Some basic facts about the distribution of sex

The Canadian Community Health Survey is an annual voluntary survey, carried out by Statistics Canada, that collects information about a wide range of health outcomes and risk factors. As part of the 2013-14 survey, 47,764  Canadians between the ages of 15 and 49 were asked about their sexual activity - whether or not they have ever had sex, and if they have had sex in the past year.

The majority of those surveyed reported being sexually active, as shown in the graph below. For example, about 88 percent of 18-19 year old women surveyed in 2013-14 (or 0.88) reported having sex in the last year.  For women, the probability of having had sex in the past year peaks at age 25-29. For men, it peaks 10 years later at ages 35-39. 

These estimates, because they are based on survey data, are subject to some margin of error. The vertical bars represent 95 percent confidence intervals - the true population mean will lie within that confidence interval 19 times out of 20. When there is little to no overlap between two confidence intervals, it is fairly certain that one group of people is having more sex than another group of people. So it is possible to say, with a high degree of confidence, that 18-19 year old girls are more likely to be sexually active than 18-19 year old boys, and 45-49 year old men are more likely to be sexually active than 45-49 year old women.

Apart from age and gender, one of the best predictors of sexual activity is marital status. A key purpose of marriage is, after all, to legitimate the results of sexual intercourse. This next diagram shows that, all else being equal, men and women who are married or living common law are much more likely to have had sex in the past year than men and women who are single, divorced or widowed.

Of course sexually active married people may not be having sex with their husband or wife - indeed the slight divergence in the married sex lines when people reach their 40s is suggestive of an uptick in male extra-marital sexual activity. There may also be selection effects at play, that is, people with strong sex drives may be more likely to get and stay married.

The above diagrams distinguish between the celibate and the sexually active but do not, for people who are sexually active, capture the frequency of sexual activity. One way of getting at sexual frequency is with information about the form of birth control used. People in long-term relationships will typically invest in easy to use, reliable and non-intrusive forms of contraception, such as the birth control pill or the Minera IUD. For a short-term relationship or a one night stand, however, a condom is a wise choice. Condom use, therefore, can be taken to be an indicator of irregular sexual activity. This next diagram presents the portion of people who used a condom in their last sexual encounter:

Men are consistently more likely to report using a condom the last time they had sex than women are. Some of this difference, however, is likely due to the fact that gay men often use condoms during sex, but lesbians do not.

For young adults, the CCHS also collected information on the use of condoms for birth control purposes - numbers which should not be affected by differences in condom use between gay men and lesbians. Again, the same pattern of greater male condom use appeared:

This graph is consistent with the idea that women in their early twenties are not only more likely than their male contemporaries to be sexually active, they are also more likely to having sex in the context of a long-term relationship, where there are regular opportunities for sex.

Recent discussions about the "incel" - the involuntarily celibate - have raised questions about inequality of sexual experience. These charts raise two points relevant to this debate.

First, there is a difference between inequality of sexual opportunity at a point in time, and inequality of sexual opportunity from a life-time perspective. A number of men who are not having sex at 18, 20 or 22 may well be more sexually active by the time they hit their late 30s, and the reverse will be true of their female contemporaries.

Second, the greatest facilitator of sex is marriage and cohabitation. Anyone who is concerned about reducing inequality in access to sex should actively work to create the conditions that allow people to marry, or enter common-law relationships. 

There is a large literature suggesting that people are more likely to get married if they have steady employment, and earn enough to buy a home with space to raise a family in. This means that there is no conflict between fighting for greater economic equality and fighting for greater sex equality. The policies that reduce income and wealth inequality are the same policies that put money in people's pockets so they can afford to go on a date. They're the policies that help people afford their own homes, where they can get down and dirty any time they want (at least until kids come along).

Comments

[Jim Sentance]

Interesting data. Just thinking the slight divergence in the married sex lines when people reach their 40s might also be due to age differences between partners. Could also explain the earlier gap in overall numbers as well of course.

[Frances Woolley]

Jim, yes, you're absolutely right. I hadn't thought of that. Good point. So many of these patterns are driven by age differences, specifically men dating younger women.

