Correlation is NOT causation

Greg Mankiw has a nice post on the topic:

It really pisses me off when people confuse correlation with causation. Here is what I just read in the Harvard University Gazette:

Researchers at Johns Hopkins School of Medicine tracked 1,055 medical students for 36 years. Compared with cooler heads, the hotheads were six times more likely to suffer heart attacks by age 55 and three times more likely to develop any form of heart or blood vessel disease.

The conclusion is clear: Anger is bad for you at any age.

No, the conclusion from the fact given is not at all clear. People who get mad more easily do so for a reason--a more stressful job, bad genes, or some other mysterious factor X. Maybe it is the X factor, rather than the anger itself, that is bad for you. Maybe reducing anger without changing X won't change health outcomes at all.

I am going to call the author and give him a piece of my mind. Just as soon as this pain in my chest goes away.

It pisses me off as well. My feeling is that human beings NEED to ascribe a cause to an effect (preferably one that validates their personal beliefs) -- however unlikely/not-exactly-correct that may be.