Scientific journals prefer to publish positive results and this can lead to false claims ending up being regarded as true facts, finds new research from the Niels Bohr Institute.