them, and stay married to them. Otherwise - well, once you're over 40, the outlook isn't good.

Note: graph created by the following stata code:

gen sex=SXB_3

recode sex 2=0

graph bar (mean) sex [pweight = WTS_M] if DHHGAGE<=9 &DHHGAGE>=5, over(DHHGAGE) over(DHHGMS) 

and then doing a bunch of fiddling about to make things look pretty.

Comments

[Mike]

I'm with you all the way up to the end, you make an assumption about the likelihood of staying married related to the sharing of chores. A quick google search returned a couple of reports saying the opposite. Not saying they are correct, they likely aren't, it is just that logic doesn't always work when talking about relationships. I think it is unlikely that any one trait/indicator has a large effect on something as complex and dynamic as a relationship between two people.

I know that in my marriage I get much better rewards from installing a garage door opener or fixing a banister than I do from cleaning up after supper.

[Frances Woolley]

Mike - Agree on the lack of evidence for a systematic relationship between equal sharing and marital stability. Basically in a more equal relationship it's easier for the woman to walk away, and more often it's the woman who walks away than the man.

The basic point, though, is that one of the key purposes of marriage as a social institution is to legitimize sex. People who are able to get and stay married - whatever it takes to do that - are more likely to get lucky (at least occasionally) than people who aren't.

[RPLong]

What's funny is that the researchers have no way of knowing how often the couples would be having sex if the man stopped doing the dishes. My guess is, however little they're having now, they'd be having even less if one party become more uncooperative.

"Happy wife, happy life." Or, in more neutral terms, "happy spouse, happy house."

It's almost certainly true that the majority of heterosexual wives want their husbands to be manly. So do the dishes like a man. Best of both worlds.

[Patrick]

Their story doesn't hold together. Baseline libido varies from person to person, and has more to do with biology than dishes and laundry. It also occurs to me that changing family circumstances and major events have bigger, but transient, effects. Most couples I know long ago worked-out who does what housework, and it quickly becomes a non-issue; too many other things that need one's attention.

[Chris Auld]

Great article, Frances. It's also worth pointing out the other endogeneity problem: household chores are not randomly assigned; it's plausible that the same unobserved factors which influence chores also influence sexual frequency (e.g., maybe good-looking men do fewer chores because we, I mean, they, can capture more of the marital surplus, and maybe also have more sex). Put another way, the authors' and NYT's interpretation that any given marriage would experience more sex if the man did less housework does not follow from the evidence presented.

[Sandwichman]

Frequency is only one dimension. How about duration? Intensity?

[Nick Rowe]

Chris Auld is stuck in spam.

[Frances Woolley]

RPLong "do the dishes like a man" - love it! There is some truth in this, too - the kind of Gordon Ramsey/Mark Bittman masculine approach to cuisine.

Chris - thanks for your comments. Yes, endogeneity, absolutely. Also spurious correlation. In this data set, women reported higher sexual frequencies on average than men did. They also reported lower male shares for core household chores. This difference in reporting alone explains some of the sexual freuqency/household chores correlation.

Sandwichman, just the idea of marital sex is traumatic enough for gentle readers. My students are always stunned/horrified by the above picture when I show it to them (to which my response is "why do you think people get married anyways?!?") Let's not go there! (or, at least, not here.)