Sometimes flags are applied to chat messages as a joke. This is a problem because every 10k rep user on the network then sees the flagged message and wastes time reading it until someone declines it. Similarly it is a problem when a flag is applied to an innocent chat message as an attack on the person.

Would it be useful to have an additional option when presented with a flagged message, to flag the flag itself as a problem? This would be separate from marking the flag as invalid, which could simply mean that the flagger has a different opinion on what should be considered offensive. The button to flag the flag would be only for cases where the flagger is believed to be deliberately abusing the flag system. This gives a way of detecting abuse without needing to divulge the identities of flaggers.

This button would be available to all 10k rep users (everyone who is involved in validating chat flags). What would be a clear unambiguous way of wording this?

It might also be worth considering where this additional flag should be sent. Direct to community managers, to site moderators, or to all 10k rep users for review?

Will it be a flag of the same type as a chat flag (and hence with the option of flagging it for abuse in the same way), or just a record in a database for SE to monitor over time?

If it's a flag that gets reviewed like any other, then if the flag abuse comes from a 10k user, that user should not be able to validate or invalidate the "flag flag". Presumably there is already a mechanism in place to prevent 10k users from reviewing flags on their own posts, so the same approach could be taken to prevent people reviewing flags on their own flags.

  • Related: meta.stackexchange.com/questions/266147/… Commented Apr 20, 2017 at 21:57
  • 3
    How would I flag malicious flag flags?
    – Jason C
    Commented Apr 21, 2017 at 1:26
  • @JasonC I think the penultimate paragraph covers that, although it doesn't sound like this will be introduced - see the answers for why we're better off without it. Commented Apr 21, 2017 at 16:25
  • He's talking about a third layer of flags... presumably a joke.
    – Catija
    Commented Apr 21, 2017 at 16:41
  • Indeed I was being facetious about flagging the flags that flagged the flags.
    – Jason C
    Commented Apr 21, 2017 at 17:22
  • No I got that it was a joke, but the penultimate paragraph really is referring to being able to flag flags of flags to arbitrary depth, which would automatically be possible if the flag flag had the same type as a flag (same dialogue when reviewed). Commented Apr 21, 2017 at 23:18
  • "able to flag flags of flags to arbitrary depth" ...that could be fun
    – Cai
    Commented Apr 22, 2017 at 9:27

3 Answers 3


I'm not a fan of this. I respect and appreciate where it's coming from but I really think it opens a can of worms that we don't really want.

The problem I have with this is that it's not always possible (if ever possible) to know that there's malice in a chat flag... particularly since you (except in very specific circumstances) have no clue who cast the flag in the first place. Unless the user submits a mod-only flag, we have no way to know why they flagged the message at all because they have no way to explain the flag. And, honestly, even if they do use the mod-only flag, they may not really explain their reasoning.

Our goal as moderators (and, yes, 10K users in chat count as moderators) is to give people the benefit of the doubt and assume that they are acting in good faith. This button would tell every 10K+ user to doubt that and encourage them to judge with only a minimal amount of information why someone is choosing to flag something. And just because you say

The button to flag the flag would be only for cases where the flagger is believed to be deliberately abusing the flag system.

doesn't mean that's how it would actually be used. People misuse just about anything they're given. If they think a flag is dumb and they want retribution on the person who flagged it, they'll click that button without thinking... and then we're invalidating flags on flags and it's turtles all the way down.

If there's really an issue with irresponsible flagging in a room, that needs to be brought up for the site mods and CMs to address more directly rather than with each flag that gets raised.

And, if my understanding is correct, the only people who could adequately even come close to adjudicating these flags on flags are the CMs and other Stack Exchange staff who have access to who flagged what.

In the end - people complain about chat flags constantly... let's not make more flags for them to complain about.


I think this kind of ignores the core issues.

We're trying to find a purely technical issue to something that's a social/cultural issue. This is roughly the equivalent of a bunch of children running around, pulling fire alarms.

It would be nice if we could remind folks flags are for serious issues, and they're making things harder for when the flags are actually needed. Its a hard slog, but there's some previous "problem" rooms that have made a great improvement on how they behave.

It feels like chat moderation isn't something that scales horizontally that well. We have a lot more people who could moderate, but it may also mean the best person to handle it may not. We do ok. Most folk think before they act.

And now.. I'm going to suggest a technical solution. Typically something that's clearly flaggable gets more than one flag - so hide any chat with less than 2-3 flags to any mods or ROs not currently in the room.

