Correlation or causation? In research, bet on the former


This article was originally on a blog post platform and may be missing photos, graphics or links. See About archive blog posts.

If trying to assess the risks and benefits of various foods, behavior and whatnot leaves you wanting to gorge on Oreos until you no longer care, small wonder. News coverage of medical and science studies, especially the headlines, often takes a few liberties when it comes to correlation and causation.

To be fair, most headlines can only hold so many words. (Note to headline writers: When in doubt, go with the all-purpose and deliberately vague ‘link.’ It’s hard to argue with ‘link.’)


So think it through. Did that study actually conclude that chocolate reduces stroke risk? (Something you’re going to be hearing a lot about as we approach Buy Chocolate Because That’s What You’re Supposed to Do Day.) Or did it conclude that people who regularly eat chocolate are less likely to die of stroke?

It’s hard to prove the former; and many, many things could affect the latter. Perhaps people who allow themselves to have a bit of dietary pleasure are simply less strung out and have fewer stroke risk factors to begin with.

Here’s David Katz of Yale University’s Prevention Research Center weighing in on that topic (and a timely review of chocolate-related research) in this ABC News story.

And here’s an explainer of cause and correlation from STATS at George Mason University. It states:

‘An action or occurrence can cause another (such as smoking causes lung cancer), or it can correlate with another (such as smoking is correlated with alcoholism). If one action causes another, then they are most certainly correlated. But just because two things occur together does not mean that one caused the other, even if it seems to make sense.’

And here are one professor’s ‘you goofed’ examples from the mainstream media. Fair enough. We each have our own ‘correlative-confused-with-causative’ favorites.


-- Tami Dennis