Soft Science: My Tech Central Station column about the shoddy science of medicine is up. Judging by this comment, things are worse than I thought:

I was a review editor for a prestigious medical journal. As I had a PhD I was asked to review certain classes of submitted papers. In 90% of the papers there were such serious flaws that the validity of the research was nil. The flaws included: lack of statistical soundness, unfounded assumptions, plain errors in interpreting data etc. In a word - shoddy science. All of the papers were published. I asked the editor what was going on. His reply: 'If we turned away papers based on shoddy work we would have none to publish at all!' I resigned.

I'm afraid that might be true. Doctors don't get a lot of training in scientific principles or statistics in medical school. When it comes to basic science, our training is largely memorization. Unfortunately, this ends up being the case in a lot of premed curricula, too. And when medical students grow up to be medical researchers they don't analyze their results themselves. They give their data to a statistician to interpret. We could do a lot better.