On the broader sense though - the people doing this are clearly doing it to tweak the mods and other users and to try to break the system. I do think that at some point, we need to consider chat flag bans, or kicking users who're abusing flags at some point (automatically) - I like Shog's solution for the chat suspension with Adam's Addendum in the comments but automatic, short (and extending) chat bans for a pattern of flag abuse feels fair - especially since that's the potential consequence of an accepted, bogus chat flag.


  • I was hoping my posting a less than ideal solution would prompt someone to come along with a better one... I like the idea of keeping the flag review to just room-local reviewers up to a certain number of flags. The only problem that comes to mind is that sometimes a room has no local reviewers, in which case I wouldn't want the flags to be left unreviewed. Perhaps local reviewers could be prioritised but distant reviewers be included in their absence? Commented Apr 22, 2017 at 13:22
  • at that point, one would assume 2-3 users would pile in knowing it would get escalated. Commented Apr 22, 2017 at 13:23
  • In a crowded room that should do it, but if there are only 4 people chatting and 2 of them are causing problems, it would be good if the system automatically brought in others rather than everyone needing to know how the system works. Commented Apr 22, 2017 at 13:26
  • Yeah, I was thinking 2 for that reason. The mental leap I have to make is the style of chat moderation I'm used to kinda breaks without at least one or two grownups in the chat. However, excessively complicated logic is bad so... I donno. Commented Apr 22, 2017 at 13:35

First a bit of background…

I never thought flagging every problem was an effective means of guiding and controlling routine annoyances in chat. Flagging is effective way to stop (or clean up) major social disasters when things get wildly inappropriate, but calling in the "authorities" every time some blowhard gets too loud or someone starts slinging around ad hominem attacks doesn't help them not get there in the first place.

enter image description here

I write about this in Crowd-Based Chat Norming. ←That post scratches at the surface of a system where folks can 'poke' a user when they feel things are starting to fall outside what is considered appropriate for the room. It would generate a series of gentle warnings to the recipient… and only escalate [to a timeout or Mod flag?] when the recipient fails to stop or move on.

But you're right…     ←This is where I get to your suggestion

Such a light-weight mechanism can also becomes a tool for a gentle ribbing. For example, using bad puns in our back-office chat rooms can get you "fired!" several times a day in chat — it's quite funny… and harmless. But in a broader public/community context, you need a way to assure such joviality doesn't cause real alarm.

That's what's great about something like a light-weight poking mechanism. Not every poke is meant to call in the authorities. But every signal provides an opportunity to move on… or to respond

report abuse

In the use case you describe above, reports of abuse notifies both parties, and hopefully that should be the end of it. But it provides an extra signal for the system to determine if it was simply a one-off problem now resolved, or if a party "abusing" the feature (on either side) simply does not get it.

Escalation by numbers gives us the ability to know what actually needs to be tackled head on, versus a lot of this cruft that should be resolved internally (read my post linked above).

But this won't work as well with flags (without fundamental changes)

If users see chat flags as a social joke, flagging those flags to report the flags will only multiply the problem. Not every problem needs to call in a Moderator/official. There are social mechanisms at play which run afoul of that "call 911" mentality — and replacing yet another "normal interaction gone wrong" with yet another reason to "tell a moderator to handle it" has not been scalable (or particularly pleasant) for some time now, and it's getting worse.

  • 1
    People say "flagged" in chat all the time without actually flagging anything... I think where this is failing is with users who think saying "flagged" is accompanied by actually doing it... or doing it instead of just saying it. Those joke flags are definitely non-useful and, probably, harmful considering the breadth of people who are notified by chat flags. But I don't know how to thwap someone who's using flags as a joke rather than just saying "flagged" in a chat room.
    – Catija
    Commented Apr 20, 2017 at 21:19
  • @Catija Good point. If saying "flagged" is the social/chat equivalent to the virtual librarian who whispers "shhhh..." in the post I linked above, then an actual poke feature (like I describe above) accomplishes that behind the scenes without everyone else having to see it. It's similar to not commenting -1 or +1 on every post. You don't do it because there are better ways to express the sentiment (voting)... without all the noise or annoyance it adds. It's all about social norming to provide tools to handle our foibles constructively without annoying everyone else in the process. Commented Apr 20, 2017 at 21:24
  • I'm not sure, honestly. I think that sometimes people use it to get others to move on to other topics (though they apparently use "tacos" in Mos and "kiwis" in The Screening Room for that purpose) but I was talking about more the ribbing, as in your puns in the workplace example. No one thinks they're going to get fired over it and it's not really a directive to stop the puns.
    – Catija
    Commented Apr 20, 2017 at 21:47

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .