According to two Harvard professors and their collaborators, a 2015 landmark study showing that more than half of all psychology studies cannot be replicated is actually wrong.