# Should poor-quality questions be deleted in general?

I asked a recent question on this meta about the deletion of a recent thread, and I think the responses there are useful but not quite as conclusive as I'd like, to a large extent because (as I realize now) the question wasn't worded anywhere near as conclusively as it should have. Still, it was very useful in putting the core terms of the debate on the table, but I'd like something more general and stable.

In that regard, Kyle's answer there is the perfect kicking off point - he makes some valid points, but I think it is extremely important that we have a clear community consensus around what does and does not fly for the community moderation on this site.

Thus my core question here is:

Typically, if a question is

• sufficiently downvoted (e.g., $$\leq-3$$)
• closed for at least 1 week (so in actual closed state, not 'on hold')
• unedited since closure

...

does that form sufficient grounds for deletion? if so, what is the basis and rationale for this?

I would ideally like to see the two alternative views on this issue (i.e. yes, they should, and no, they shouldn't) provide answers here, so that the voting on those can give us a better idea of how the broader site community feels about this.

The reason I think this is extremely important is that question deletion is one of the places where community moderation is lacking the most in terms of mechanisms for checks and balances:

• To the bulk of the moderation-active population on this site, for the foreseeable future, if a question is deleted, it drops right out of the map, and its deletion is basically undetectable.
• The analytics on these statistics are absolutely terrible. Even if one does have the reputation to access the stats (i.e. the 10k Tools page), they are very hard to read and extremely hard to use effectively as a tool to audit what questions are and are not getting deleted (example).

I don't think this means that nothing should be deleted, but I do think it means that we need an affirmative site consensus that some questions should be deleted, with clear boundaries on what that class of questions is.

In particular, it also means that statements like

I reckon my standards are higher than average

should be treated with extreme care in this context. Everybody is entitled to their own standards, and for e.g. question closure, that is perfectly fine: if people disagree, they can detect it and vote to counteract it. Since here the latter mechanism is broken, I think we need a tighter ship in terms of standards. In that regard, I think the criteria set out by Kyle and quoted above are a great description of the class of questions involved, but I think we still need a wider agreement that they are actionable.

For full clarity:

• This is a discussion about deletion, not closure. The on-topic-ness of the question involved needs to have already been decided, in the negative, for the considerations in this thread to begin to be applicable.

In particular, if your answer concludes with "such answers should be closed" or "such answers should not be closed", you're not answering the question I'm posing, by a country mile.

If your answer includes "such answers should/shouldn't be deleted", then you're in the right track.

As for my personal take - I don't have strong feelings either way. But deletion is a very delicate issue (particularly for some unrelated classes of questions) and I care about having strong community processes where it's on the line.

• The system automatically deletes questions under certain circumstances, and the moderators generally take the stance that those criteria are good enough that we don't need to use our binding deletion votes on low-quality questions. If non-moderators want to delete such questions by three-person consensus, that's a different matter. (I haven't posted this as an answer because it sounds like you're mainly targeting the latter issue.) – David Z Oct 24 '18 at 19:19
• @DavidZ I think you should post that stance ("the automatic systems do a fine job and don't need much additional help clearing cruft") as an answer so it can be voted on. – Emilio Pisanty Oct 24 '18 at 20:32
• @EmilioPisanty Sure, I'll do that today when I get time. – David Z Oct 24 '18 at 20:41
• @EmilioPisanty I think that DavidZ is saying that the diamond moderators tend to trust the roomba, rather than unilaterally deleting. That is different from suggesting that the broader community should trust the roomba. – rob Oct 25 '18 at 13:17
• @rob Sure. I'd still like to see that position argued for, though (maybe in the form "this is what the mods do, let's see what folks think in terms of whether that should be a broader policy"?) so it can be voted on. – Emilio Pisanty Oct 25 '18 at 13:23
• While I agree that such questions should be deleted, it isn't fair for the users who have spent the effort to post answers that were sufficiently voted up to lose their earned reputation as a consequence. The main difficulty here is the fact that users' reputation are tied to the existence of their posts instead of the users themselves. – user7777777 Oct 25 '18 at 15:39
• @user7777777 There's a strong argument to be made that if folks are answering absolutely-terrible questions without also fixing them into on-topic forms at the same time, then their claims to holding on to those answers are significantly weakened. – Emilio Pisanty Oct 25 '18 at 15:44
• @Emilio Pisanty: Yes, on the other hand, it can also be argued that it's unfair for simple answers to this kind of question to "easily" gain so many up votes and reputation. In other words, we shouldn't be encouraging new users to view answering simple low-quality questions as a quick way of gaining reputation (and therefore promote such acts). However, the actual worth of the post is largely a matter of personal judgement. – user7777777 Oct 25 '18 at 15:57
• @EmilioPisanty Done, sorry that took so long - I was occupied with some other things yesterday. – David Z Oct 26 '18 at 6:07
• It is a annoying. Now this one, where I had added an answer: physics.stackexchange.com/questions/437196/… – Pieter Oct 28 '18 at 8:53
• @Pieter That's a system deletion, not a 10k+ user-initiated one. I don't understand that deletion and it looks to me like it shouldn't have happened - I'll investigate it in more detail if I can find the time. – Emilio Pisanty Oct 28 '18 at 12:15
• @Pieter Ah, got it. That post was removed via reason 11 in the MSE faq thread (explained in more detail here) - its owner user was removed, and it was removed because it had negative score and neither of its answers had upvotes (and, in addition, it was closed). That channel leaves a clear trace on the timeline (screenshot) but you do need 10k+ to access it. In any case - it's not the class of deletions I'm addressing here. – Emilio Pisanty Oct 29 '18 at 11:40
• @Pieter No worries. – Emilio Pisanty Oct 29 '18 at 12:24
• Something loosely related regarding users loosing rep after questions they answered are deleted is discussed in the comment thread under this Meta SE answer. – user191954 Nov 18 '18 at 12:46

