Have 1 in 5 Americans Been in a Consensual Non-Monogamous Relationship?

There is nothing with which modern relationship journalism seems so peculiarly infatuated as non-monogamy. Call it "polyamory," "swinging," or "consensual non-monogamy" (CNM)—if reporting is to be believed, it's everywhere.

The latest contribution to the CNM craze comes from CBS, which last weekend debuted a new documentary on "[f]ighting the stigma of consensual non-monogamy." In promoting the show, the network tweeted out the eye-catching claim that "1 in 5 Americans have been involved in a consensually non-monogamous relationship at some point in their life." CBS is far from the only outlet to push the "one in five" claim: it's appeared in Rolling Stone, Quartz (as cited by NPR), Time, Men's Health, and Psychology Today, among others.

Where does that number come from? Essentially all of the articles point to the same source, a 2016 study in the Journal of Sex & Marital Therapy by a group of researchers at the Kinsey Institute (hereinafter collectively referred to as Haupert et al.). The abstract of the study does indeed confirm that "more than one in five (21.9% in Study 1; 21.2% in Study 2) participants report engaging in CNM at some point in their lifetime."

The study itself is a straightforward survey. Haupert et al. used two waves of the "Singles in America" study, an annual survey of single American adults administered by Match.com through U.S.-based research firm ResearchNow. Respondents to the first survey were over 21; respondents to the second survey were over 18.

Wait a second—all the respondents were single? Yes: the first wave covered "those who were legally single at the time of the survey," meaning people who were single, casually or seriously dating, cohabiting, or engaged. The second wave covered "only those who were either single and not seeing anyone, or single and casually dating."

If your sample is only of single people, then your conclusions only generalize to the population of single people. Haupert et al. do try to argue that their "ever practiced" framing means that their findings might apply to married people, under the principle that all married people were once single: 

while many married Americans may have engaged in CNM, our focus on singles allows for widely applicable results, as so many U.S. adults are single for some duration of time. Further, those singles who go on to marry undoubtedly carry their prior relationship experiences with them, laying the foundation on which they build future relationships.

But, as decades of research have shown, married people vary systematically from their single peers. Among other factors, they are whiter, wealthier, and more religious. It is entirely plausible that a sample of entirely single people overrepresents a preference for polyamory—indeed, that they have not selected out of singlehood and into stable monogamy is one such indicator.

So, the most that Haupert et al. really allows us to say is that 20% of single Americans have experienced polyamory at some point in their lives. But is that what it lets us say? Does the study allow us to conclude, to paraphrase Mel Magazine, that "roughly 20 percent of [singles] say they’ve engaged in some form of a consensually non-monogamous relationship such as polyamory, swinging or opening up[?]"

Well, no.

According to the study, "[a]ll participants were asked if they had ever had an open sexual relationship." What's an open sexual relationship? "An agreed-upon, sexually non-exclusive relationship." 

This language could, of course, describe "swinging" or "opening up." But it could also quite plausibly describe casual dating, in which singles knowingly date, and sleep with, multiple people at once. Such relationships are perhaps, strictly speaking, a-traditional, but they do not meet most people's intuitive definitions of "polyamory," or even "open relationships" (which connotes a degree of romantic, but not sexual, commitment—a nuance uncaptured by the question).

In point of fact, some CNM relationships do not meet the definition of "an agreed-upon, sexually non-exclusive relationship," because "non-exclusivity" and "monogamy" are not the same thing. If three people all agree to be sexually exclusive with one another—a "throuple"—then they are all in a sexually exclusive relationship, and therefore do not meet Haupert et al.'s definition of CNM.

There's at least one other reason to be suspicious of Haupert et al.'s finding. Their methodology notes that they deliberately oversampled "homosexual men and women." In fact, 15.3% of study 1 and 14.3% of study 2 respondents self-identified as LGB (lesbian, gay, or bisexual). That's substantially higher than the population-wide prevalence of LGB people, which is generally pinned at 3 to 5%.

Previous research cited by the paper has shown, and Haupert et al. confirm, that identifying as lesbian, gay, or bisexual is associated with a significantly higher likelihood of reporting engaging in consensual non-monogamy. (It's one of two factors, alongside being male, that shows up as statistically significant in their regressions.) In other words, the study substantially oversampled the very subpopulation they then find is far more likely to engage in CNM. 

It's entirely possible that the researchers accounted for this by reweighting LGB respondents in their point estimates. But if they did, we wouldn't know. The paper includes no crosstabs, and in fact does not even explain how the 20% figure was estimated besides, one infers, bare division. The only efforts at representativeness in design Haupert et al. seem to have undertaken is to weight "recruitment targeting based on demographic distributions" seen in the Current Population Survey—a monthly survey conducted by the Bureau of Labor Statistics, which does not ask about sexual orientation. 

To their credit, Haupert et al. are honest about the limits of their findings. But that has not stopped dozens of journalists from using their research to perform a magic trick. At best, the study shows that one in five single Americans have engaged in CNM; more likely, it shows that one in five single Americans have engaged in a casual sexual relationship, with a subset of those engaging in CNM; possibly, 20% is an artifact of sampling choices. But before the eyes of thousands of readers, this figure has been transmuted into "1 in 5 Americans have been involved in a consensually non-monogamous relationship." Isn't that magical?

As always, the reality is probably more boring. Some single people engage in non-exclusive relationships; a smaller, unmeasured share probably engage in more formal "polyamorous" or "consensually non-monogamous" relationships, and that share has probably risen slightly. 

That's the conclusion of the 2018 “i-Fidelity” survey, which was conducted by YouGov for The Wheatley Institution at BYU, and found that 12% of respondents had ever engaged in an "open sexual relationship," defined as "an agreed-upon, sexually non-exclusive relationship with more than one partner." The study explicitly listed "polyamory, consensual non-monogamy, ethical non-monogamy, swinging" as examples, although it is possible it suffered to a lesser degree from the ambiguity highlighted above. In general, the study found CNM was more popular with young people, but that even among Millennials, fewer than 20% had ever tried it.

Polyamory may sound fun and exotic, but most of us don't live such fun and exotic (and complicated) lives. By their 30s, most Americans (80%) are either married or single, with little evidence that "alternative" structures are filling the gap for a significant share of adults. As Dr. Alan Hawkins recently put it, "the norm of marital monogamy is not crumbling" after all.

Charles Fain Lehman is a staff writer for the Washington Free Beacon, where he covers crime, law, drugs, immigration, and social issues. Reach him on twitter @CharlesFLehman.