Carl Colglazier, Nathan TeBlunthuis, and Aaron Shaw in their study, The Effects of Group Sanctions on Participation and Toxicity:

In this study, we investigated the effects of defederation events on the activity levels and toxic posting behavior of accounts in the Fediverse. The results indicate that such events produce asymmetric effects on activity for affected accounts on blocked servers versus those on blocking servers with no increase in toxicity for any groups. The results also highlight the potential of decentralized social networks and their unique mechanisms, such as defederation, in providing communities with tools to manage content moderation and other aspects of online interactions.

It’s a dense read, but could be interesting if you’re into federation. The interesting thing to me is that defederating did change one thing, but maybe not the thing you’d think. The study found that the level of toxicity on a server did not change after defederation happened, and this was true for the server doing the defederation as well as the one that was defederatred. What did change was the level of activity on each server. Things remained constant on the server that did the defederating, but there was a marked drop in activity on the server that was defederated.

There are surely many more variables to look at, but I really found it notable that the overall level of toxic behavior on a platform wasn’t improved by removing bad servers from the mix. I’m not saying that defederation is useless by any means, but I do think it’s worth recognizing that you can’t give a Mastodon server good vibes simply by blocking all the servers you don’t like, that’s something you need to foster in other ways.