Friday, October 20, 2006

Can I say it again?

Correlation does not constitute causation. This morning, I was greeted with the following article on IHE: Want to Get Ahead? Get Hitched. Why is it that journalists (or, gasp, researchers!) insist on making broad sweeping statements of causality, based on mere correlation? This reminds me of a classic article I read years ago (I'm talking late 70s, perhaps early 80s) in one New York city newspaper which proclaimed: "Calculators Cause Pregnancy!" As I recall it, it was an hilarious lampoon of just this sort of gee willacres, did you ever notice that traffic lights cause accidents?

It's a very pervasive thing, however, and distorts so much of our education, especially of science. No wonder so many people are confused about evolution. No... unless we stipulate an ultimate cause or shall I say Causer we simply can not say things like "ladybugs secrete a foul-tasting liquid through their legs in order to make themselves distasteful to predators". Sorry, that's stipulation, not science. Fine if you believe it, but it's still not science. Let's try: "Interestingly, ladybugs secrete a foul-tasting liquid through their legs, which evidence suggests has had a dampening effect on the desires of predators, likely aiding ladybugs in the goal of species survival."

Sheesh! Again, for the record: Correlation does not constitute Causation.

No comments: