“Why do purely observational studies fail so often despite finding such clear associations? The diet soda example tells the tale. All of those alternative theories I mentioned can be boiled down to a single, devastating possibility: what if diet soda drinkers are just fundamentally different from regular soda drinkers, in any of the ways I mentioned, and this difference colors everything about the way they live and behave? Scientists call this the selection effect, or selection bias. When human beings are free to behave as they always have—free to willfully choose their behavior—there is no meaningful way to find a control group of comparable subjects.”