You should re-read it. It offers that as a possibility, with terms like "may" and "could have". It does not conclude it did happen; it notes that some users migrated, but that's not at all surprising.
The hardcore folks are likely to always wind up somewhere, but driving them off Reddit likely makes it more difficult to recruit less initially strident users. As the study indicates, "the ban worked for Reddit".
It has the effect of preventing violent content from using adjacent nonviolent content (i.e. baby and pet pictures, nonviolent political advocacy) on the mainstream platforms as a kind of "human shield".
Of course, this is a disappointment for anyone who finds the normalization of violence-provoking content desirable.
The problem is that there already exist plenty of forums on the internet for non mainstream but non violent speech. The only speech seeking a home is the violent speech.
One of the conclusions of a recent purge by Reddit was that it just pushed the banned users into even more radical spaces online[2].
Of course we all know about The Streisand effect, and one article suggests that censorship just draws more attention to the banned content[3]. If we assume that ideas are somehow "contagious" or "infectious"[4] then we're just exposing people to them even more.
> In any case, the only news on the topic that I’ve seen has been that election misinformation on Facebook/Twitter has dropped by 70% since the deplatforming happened on those platforms.
But the deplatforming didn't make those people go away, it pushed them to platforms like Gab and Parler, right-wing echo chambers. This is like an extreme version of a filter bubble. Remember, millions of people supported what happened at the capitol[5], and there is zero hope of de-radicalising people if the left and right aren't talking. If anything, both sides will get more extreme.
[0] https://en.wikipedia.org/wiki/Deplatforming#Twitter (see references)
[1]https://www.wired.co.uk/article/twitter-political-account-ba...
[2]http://comp.social.gatech.edu/papers/cscw18-chand-hate.pdf
[3] https://ro.uow.edu.au/cgi/viewcontent.cgi?referer=https://th...
[4]https://www.nature.com/articles/s41599-020-00546-3
[5]https://www.statista.com/chart/23886/capitol-riot-approval/