Easy As ABC

A really, really common style of headline is: “New Discovery: A Causes B!”

And my immediate thought is always, “What if C causes A and B?”

“Correlation does not imply causation” is old news, but it’s unexciting news. So it doesn’t get spread around nearly as much as it should. When you read the headline “A Causes B,” usually the article actually says something along the lines of “we observed a lot of A and B together, so we’re assuming one causes the other.”

There are plenty of ways to correct for this, of course – but rarely are such methods actually employed. It’s faster and more fun to just rush off with a bold claim. I recently read a story that claimed that toxoplasmosis (the weird parasite-borne disease that skews your behavior towards liking cats, and yes that’s a thing) also caused people to become more entrepreneurial and less risk-averse, because among the population of Denmark more entrepreneurs had toxoplasmosis than the rate in the general population.

So sure, one theory is “A Causes B;” in this case, “toxoplasmosis causes risk-seeking behavior in humans.”

But here’s another theory: “C causes A and B;” in this case, “a high innate risk tolerance causes humans to both be more entrepreneurial and interact more with cats.” One of the most common ways you can get toxoplasmosis is from a cat’s litter box – might it not be reasonable to say that more risk-averse people are more likely to be more cautious when dealing with cat feces?

I’m not saying I’m definitely right or anything. Like I said, there are controls you can put on experiments to rule that stuff out – but people rarely do, and it doesn’t look like they did it here.

There’s a real danger to taking “A Causes B” stories at face value. The implied second stage is usually “Since A causes B, if you find some way to artificially increase A, you’ll get the natural result of increased B.” But if A doesn’t actually cause B, you’re wasting resources trying to artificially inflate it.