Most published research is wrong because scientists gather data first and then mine it for patterns — like drawing a bullseye around a bullet hole. Cardiovascular drug trials were almost all positive before 2000, then almost all negative after, not because medicine changed but because funding agencies finally required pre-registered predictions.