[Linda Welling]

The age difference was also my first thought. If you look at a 3-5 year gap, does that take care of much of the differences?

[Frances Woolley]

Linda - it's hard to say - the CCHS, like just about every other StatCan survey, has only one respondent per household, and the CCHS just collects info on one person. So I can't do any more than you can, really, and just eyeball the differences.

You did prompt me to run a regression analysis, and there are quite a few significant predictors of sexual activity - immigrants are less likely to report having sex than non-immigrants, people living in Quebec and Newfoundland are significantly more likely to report having had sex in the past year than people in the rest of the country, and people who are moderately physically active or physically inactive are less likely to report having had sex than people who are physically active. Once all of those controls are thrown in the regression equation (marital status, etc etc), the likelihood of having sex in the past year peaks for women and men at the same age - 18 to 19 years old (though that might change if I put a marital status/age interaction term in the regression equation).

On facebook someone linked to this site - https://rationalwiki.org/wiki/Manosphere_glossary. That site appears to articulate a couple of hypotheses. The first is that young women are more sexually active than young men, and this seems to be supported by the data. The second hypothesis is that women have fun, no-strings-attached sex with "alpha males" when they are young and hot, but then when they want to settle down, they find some poor "beta" who will support them while they raise their kids (and whom they can then take for half his assets in a divorce settlement). It's much harder to find support for this idea in the data. The data on birth control methods is much more consistent with the idea that younger women are in steady relationships with men who are slightly older than they are.

[Colin Percival]

I think the "use of condoms for birth control purposes" numbers need to be taken with a caveat: Men and women won't necessarily agree about why they're using a condom. In particular, *women know if they're on the pill* -- so there are probably many cases where a pill-taking woman would say that she's using a condom just for protection from diseases while her partner says he's using a condom for protection from both diseases and parenthood.

The use of "last sexual experience" could also introduce problems, in that it introduces multiple-counting: It's possible that many women had their most recent sexual experience with the same man; if that man doesn't use a condom while other men habitually do, you would get the same pattern of higher condom usage among men in "most recent sexual experience" counting.

[david]

Couple of thoughts:

- the demographic transition has an especially harsh impact with hypergamy - the ratio of slightly older heterosexual men to slightly younger heterosexual women is itself not constant. If the postwar boom saw a sexual revolution, one wonders whether a cultural reaction is only to be expected with an aging population. Was birth control technology really the only innovation there?

- in a bilateral pairing model, even a slight population imbalance implies huge elasticity in the "clearing wage". This applies whether it's marriage or kidney donors.

- we can detect how large #2 really is by considering the consumption of substitutes, which happily leaves a large paper trail - i.e., we should look to cross-nationality marriage formation in the census, one supposes as it varies by age and gender.

[david]

The observation on the elasticity is dimly due to Becker, I think, but don't quote me on that.

[Frances Woolley]

David - interesting thoughts, thanks for posting them.

"the demographic transition has an especially harsh impact with hypergamy"

I think you're right, but I don't know how right - that is, how strong these effects are. I've been toying around with trying to estimate how the pool of potential partners for people of various ages has been changing over time, but I can't figure out how to do it without making some pretty arbitrary assumptions about who is considered an acceptable partner. Also there are complex knock-on effects - so 60 year old male baby boomers partnering with 40 year old female gen-x-ers will cause 40 year old male gen-xers to turn to younger cohorts to find partners, etc.

"the consumption of substitutes" - how well does porn + video games + a parent to do laundry and make dinner do as a substitute for a real-life partner? People in their 20s today have grown up with vastly greater access to pornography than any previous generation. Could it be that part of the rage about not having expectations met stems from grossly unrealistic expectations shaped by movie-sex?

Cross-nationality marriage formation wouldn't really do it, especially in Canada, because of changing immigration patterns. Also travel, internet makes proximity much less important in terms of the marriage market.

As an aside- women born at the leading edge of the baby boom have had a pretty rough marriage market too - my gut feeling is that some beliefs about, e.g., divorced women finding it hard to find new partners are shaped by the experience of that generation, and don't obviously apply to women born in the late 60s, 70s, 80s.