## 3 Answers

The system automatically deletes questions under certain circumstances (using a process semi-officially known as "the roomba"). In this context, the most relevant criteria under which questions are deleted are essentially these:

• The question has been closed and not edited for 9 days, has no pending reopen votes and net score zero or less, all answers have net score zero or less, and there is no accepted answer
• The question has negative score and no answers and is at least 30 days old
• The question has zero score (or a score of one if the user's account is deleted), less than two comments, less than 1.5 average views per day, and is 365 days old

The (diamond) moderators generally take the stance that the roomba, using this and other more niche criteria, does a decent job of keeping the site clean under most circumstances. For that reason, we generally don't use our binding deletion votes on low-quality questions; we figure that if the question is low quality enough that it would deserve deletion by moderator fiat, it should meet one of these criteria eventually and get automatically deleted anyway. And it's quite rare that we think a question is harmful enough that we can't wait the 9/30/365/whatever days until the roomba gets it. (This doesn't preclude deleting questions for reasons other than being low quality. That we still do.)

Now, keep in mind that this is how the diamond mods see things. Non-moderators with the power to cast delete votes don't have to take this approach; they can definitely be more proactive in voting to delete if they want to. We rely on the fact that a post has to receive a consensus of three community delete votes to actually be removed to prevent excessive purging of posts from the site. Nevertheless, the community could choose to adopt a similar approach to the diamond mods, so I'm posting it as an option to be voted on appropriately.

