When reviewing first posts in particular, it is my understanding that one should check if the question is a fit for the site, among which if the question makes sufficient sense. I take these tasks fairly seriously, and have recently been confused. What constitutes "low quality"?

To make this concrete, the 2 cases that prompt me to look for guidance. I flagged these 2 for low quality, and add the feedback received after my flag was declined:

vector operations between a two-dimensional and a three-dimensional vectors ("flags should only be used if a question demands moderator attention" - I would understand if a moderator thinks it is "not low quality" (although, how, I do not know, but one can disagree); but how am I supposed to understand this?)

https://math.stackexchange.com/questions/318792/a-characterization-of-probability-theory ("moderator found no evidence to support your flag" - question was closed shortly thereafter, probably as it is not an actionable question, which is why I flagged).

As said, I take my flags very seriously, and don't want to be on the record as someone who flags many unjustified posts. Reviewing first posts has an answer box "low quality", which is a subjective criterion which I took as "it's low quality, alright, flag, then let a moderator decide what to do" - but to at least bring it to a moderator's attention I had thought to be encouraged.

I happened to visit the StackOverflow voting page today, in which candidates underline their number of flags as evidence for how much they care. To understand this properly, I do not consider flagging as preparing a future candidacy (because that will never happen); but it seems at SO flagging is considered a good thing, whereas I cannot help feeling that the most recent feedback sounded as if I was told to not bother the moderators.

It seems to tell me that only absurdly bad posts should be flagged (or spam maybe). I feel disinclined to ever flag 'low quality' again, and lean towards just waving through new posts if it just leads to me having a bad flagging record.

  • $\begingroup$ The helpful flag parameter in the voting page is automatic, actually. The software puts it there. $\endgroup$
    – Asaf Karagila Mod
    Commented Mar 5, 2013 at 6:37
  • $\begingroup$ @AsafKaragila: I do not understand what you are referring to? $\endgroup$ Commented Mar 5, 2013 at 6:38
  • $\begingroup$ Second to last paragraph. $\endgroup$
    – Asaf Karagila Mod
    Commented Mar 5, 2013 at 7:42
  • $\begingroup$ What exactly should be wrong with the probability question? $\endgroup$ Commented Mar 5, 2013 at 10:24
  • $\begingroup$ This is too short to really be an answer, so I'm posting as a comment: Perhaps the moderators want you to flag as "not a real question" or one of the other non-"needs mod attention" (which, to my understanding, are also visible by 10K users) instead of throwing a flag only the mods can see. $\endgroup$
    – apnorton
    Commented Mar 5, 2013 at 14:40
  • $\begingroup$ @MichaelGreinecker: It is a question that either has as answer "I agree", or it only serves to start debate on a largely subjective matter. I have seen questions like this being closed frequently (see also QY's comment, which I read to mean something like this). Formerly, I flagged "low quality" to alert those who can vote to close to a question they might vote on to close (as it happened shortly thereafter). "Decline" was for me a response for an untrainable such notice; to not actually agree and not close unilaterally was never the point for me. $\endgroup$ Commented Mar 9, 2013 at 19:52
  • $\begingroup$ Unreasonable (typo: untrainable). $\endgroup$ Commented Mar 9, 2013 at 19:54
  • $\begingroup$ @gnometorule I very much disagree. One can explain that there are approaches to probability that are not based on measure theory, point to discussions on the scope of the measure theoretic approah by people such von Mises and de Finetti, and one can explain why probabilists are in practice mainly interested in the distribtion of random variables. Terry Tao wrote that only concepts invariant under extensions of the measure space deserve to be called "probabilistic".... $\endgroup$ Commented Mar 10, 2013 at 8:22
  • $\begingroup$ ...I don't consider such an answer overly subjective, unless one thinks a question of various ways to formalize a derivative and their underlying motivations to be an overly subjective question. I certainly think that I could have given a useful answer that is not based on my own prejudices but knowledge of the history of probability theory and the discussions around it. $\endgroup$ Commented Mar 10, 2013 at 8:25
  • $\begingroup$ @MichaelGreinecker: And I don't even really disagree. I didn't use to roam this site looking for something to flag; I came across it when I did my 10+ reviews daily of first posts. I actually liked the question on its own merits, but remembered similar questions being closed frequently (implying a minimum consensus of 5). My understanding of doing a solid review was to point those who can close, to posts they might want to close. And the feedback in such cases was almost always 'helpful' - meaning not 'I will close', but 'thanks for letting me know', so feedback here is inconsistent. That $\endgroup$ Commented Mar 10, 2013 at 17:19
  • $\begingroup$ is what my question was about: how providing someone with 10k+ (I think, not only moderators) with information is "not helpful." I flagged more than the average user, but I also reviewed most days my maximum amount (20) of possible first posts, so that is how it should be, in my understanding. $\endgroup$ Commented Mar 10, 2013 at 17:22
  • $\begingroup$ @gnometorule Thank you for the explanation. $\endgroup$ Commented Mar 10, 2013 at 17:23

2 Answers 2


I'm answering this from a network-wide perspective (I'm a moderator on Skeptics and Biology), the attitude towards moderators using their supervotes is markedly different than on most other SE sites. This might affect also the attitude towards which posts should be flagged.

