The Real Science Crisis
The scientific method is the crown jewel of science. But is modern research on shaky ground?
If you’re a statistician, you don’t have that many ways to get noticed. Usually, your opinion is couched in careful language and framed in mathematical abstractions like confidence intervals and p-values. You don’t talk about event A causing result B, but about the chance that event A might have caused result B.
But if you want to step out of the shadows for a moment, there’s no bolder move than joining 800 of your colleagues and signing your name on a blistering critique of sloppy experimental reasoning. Especially if that critique is published world’s leading science publication, Nature. And especially if that critique is aimed at researchers’ favorite way to separate successful experiments from dumpster-fire failures.
This is the controversy that broke out in Nature’s hallowed pages on March 20. The statisticians’ letter didn’t mince words:
“We agree, and call for the entire concept of statistical significance to be abandoned.”¹
Statistical significance is a way of comparing experimental results against pure chance. It’s the first step in a chain of reasoning that allows scientists to…
