About the SA Blog Network

Roots of Unity

Roots of Unity

Mathematics: learning it, doing it, celebrating it.
Roots of Unity Home

Sex Makes You Rich? Why We Keep Saying “Correlation Is Not Causation” Even Though It’s Annoying

The views expressed are those of the author and are not necessarily those of Scientific American.

Email   PrintPrint

Sex and money: the Bearina IUD, a conceptual intrauterine device design that would incorporate a copper coin. One of the most effective forms of reversible contraception is the copper IUD. For more information, click the picture. Image: Ronen Kadushin.

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.”

Evelyn Lamb About the Author: Evelyn Lamb is a postdoc at the University of Utah. She writes about mathematics and other cool stuff. Follow on Twitter @evelynjlamb.

The views expressed are those of the author and are not necessarily those of Scientific American.

Rights & Permissions

Comments 15 Comments

Add Comment
  1. 1. AllenKnight 10:49 am 08/20/2013

    A classic sociological study found poorer people have more rigid behaviors, and so have sex less often than their wealthier counterparts. There is no evidence that sex makes you rich, unfortunately…!

    The reference is:

    A Slum Sex Code
    William Foote Whyte
    American Journal of Sociology
    Vol. 49, No. 1 (Jul., 1943), pp. 24-31

    Link to this
  2. 2. rshoff 11:05 am 08/20/2013

    Evelyn, thank you so much for blogging on this topic. The whole idea drives me stark raving mad! Ok, well maybe just ‘bananas’.

    You hit the nail on the head when you said…. “Articles like Gawker’s frustrate me not only because they pander and oversimplify.” Squarely on the head I might add.

    Btw, does gawker mention sex can also destroy lives by leading to illnesses like HIV and other stds? Does he mention it can lead to badly timed pregancies or divorce? Does he take ANY social responsibility?!

    I guess people are removed from the self selected statistical sex pool after they suffer devastating losses as a result of believing they are Adonis.

    Link to this
  3. 3. JPGumby 11:12 am 08/20/2013

    Good article!

    On of my personal disappointments is that interesting articles are often summarized in ways that are misleading and leave out all the actually interesting parts.

    Link to this
  4. 4. rshoff 11:18 am 08/20/2013

    Regarding the study. It seems to me there may be several factors including:

    1-Sex is one of many aspects of social and professional networking. Unless your hiring a prostitue, there are factors other than sex that connects the partners. This networking may be loosely related to income/status/success. In this case, it’s the networking, not the sex.

    2-Healthy people probably have a higher incidence of sex. They probably have an easier time competing in the job market too. This has to do with health, not sex.

    3-Attractive people probably have a higher incidence of sex (assumption here). Attractive people are able to market their sex appeal and leverage it into income. After all, all prostitues don’t walk the street. Many are in the board room or box office, but prostituion it is.

    So there.

    Link to this
  5. 5. rshoff 12:18 pm 08/20/2013

    @Allen – Thanks for the reference. Perhaps people in the lower socio- economic groups are more rigid out of necessity and not out of choice (how can we separate the two ideas?) because being poor is dangerous and hard. There is not a lot of room for error, and when error does occur, there’s no money to throw at it for a solution….

    Link to this
  6. 6. TerrieV 12:23 pm 08/20/2013

    Thanks for the excellent blog post, Evelyn. I can get very frustrated with scientific reporting in the popular press, especially knowing it is the only way most people have exposure to research.

    Also, thanks for discussing the nuances of the “correlation dose not equal causation” mantra. As someone who teaches research methods and statistics to others, I appreciate how you present the complexity of how we understand relationships; essentially we should approach new findings with a balance of skepticism and curiosity. Papers like the one discussed above can be useful in generating hypotheses and may eventually lead to the discovery of causal mechanisms, but themselves don’t “prove” the nature of any relationship. Unfortunately this complexity in research and reporting can simultaneously create fervor and indifference in the public, depending on their level of scientific cynicism and how the press reports certain research findings.

    Link to this
  7. 7. rkipling 12:27 pm 08/20/2013

    Dr. Lamb,

    You are a voice in the wilderness, but keep at it.

    Link to this
  8. 8. denke42 3:20 pm 08/20/2013

    Germane to this discussion:

    This comic is available on a T-shirt, of which I have two, given me on the same day – one by my wife, the other by my son. Though the gifts were correlated, neither caused the other: it was Christmas and both knew I wanted the shirt.
    When I wear the T-shirt, I find it sad that, though many are interested by it, few understand the joke. I fear that this bespeaks a widespread failure to understand the logic involved, which bodes ill for our democracy. If citizens don’t grasp this basic principle of inference, how can they make wise decisions for themselves, let alone for their polities?

    Link to this
  9. 9. jgrosay 3:32 pm 08/20/2013

    ?people with higher income tend to have more partners’. It sounds simple: people with higher incomes would have more opportunities to socialize, thus mating, and, at least for women, it’s better choosing a richer man that would have more chances of guaranteeing the welfare of women and her eventual offspring, this is deep in the people’s minds, and can be considered also a manifestation of positivism, or rather materialism, but thousands of years after Commandments were revealed, there’s still crime, sexually transmitted diseases and wars, it’s hard thinking some things will have great changes. ‘If I were a richman, doody, doody, doody, doo…’

    Link to this
  10. 10. jgrosay 3:33 pm 08/20/2013

    Spike Lee’s ‘Jungle fever’ is a nice movie.

    Link to this
  11. 11. RockyBob 4:14 pm 08/20/2013

    Regarding “correlation is not causation”, I think most readers do not fully understand the usual requirements for demonstrating causation, and, therefore, do not truly appreciate why correlation is NOT causation. The absolutely crucial importance of “ruling out all reasonable alternative explanations” goes a long way toward bringing home the difference. If any reasonable alternative remains unaddressed, causation cannot be established. An entirely all-to-common unaddressed alternative is the claim that the subject and control populations are different. Multivariate analysis never fully accounts for all differences. Only by randomly assigning members of a single population to subject/control groups can this problem ever be eliminated.

    Link to this
  12. 12. Hitchiker of the Galaxy 4:52 am 08/21/2013

    People also didn’t notice that the mean difference was only 3% of income. Which is much smaller than the difference of earnings between poor and rich.

    Looking at the magnitude of change is another way to spot mis-interpretation of results in popular press.

    Link to this
  13. 13. Hitchiker of the Galaxy 7:18 am 08/21/2013

    (my previous comment appears to be lost).

    Difference between groups was relatively tiny, just 3% of income. Nobody noticed that it is much less than pay difference between average guy versus rock star or rich old CEO.

    Looking at the magnitude of change is another good method of spotting misinterpretation in news. BTW, Scientific American blogs are also regularly guilty of it.

    Link to this
  14. 14. Evelyn Lamb in reply to Evelyn Lamb 3:53 pm 08/21/2013

    Hitchiker (and anyone else whose comments seem to have disappeared): Comments are moderated, so it can take time for them to appear on the post.

    Link to this
  15. 15. rshoff 7:06 pm 08/21/2013

    Rockybob – You bring up a good point. Often times when I come up with alternate explanations for something, it’s simply to demonstrate that the original thesis is only a hypothesis. When I bring up those alternate reasons, it seems the people I’m talking with tend to think that I’m trying to propose my theory instead. In fact, I’m only trying to demonstrate what you say. As long as we can come up with alternative rational explanations, proven or not, the initial hypothesis cannot be presented as fact. Sometimes we have to roll our eyes when it is so simple to come up with rational alternates to interpretations to a study.

    Link to this

Add a Comment
You must sign in or register as a member to submit a comment.

More from Scientific American

Email this Article