In general, I find the most useful way to think about flagging is to think about what you want to happen to the post you flagged. Usually, you flag for a moderator because you want something to happen to the post that you can't do yourself. If you don't have the power to close questions, you can flag them. If you can't delete posts (or even if you can, because on most sites deletion by the community is very hard to achieve), you can flag.

Flags are not only handled by moderators, flags that request that a question is closed can be seen and handled by the community (10k+ users). If a moderator thinks that the question is at least problematic, but doesn't want to unilaterally decide to close, they can just leave the flag in the queue and let the community handle it.

Another category of flags are custom flags that inform the moderators about problematic user behaviour, because dealing with such issues is something that regular users can't do, as they don't have the necessary tools to do this.

The "very low quality" is by far the most problematic of all flag reasons. It shouldn't be used for somewhat problematic posts, it is meant for really bad and unsalvageable posts. If you think the post is so bad that a moderator should just outright delete it, then a "very low quality" flag would be appropriate.

I personally find that "very low quality" flags are unnecessary in most cases, and you usually should use other, more descriptive flags. For questions, I would flag with one of the close reasons, a very low quality question fits into "not a real question" most of the time. For answers, very often such a low quality answer is not an answer at all, and should be flagged as "not an answer".


I looked at the post about vector operations and could not understand the reason for flagging it. It is a reasonable and natural question, much better than verbatim copies of homework washing up to the shores of Math.SE every day. I wish more of my calculus students asked questions like this (not necessarily at Math.SE, they can ask me).

Because "low quality" is a somewhat diffuse notion, it may help to consider a practical aspect before flagging: what would you do if you were the moderator seeing this flag?

  • Would you close or delete the post at once? (Moderators cannot simply vote to close, their decisions go into effect immediately).
  • Would you suspend the user who asked the question, or destroy their account?
  • Would you migrate the question to another site?

If none of the above actions are called for, perhaps the post indeed does not need moderator's attention. This does not mean it's perfect, but whatever needs to be fixed can be fixed by ordinary users such as yourself. You can edit the post, comment to prompt a clarification from the user, vote to close (not yet, but soon you'll be able to), or simply vote. The first post review queue displays the voting arrows specifically to encourage the reviewers to vote on the newcomers' posts.

  • 1
    $\begingroup$ This sounds reasonable, but it still amounts to discouraging flagging except in the most extreme cases. The second post was closed very soon after I had flagged it, but on my record, I raised an unreasonable flag. But if this is the site understanding of the process, then it is helpful to know to err on the side of not flagging. $\endgroup$ Commented Mar 5, 2013 at 6:01
  • $\begingroup$ As to migrating questions, my conclusion from having asked about it (and from having another decline on my record as I suggested a transfer of a question that still has no feedback here to Cross-Correlated) in meta.math.stackexchange.com/questions/8637/… is that it is discouraged too (there were conflicting answers) as it is disruptive. Just a comment on your third bullet, 5pm, to have this in one place. $\endgroup$ Commented Mar 5, 2013 at 6:12
  • $\begingroup$ @gnometorule I'll try to put it in a more encouraging way: this site is moderated by you. You can always "flag" for your own moderator's attention, meaning that you can decide to take whatever action the post warrants (and your rep level allows you to perform). The second post was closed by five users, but not by a moderator (who obviously saw the post and commented on it). This is typical for Math.SE, where moderators let the users make such decisions. Unilateral closures by moderators also happen, but in somewhat unusual situations. The probability post did not need moderator's action. $\endgroup$
    – user53153
    Commented Mar 5, 2013 at 6:13
  • $\begingroup$ Well, please don't take what I say in the wrong way as it could sound combative, but is not meant so. That's (very good) sophism. :) As 5 users closed the question, it didn't need moderators attention. How can I disagree with that? It's describing a fact. But above you say I should flag what I would close. I would have voted to close (I didn't even dislike the question, but what answer was to come out of it?). As it happens, I just found an old post by Asaf which then stirred some controversy. It's not the same, but I think bottom line is it's subjective, and I'll stay away from flagging. $\endgroup$ Commented Mar 5, 2013 at 6:22
  • $\begingroup$ +1 for answer though as you took your time to reasonably lay out one understanding of the process, which is appreciated. :) $\endgroup$ Commented Mar 5, 2013 at 6:25
  • 1
    $\begingroup$ @gnometorule Glad if my answer clarified something, though it's not any official position. The cleanest solution of this whole thing is to move on, answer another dozen or two of questions, get 3K rep, and you'll find that you don't need the flag button all that much. $\endgroup$
    – user53153
    Commented Mar 5, 2013 at 6:39

You must log in to answer this question.

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