• Exactly this. Why spend precious mod / reviewer time where machines can do the work? Unless someone enjoys the process that is. – Dmitry Grigoryev Oct 26 '18 at 7:55
• @DmitryGrigoryev because machines are slow? because I'm bored at work & clicking "delete" takes all of half a second? because cleaning the site of low-quality work is actually a good thing and therefore there is some satisfaction in helping the site become better? Do you want me to go on? – Kyle Kanos Oct 26 '18 at 15:21
• @KyleKanos 1) Machines are fast, at least software algorithms are faster than you and me. If the automatic system delays deletion for 9, 30, or 365 days, then I'm guessing there's a reason behind it, maybe to give time to the involved people to make the necessary changes. 2) Hopefully it takes you more than half a second to actually read and understand the questions that you are deleting, i.e. deleting a question just because it has a few downvotes does not qualify as good moderation practices IMO. – user190081 Oct 26 '18 at 16:01
• @KyleKanos 3) As anna v commented on recent discussion: Low quality posts may reflect an average non physicist's ignorance, there is value on leaving some of these questions on the site as they may educate future people on their own misconceptions. – user190081 Oct 26 '18 at 16:01
• @user190081 reasons that can be circumvented by those with >10k rep by the site's design. Why is it that I am required to read & understand a terrible question (deemed as such by the community) but a machine isn't? That sounds speciesist. If the post fails to live up to community standards, it should be deleted regardless if whatever it says. I'd say that letting any question by without care of standards is worse moderation practice that one could possibly have (because it's basically no moderation). – Kyle Kanos Oct 26 '18 at 16:06
• @user190081 and if you really don't like that questions (and answers) can be deleted, you should post your concerns on mother meta (and see how many downvotes you accrue), rather than as comments on posts (which can't accrue rep or be subject to downvotes) – Kyle Kanos Oct 26 '18 at 16:08
• @KyleKanos "Why is it that I am required to read & understand a terrible question but a machine isn't?" - Because the machine is not deleting the question by evaluating the question itself, it uses the feedback of the community (as explained in the criteria posted by David Z) to decide to delete a question. You, on the other hand, are expected to use your personal criteria to issue such deletion. – user190081 Oct 26 '18 at 16:24
• @KyleKanos Also, I have never implied that we should let any question without care, IMO if a question can be fixed, then help to fix it, if it is useful to the community (even if it is low quality) then do not delete it. I dont know here are you getting that I do not like questions being deleted, I understand that deletion is necessary to keep the site clean. What I dont like is deletion under arbitrary and poorly-based standards such as 3 downvotes without even looking at the question. – user190081 Oct 26 '18 at 16:25
• @user190081 I'm not sure how I'm not using feedback from the community either: score <=-3 & closed & unedited since closure is not something I can do singlehandedly. I'm not voting based on the question itself, but in the community's response to it. – Kyle Kanos Oct 26 '18 at 16:27
• And these are not arbitrary standards either, a bit more strict than the Roomba in some ways. At -3, the question is removed from the main view so you have to go looking for it (e.g., questions tab), so anything at it below that is decidedly bad by global SE opinion. If it's closed, then the community has decided it's a bad question. If OP or others have not tried improving it, it's abandoned. Hence, it should be subject to deletion. If you don't want low quality questions deleted, then improve the low quality posts & I'm more than happy to keep it around. – Kyle Kanos Oct 26 '18 at 16:31
• @KyleKanos Wait a minute, I think I misunderstood your answer. What I did't like about the deletion criterion that you posted before was that, as I understood it, you were deleting questions that had: -3 or closed or unedited since closure. I have concerns specially the 3 downvotes one. I have no trouble at all is what you mean is -3 and closed and unedited since closure. Is this the case? – user190081 Oct 26 '18 at 16:33
• @user190081 yes, these are and conditionals. – Kyle Kanos Oct 26 '18 at 16:35
• @KyleKanos Then I must apologize for the misunderstanding. I still think it can be argued that even if a question is already closed, and the author shows no interest in fixing it, but it is useful for future users, it might be a good idea not to delete it. I believe that these questions will come time and time again since they are usually either common misconceptions or questions that naturally arise when one is learning physics and does not have enough knowledge to see a more complete picture of a particular problem. – user190081 Oct 26 '18 at 16:42
• @user190081 no apologies necessary. I probably could/should make the post more clear about the and conditionals. – Kyle Kanos Oct 26 '18 at 16:45
• What is this "roomba" that you speak of? – Peter Mortensen Oct 27 '18 at 14:17

Some poor quality questions should be systematically deleted. Examples include and this one:

Other questions that should be deleted are closed questions based on false or nonsensical premises. An extreme example would be

I understand that pigs can fly.

I understand that all pigs are the same.

Assuming this holds how can one keep pigs in a pen?

Such questions should not be answered. Period.

My position on this is broadly in line with Russell's teapot argument: it is for posters to make sure the questions are sensible, not for answerers to debunk incorrect claims or personal theories by posters.

In particular, suggesting there is a contradiction - and we've seen plenty such questions especially in the context of special relativity - without proper research on the part of the poster cannot be constructive to the site. Indeed, in such cases the most appropriate course of action would be not to answer but to insert a comment suggesting a reference/book so the poster can correct his/her misconceptions (although in the case of flying pigs it would be a challenge to find even that!).

While I'm not sure there is some sort of quantitative threshold (in terms of down votes) to be met in general, closure is usually a good hint that something's not right with the original post: there is ample opportunity to correct misconceptions, if only through comments. Sufficient time for a poster to make appropriate corrections should be granted, although sufficient is somewhat vague and may depend on the actual post and the "seniority" of the poster: the two questions liked above should be IMO deleted as soon as feasible.

Finally, sure we learn through mistakes but we have plenty of examples of closed questions for people to assess what makes a good or a bad question: we can afford to purge a great many duplicates and poorly received questions.