[david]

I did not think about the basement-dweller hypothesis, but it would seem to have the wrong sign, at least insofar as a male-biased withdrawal from the marriage market. That would imply women bidding up to obtain men, who have instead decided to boycott date night. Unless we are not really thinking of substitutes in the typical sense, but instead in terms of some kind of prospect theory.

It's a bit like thinking about NEETs. If the charge is: they've all withdrawn from the labour market to play video games and/or eat millennial avocado toast, then we would be quite surprised to hear of more political demands for low-income work at the margin.

The hypergamy/demographic transition notion implies that women on the marriage market during the baby boom really got the short end of the stick. Which brings to mind, you know, there's a bit from Kerouac's On The Road where Galatea calls out Dean Moriarty for lying to women in order to sleep with them, lying that he will marry them when he gets them pregnant and instead disappearing with their money, and then returning to hit them if they dare to sleep with other men. For this, Galatea is the one who is condemned by Kerouac for being unreasonably old-fashioned.

[Frances Woolley]

David - "Unless we are not really thinking of substitutes in the typical sense, but instead in terms of some kind of prospect theory." - yes, that's exactly my point.

Though one can argue that there is also some amount of women bidding up to obtain men - e.g. 60% female campuses (1.5/1 female/male ratio) arguably facilitate "bro" and hook-up culture.

"then we would be quite surprised to hear of more political demands for low-income work at the margin." - Are we hearing such demands? I hear demands for income guarantees (e.g. UBI) and increased minimum wages.

"The hypergamy/demographic transition notion implies that women on the marriage market during the baby boom really got the short end of the stick". Perhaps I'm misunderstanding what you mean by "on the marriage market", but I'm having problems understanding your reasoning here. Baby boom is 1945 to 1962, say 1945 to 1960 to make it easy. If women enter the marriage market at age 20, the women on the marriage market during the baby boom are those born 1925-1940. These women (assuming 5 year average age difference) are looking to marry men born 1920-1935. Not quite obvious how it played out for that generation. Birth rates plummeted during the great depression. That means there were fewer babies/women born 1925 to 1940 than babies/men born 1920 to 1935. This would make it good for women on marriage market during baby boom. But then a bunch of guys got killed in WWII. My guess is that WWII would have more influence than anything else on those marriage rates, but I don't know - that depression fall in birth rates was pretty big.

[david]

With prospect curves I am hesitant to predict what a robust matching model dynamic would look like, really... to be frank I would have no idea. Same problem as prospect theory in general: binds practically nothing.

If workplace effects are strong here, we should be correspondingly skeptical of national effects in noisy data aggregated at a provincial or national level... it seems implausible that the proportion and magnitude of gender-unbalanced workplaces are nicely symmetric, and then there's age to consider too.

Zero hours contracts have become an issue in the UK and NZ (not sure about Canada).

Depression/hypergamy vs WW2: true! I suppose it would come down to picking through the marriage data and census pyramid. I defer to your expertise there. One thing, though, when fertility dropped during the Depression we are still looking at a drop to a TFR of 2.5+ - still well above replacement. This makes sense: we know the Canada population ex migration kept growing, WW2 or not.

[Jacques René Giguère]

Some fast thoughts from the airport lounge.
Baby boomers (1946-1964) were the one to invent the "sexual revolution", that is women lowering their standards viz monogamy and shortening their skirts. By 1983, popular magazine like Newsweek (who did an issue on that) talked about a new conservatism, that is younger women born after 63 pairing with men born slighly earlier ( thus in greater numbers). No new morality, just basic supply and demand.
Birth rates plumeted during the Depression, thus this generation (my parents) having the great success of the 30 Glorious and bemoaning how the young can't make it like they did.
All of this is clasic Easterlin (with more numbers.)
Better and more available porn may not be the best substitute for real relationship but if it's the only thing that relative income and population ratio leaves a generation, they'll have to make do.
As for the men-women ratio on campus, there is LTG "lesbians till graduation"...