Why Science is Often Bad
A guide, for the rest of us.
The Frontier Psychiatrists is a daily newsletter for general audiences. It covers sophisticated material but is not written with only a specialist audience of scientists in mind. I read a lot of science while preparing this newsletter for my readers. It won't come as a surprise, but a lot of science? It's bad. Much of it is beyond bad writing—that is, the published results will not turn out to replicate! It’s factually false. Much also falls apart on subsequent analysis—I'm looking at you, STAR*D1. There are understandable reasons for this that are deeply human. Scientific journals have a massive bias towards publishing interesting or exciting things.
“New intervention does nothing different than previous interventions” would be the title of papers people are bothered to publish. We should wish for those things to be published so scientists can access the information! My “journal of negative results and duplications of prior work” will do very poorly, when it comes to an im…