Your model fails to admit the possibility of researchers who do better or worse work.