The Problem With Scientific Evidence