I am facing the following decision:

1) We can remove a user that violated our rules in the comment section from commenting further. It would be the easiest way of implementing it. I just see the problem that we exclude the user from a discussion without banning him. Our systems and of course also our moderators do make mistakes so I am a little bit scared to do that and to destroy discussions. Also for one minor rule violation it would be quite harsh to block a user from commenting for good...

The good side would be that trolls etc wouldn't be able to continue breaking the rules in the particular thread they chose to do it and try to behave better in the next thread they want to participate in.

2) If we just remove his/her reply we need to write a little bit more code and in this case I fear that the trolling won't stop until we actively ban the user. Here a user could still take part and defend his/her stance and help to create a good discussion in the best case. In the worst case the user can keep trolling/abusing until a mod steps in.

I honestly know of no community that did the 1) approach so I am reluctant to do this.

One of the problems is that I don't have enough information about what could happen in either approach (especially Nr1) and also not the data to look into it. This is why I would like to have some input on what dynamics both could lead to.

The next is that I don't really know what the goal is here. I mean stop abuse in comment sections is one of course. But also enable users to have good, meaningful and also controversial discussions.


You have a user who is commenting inappropriately. If you prevent him from commenting then you solve your problem at the cost of any constructive comments he might have posted. You seem reluctant to do that, but also consider the cost of not blocking him: you're telling your community that it's ok to break your rules. Not only will this user likely continue, but others who know you could have blocked him and didn't will be confused and/or will follow in his footsteps. That's no good.

An approach I have seen work well on Stack Exchange is timed restrictions. A user who asks a lot of bad questions is temporarily blocked from asking more until he fixes some of his existing ones. A user who raises a lot of noisy flags is temporarily blocked from raising more, usually for about a week. A user who has chat messages flagged (and validated) as offensive is kicked out of chat for half an hour. A user who has too many bad reviews is blocked from reviewing more for a couple days (longer if he does it again). The blocks are designed to both protect the site and educate the user about what he did wrong.

The blocks I've described so far are all automatic, built into the system, and are focused on particular functions. A user who's question-blocked can still flag; a user who's chat-blocked can still ask questions; and so on. It sounds like you need a comment-block along those lines.

Stack Exchange moderators can also suspend users temporarily. This prevents all participation and, again, is meant to be temporary. But here's the key point: each successive suspension is for a longer period of time. If a user does something wrong, gets suspended for a week, and comes back and is a productive community member, then great -- we don't hold grudges, and it's water under the bridge. If he doesn't learn and gets suspended again, though, then the next one is typically for a month. And if he doesn't learn from that? Usually the next one is a year.

Progressively lengthening the suspension sends stronger and stronger signals to the user while lowering the rate at which you have to deal with problems he causes. You could have all suspensions be for, say, a week, but then you'll be constantly playing whack-a-mole with persistent trolls. So start small but be willing to raise it if needed. A week for the first suspension is long enough to get the attention of a user who wants to participate productively while being short enough that he can reasonably wait it out.

It sounds like you have permanent comment blocks. Consider instead implementing temporary comment blocks with increasing penalties for repeat offenders. A user who makes one mistake isn't out forever, but a user determined to cause you trouble doesn't require ongoing monitoring and maintenance.

  • one question: Do you count the successive suspensions generally (1st for flags, 2nd for comments, etc) or do you distinguish between the different suspensions. So when I abuse flagging the first time and after this commenting, will I get 2x first time offender temporarily bans to learn each function or will I get treated as a 2nd time offernder? – loiro Nov 4 '16 at 11:39
  • 1
    The blocks that are built into the system (question-blocks, flag-blocks, etc) don't interact with each other. Suspensions (the ones described in the link that block all site activity, not just specific functions) are manually imposed by moderators. They are thus subject to human judgement. Typically moderators follow the week-month-year progression regardless of what specific behaviors were involved, but if mods feel that's unwarranted, they can adjust the times. I once had somebody who'd been suspended twice early on, behaved well for two years, and then acted out; I didn't jump to a year. – Monica Cellio Nov 4 '16 at 14:01

It sounds as if not only that user is misbehaving on that thread, but so are others in response.
In that case consider blocking the entire thread for comments. Leave a notice or a first/last comment explaining why it is blocked.
The notice could contain a request like avoid ... in the future ...

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.