That's the point of the study. The concern was that by banning particular subreddits, they'd simple disperse the activity on those groups across the site. But that's not what happened, if you study the data site-wide.