NPR's Richard Harris on the dangers of sloppy science