I think this has already been done with r/politics. I've created new reddit accounts to argue with people on there a few times and find them being banned within a few weeks for the most trivial of infractions, simply comments like 'I find it hard to believe someone would express that belief in good faith' get you banned, and as you don't get a ban without a report someone has to be reporting comments that no reasonable person would report.
I came to the conclusion - after a while - that it's a secretly walled garden, that basically 100% of the content is automated, probably by bot-runners aligned with the moderators, and that users that stray into it are intensely surveilled until a plausible reason exists to ban them. There is zero metacommentary allowed and making a post 'about the moderation of this subreddit' gets you banned for metacommentary. You now also aren't allowed to question user activity, so if someone's account history is 100% botlike you get a ban for pointing it out.
There have been government studies on this as well, like "Containment Control for a Social Network with State-Dependent Connectivity" in cooperation with the Air Force Research Laboratory out of Eglin AFB in Florida: https://arxiv.org/pdf/1402.5644.pdf
…which coincidentally was outed as Reddit's "Most-Addicted City" in 2013: https://web.archive.org/web/20160604042751/http://www.reddit...