Thursday, April 10, 2008


Correlation does not imply causation.

The media rarely understands this, but looking at an article on eggs today it was particularly bad. It talks about people eating a lot of eggs being more likely to die than those who did not. Possible I suppose, but a remarkably hard question to answer. The article looks plausible until it gets to the last few paragraphs:

"Men who ate the most eggs also were older, fatter, ate more vegetables but less breakfast cereal, and were more likely to drink alcohol, smoke and less likely to exercise -- all factors that can affect the risk of heart attack and death."

I haven't looked at the real paper. However I would be willing to bet that people who drink and smoke more, and exercise less are more likely to die. Oh, and this group just happened to eat more eggs. Sounds to me like a classic example of a random correlation that is the result of cultural values. Sure they might have done some clever things to account for this. Given the track record of epidemiology I am not betting on them doing a decent job though.

No comments: