Correlation/Causation (When It’s Convenient)

A while ago, news came out that eating processed meats (like bacon and sausage) is strongly linked to developing cancer. It doesn’t cause it, but there’s a correlation.

My natural inclination at the time was to make a few mildly amusing vegetarian-friendly jokes and move on. I already don’t eat that stuff, so it didn’t really impact me in any strong way. Plus, I don’t judge others for making their own food choices. So it wasn’t something I planned to harp on.

But oh my god, y’all. The way people reacted was intense. Suddenly individuals who never normally bother to question anything from studies were dissecting every word of every article to point out the differences between correlation and causation.

Because statistics are boring unless they mean someone might take away their bacon.

If only people cared this much about correlation vs. causation when it comes to other things, like fat acceptance. Everyone is perfectly happy to claim that being fat causes all these terrible things and will make us die early. They pay no mind to the difference between a definitive cause and a spurious relationship. No one digs into the details of a study, checks the funding source, or even clicks beyond a headline when it’s supporting the fat = bad myth. Only when it might mean no hot dogs.

Threaten someone’s meat consumption and suddenly they’re a statistical genius; point out that their fat hate is based on bad science and bias and it’s “everyone knows” and that we should never question a doctor. It’s a double standard.

Scientific skepticism is good, but you have to use it everywhere. Apply the fervor with which you sought a loophole in the sausage-cancer link (no pun intended) to things that can actually help people, like combating misinformation about weight and health.

Adapted from a post I published in 2015 on a different blog.

Add a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.