I would expect the distribution of inconsistencies and errors to be the same for both sets of topics, in the absence of any other information.