Two words are key: correlation and causation. Correlation is a term that establishes the existence of a connection between two phenomena, i. e., it tells us clearly that variations in one of them are reflected in variations in the other. But it does not say if one of those phenomena causes the other. That's why the concept of causation exists. And the key point is that correlation not necessarily implies causation.

Seems obvious. It's not. And the best proof can be found, as is often the case, in the online world.

Just yesterday, a wire from EFE agency said that women with higher emotional intelligence quotients enjoy sex more ("Las mujeres emocionalmente inteligentes disfrutan más del sexo"). The source? A study made in King's College London and published in the Journal of Sexual Medicine.

The wire said: "'Emotional intelligence seems to have a direct incidence in the sexual workings of women, by influencing their capacity to communicate their sexual expectations and needs to their significant other', explained psychiatrist Andrea Burri, who directed the research" ("'La inteligencia emocional parece tener una incidencia directa en el funcionamiento sexual de las mujeres, al influir en sus capacidades para comunicar sus expectativas y deseos sexuales a su pareja', explicó la psiquiatra Andrea Burri, directora de la investigación"). From this it's easy to deduct a casual relationship: emotional intelligence influences a woman's capacity to communicate her partner what she wants. Ergo, emotional intelligence comes first (pardon the pun), and it allows her to enjoy more.

But hold on to your horses. Christie Lynn, who writes the blog "Observations of a nerd", has weighty objections. Please see "Does high EQ lead to more orgasms?".

Lynn says that the study falls victim of the quoted confusion between correlation and causation: "To link EQ to orgasm frequency doesn't mean that having a higher EQ allows a women to have more orgasms - not necessarily. It could even be the other way around, where more orgasms lead to better emotional stability and high EQ scores".

In his own blog, Greg Laden also takes a stand: "More emotional intelligence = more orgasms". He quotes straight from the paper: "Low emotional intelligence seems to be a significant risk factor for low orgasmic frequency".

After some comments, Laden ends up saying that, without a particular reason for it and regarding causality, "I'm going with 'orgasms increase EQ, IQ and general health and well being'".

What this means is that in this case we see an error that comes from the paper itself. The research authors (Burri, A., Cherkas, L., & Spector, T. (2009). Emotional Intelligence and Its Association with Orgasmic Frequency in Women Journal of Sexual Medicine DOI: 10.1111/j.1743-6109.2009.01297.x) assume the correlation implies a cause-effect relationship, one in which orgasms are the effect and emotional intelligence is the cause. Tsk, tsk, tsk.

Then there are the moments in which the problems come not from the paper but from the journalist. In "Facebook and Procrastionation", the Columbia Journalism Review tells the story of a poster presentation, as told by Earle Holland, from Ohio State University. Last april, the university's communications office released a text regardign the poster, clearly stating its weak points, its temporary nature, and so on.

What did the report say? That a small sample of students was analyzed and a correlation was found between the use of Facebook and low grades. The study did not suggest cuasality. But somebody would do it for them. London's Sunday Times wrote that:

"Research finds the website [Facebook] is damaging students’ academic performance. … Facebook users … are more likely to perform poorly in exams, according to new research".

Uf. Post mortem analysis would reveal that the published account was not what the reporter wrote, and there were also embargo problems, and so on. Part of this was also true from the previous example about orgasms: a reader commented that the correlation level (as measured in Spearman's coefficient) was so, so modest, "one has to wonder if there is really anything worth reporting".

So, to get to the point. Correlation is NOT causation. The fact that a nexus exists says nothing about what's due to what. It's very important as a journalist to ensure as much as possible that the story respects the spirit of the research. And if the research has a bias, as in the mentioned story, well, then you have to decide if you publish it or not. Lastly. it's also important to learn the ABC of statistics to have a clear notion of what terms such as these mean (taking into account the marvelous caveat emptor that Stephen Jay Gould gave us, alerting against the misuse of statistics in order to defend prejudice. See The Mismeasure of Man).

Happily, the two editions of the National Association of Science Writers' reporters guide touch this issue. The first edition (sorry, Amazon doesn't list it anymore) includes a chapter called "Coping with Statistics", by Victor Cohn, who also wrote News and Numbers: A Guide to Reporting Statistical Claims and Controversies in Health and Other Fields. The second edition includes the chapter capítulo "Understanding and using Statistics", written by Lewis Cope, Cohn's co-autor in the just mentioned book.

So I close this text with some of Cope's words: "Association alone doesn’t prove cause and effect. The rooster’s crowing doesn’t cause the sun to rise. A virus found in patients’ bodies may be an innocent bystander, rather than the cause of their illness. A chemical in a town’s water supply may not be the cause of illnesses there". Amen.

PS. After a few days, thanks to Bora Zivkovic (via Twitter, of course), here's two important explanations by Dr. Petra Boynton:

• Do 'emotionally intelligent' women have better sex?
• More on emotional intelligence and women's sex lives