Study shows one way that scientific progress is broken | Ars Technica