NPR is reporting about a study that shows a correlation between skinny kids and full fat milk. They then make the leap that the high fat milk is causing the skinny kids. This is the simplest case of error bias I’ve ever seen. Parents who have skinny kids give their kids full fat milk. The hope is that the full fat milk will help the kid gain weight. So, of course there’s a correlation between high-fat milk and skinny kids. It is causal — skinny kids are given full fat milk! The study has reversed the cause/effect. Here’s the concept written in symbolic logic: skinniness ->full fat milk. This, with logic, we can also say ! full fat milk –> !skinny kid. The logic maps the correlation perfectly, but only when you realize the causality chain.