• Hmmm. Your links to examples of poor questions are broken. Does that make this a bad answer? :) – Bill N Oct 30 '18 at 18:57
• @BillN the questions have since been deleted. Let me fix this as best I can. – ZeroTheHero Oct 30 '18 at 19:32
• I figured that's what happened. Just poking at the answer in jest. – Bill N Oct 30 '18 at 21:15
• Both of the examples you've provided were deleted by the roomba, so in a sense the data you've provided lend more support to the position that we should just let it do its job and keep user-driven deletions down. I imagine that's not your position, but it would be good to make this answer more specific to the types of situations where the roomba's cleanups are not sufficient and 10k+'ers should step in. – Emilio Pisanty Nov 3 '18 at 14:43
• @EmilioPisanty the observation that both of the specific examples above were deleted by roomba rather than users speaks to the need for users to become more involved more quickly so as to pare down the number questions harmful or not actively useful to the site. The site remains a treasure of good questions and good answers, but I’m afraid the rate of accumulation of bad/duplicate/erroneous question is greater than the rate for good questions. – ZeroTheHero Nov 3 '18 at 20:57
• @ZeroTheHero The posts you linked to were deleted by the system within nine days and they were removed from front-page visibility within one day; since both the SE recommendation and search engines as well as external searches do some heavy punishing to low-score posts, the link will have been shown to a really minuscule population. I really don't see how a window that short can be considered harmful enough to call for an active campaign against that class of questions. – Emilio Pisanty Nov 6 '18 at 11:08
• @EmilioPisanty I actually initiated these deletion: I do not know if this has any effect on roomba. Regardless, I don't rely on roomba to get rid of obviously lame questions. I think we can both agree these kinds of questions will be deleted: whether it's quickly enough or not is secondary. It seems the core issue (or at least what drove your initial post) is with not-so-extreme examples. I actually think there are "hypothetical" questions that harm the site, especially those based on false premises, and those should be removed, but I'm only one vote. – ZeroTheHero Nov 6 '18 at 13:04
• There are two delete votes on both of your links but they are unrelated to the roomba's action - that would have deleted both questions regardless of any existing user-cast delete votes. – Emilio Pisanty Nov 6 '18 at 16:45
• I completely disagree that "whether it's quickly enough or not is secondary" - the whole point is whether user-initiated deletions are sanctioned by the community despite the fact that they're invisible and extremely hard to audit and to reverse. The only argument for user-initiated removal of questions like the ones you've linked is that they're so harmful within their system-determined nine-day lifetime that they require user-initiated deletion asap, but there's no such argument in your current answer. – Emilio Pisanty Nov 6 '18 at 16:47
• But yes: the questions you've linked are not very representative of the class of questions I intended to treat in this thread (which is primarily focused on stuff that wouldn't be removed by the system in the first place), and I'm kind of annoyed at that unmarked derailing of the conversation. I don't really see the relevance of your current links to the primary purpose of this thread, and as I asked in my first comment, it would be good if you could link to examples which are more relevant and provide clear arguments for why those should be deleted too. – Emilio Pisanty Nov 6 '18 at 16:50
• @EmilioPisanty Sorry mon. I'm totally swamped with multiple urgent requests arriving in the last 3 days. I don't want you to think I'm ignoring your comment. Not sure when I can reply though. – ZeroTheHero Nov 7 '18 at 14:14
• @Zero I don't see this conversation as time-critical. It needs to happen, but it doesn't need to happen urgently. – Emilio Pisanty Nov 7 '18 at 15:55
• I’ve edited my answer so the prominence of outlier questions is not so great: they are part of the types of questions that should certainly be deleted, but IMO that’s not enough. I’m afraid I find there’s a huge number of questions that are not useful to the site: as a resource, the site needs to be easily searchable, and in this way obvious duplications, misinformed questions etc should be IMO eliminated. Examples include posters claiming to have found an inconsistency in this or that derivation, variations on twin paradox, parallel/series circuit questions – ZeroTheHero Nov 24 '18 at 15:49
• ...It’s unlikely that much more can be said of the latter two types: posting such questions reflects lack of prior on the part of the OP; such questions are typically not eliminated by roomba but still clog the system. – ZeroTheHero Nov 24 '18 at 15:49

Under the qualities listed above, I don't think there is really any question (hah!) that such questions should be deleted, eliminated, terminated, however one wants to phrase getting them off our books.

1. The down vote total suggest it was not well received by the community. So, why should it be retained?

2. It is unedited, so the OP hasn't bothered to try and respond/fix/reword/whatever. If they don't care, neither should we.

3. After a week nobody else cared enough to step in and try and rescue it. So, the question is pretty conclusively not considered useful to anyone.

Taken together, there really isn't a good reason such questions should stay around in any form. They add nothing of any value

• @EmilioPisanty - you know, I could try and claim a lack of coffee, but I have actually drunk a lot of it today... Yes, I will edit to note they should be deleted. They are just random flotsam and jetsam on the site. – Jon Custer Oct 24 '18 at 18:26
• Thanks for editing. You're precisely in the population (moderation-active <10k) I was hoping to hear from. – Emilio Pisanty Oct 24 '18 at 18:47
• @EmilioPisanty - uh-oh, now I'm afraid of becoming part of a target group for pop-up ads: "Not yet a 10k user on Physics SE? Take our online course on effective answer writing for only \$99.99 plus tax!". "10 Top Tips for rep enhancement!". Etc... (add smiley face here). – Jon Custer Oct 24 '18 at 22:47