On Saturday, my Twitter feed alerted me to a totally non-sensationalistic Gawker article called More Buck For Your Bang: People Who Have More Sex Make The Most Money. "Scientists in the adonis-laden European country [Germany] found that people who have sex more than four times a week receive a 3.2 percent higher paycheck than those who have sex only once a week. God forbid you don't have sex at all," writes Max Rivlin-Nadler.
In the comments section (yes, I know I shouldn't read them), aside from unnecessary pictures of scantily clad ladies and older white men in suits, several people hypothesize that causation probably goes the other way: people who make more money are able to have more sex. "People who have the most sex make the most money? Or is it that the people who make the most money have the most sex?" wrote one commenter.
The relevant study, "The Effect of Sexual Activity on Wages," (pdf) was published in the Institute for the Study of Labor (IZA, based in Bonn, Germany) Discussion Paper Series. Incidentally, the study is based on a survey of households in Greece, not Germany as the Gawker article states. A disclaimer at the bottom of the first page of the paper states, "IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper should account for its provisional character." So much for that one on the Internet!
The paper, studies the complicated relationship between various behaviors and characteristics—age, amount of sexual activity, personality traits, type of job, health, work experience, and so on—and wages. The author, Nick Drydakis of Anglia Ruskin University in the UK and the IZA, seeks to add to the literature on the correlation of health and well-being with income. He found that a regression model that did incorporate sexual activity levels in addition to health, age, and so on was slightly more accurate at predicting participants' income than one that didn't.
"It is unclear whether this correlation represents a causal relationship," wrote Drydakis in the paper. "In this study, we hypothesised that because the medical and psychological literature suggest that sexual activity is associated with good health, endurance, mental well-being, mental capacities and dietary habits, it could be perceived as a health indicator, which might influence returns to labor market activity….The patterns found in this study strengthen this reasoning," (emphasis mine). It's likely that health influences both sexual activity levels and income, and sex may improve certain aspects of health as well. The causal chain is likely very complicated and filled with loops.
Back in October 2012, Slate posted an article by Daniel Engber that kind of rankled me: The Internet Blowhard's Phrase: Why do people love to say that correlation does not imply causation? I'd like to think that I'm not an Internet blowhard, and I do stay away from the comments sections of most news stories unless I simply can't help myself, but I think articles like Gawker's report on Drydakis's paper are perfectly reasonable places for people to sputter "correlation is not causation."
To some extent, I understand Engber's criticism. A constant chorus of "correlation is not causation" doesn't really do much good, and it does get old. I don't want people to stop doing the studies that find correlation but don't present a causal mechanism, and I don't want news outlets to stop reporting on them. These studies do tell us useful and interesting things, and sometimes they can help us understand possible causal links between behaviors and outcomes. But nuance and skepticism are necessary in reporting and understanding the results of studies like this one.
Articles like Gawker's frustrate me not only because they pander and oversimplify. They also make it far too easy for people like me to roll our eyes, mumble, "Correlation is not causation," and dismiss an entire study as bogus. "Yeh, this seems like one of those studies that was commissioned just to meet publication quotas…" wrote a commenter on the Gawker article.
In fact, the paper is pretty interesting! It's interesting to see what variables Drydakis looked at and how economists go about disentangling mutually influential traits. He also refers to a lot of interesting research on various aspects of the interplay between health, sexual activity, and income. This article isn't the last word in the study of sex and income, and it isn't free from problems; the data set comes from a survey, which is not necessarily reliable. On surveys like this, men often over-report and women often under-report the amount of sex they have. I don't have the full data set, but in the Greek survey, men reported having more sex on average than women, and it's not clear whether or not the discrepancy can be attributed to different sexual behavior in homosexual men and women. (In principle, in a representative sample of people who are remembering and reporting their behavior accurately, heterosexual men should report the same number of heterosexual encounters as heterosexual women do, on average.)
I am also curious about how health problems and disabilities were categorized. The paper indicated that within the group of people who are health impaired, those who had more sex had higher wages. But from the data in the paper, I am not sure whether people with similar disabilities or health problems were compared with each other, or if all of them were lumped into the same group. If people with very different health impairments and disabilities are in the same group, the regression may not be able to separate differences based on sexual activity from differences based on health. Despite these issues, I found the paper much more interesting and robust than I assumed it would be based on the Gawker article.
It's hard to write popular science articles that dive into the subtleties of papers like this one in an easy-to-digest way, but in my opinion, it's worth trying to do. Such articles do a better service to both science and our readers than sensationalistic sound bites that make me glare at my computer thinking, "Correlation is not causation."