24
$\begingroup$

Currently, it takes 5 close votes from users with the vote-to-close privilege ($\ge3000$ reputation), or 1 vote from a moderator, to close a question. If the question is being closed as a duplicate of an existing thread, it can also be closed by a single user with a gold tag badge on one of the tags on the question, if the OP had put that tag on the question originally. On the other hand, if there are 3 leave open votes on a question in the queue, it will not be closed and will exit the queue. However, right now (6/25/20) we have 47 threads sitting the the close vote queue. That seems to be typical. Over the past few years, the queue has occasionally gotten into triple digits; it has not gotten to 0 and stayed there long enough to be recorded. If neither voting threshold is reached, the thread will 'age out' of the close vote queue after 4 or 14 days (depending on whether the question has $\ge 100$ or $<100$ views, respectively). I believe there have been many threads that have been unresolved and simply aged away, but I don't know the number. To the extent that threads have been closed or left open consciously, it has been primarily due to action by voters with the power to unilaterally resolve the issue. To a large extent, threads have been resolved due to the efforts of Peter Flom, who has completed more than 15 thousand reviews (thanks, Peter!).

This state of affairs seems suboptimal to me: Many threads are not resolved, those that are often take too long to reach resolution, and too much of the burden and results have been shoulders of too few. I am now wondering how we should move forward. I see a couple options:

  1. We could continue with the status quo.
    This does not seem like a great option to me, but I don't know if others share that opinion.
  2. We could try to somehow get users with sufficient reputation to engage with the close vote queue more.
    How?
  3. A third possibility is to petition SE to lower the threshold for closing questions from 5 votes (presumably to 3).
    Stack Overflow did this, and while their queue remains very long, it did improve. Another 14 of the smaller SE sites (I don't have the list) have had this policy change. I gather they have had improvements as well. One thing to note about this option is that, if we elect to do this, it is unlikely that the change will be implemented very quickly; the SE staff who do this sort of thing has been reduced, even though the network has grown. Nonetheless, it can be done, if we choose to move forward.

Update:
Something to bear in mind is that we are discussing a very specific proposed change: Lowering the threshold for closing questions from 5 to 3 (I don't believe that even changing from 5 to 4 is an option). This change is on the table because it is something SE has decided they will do (eventually) for sites that want it. It may be that some other changes will come along with this (e.g., @StephanKolassa's observation that the number of close votes a user can cast was increased on SO), but even that is speculative. People have proffered lots of other ideas that may help if implemented in addition to or in place of changing the threshold, but those would ultimately have to be the subject of other discussions on meta.SE, approved by SE's hierarchy, assigned to developers' queues, etc. It is fine to spitball additional options here, but we need to be clear on what we're discussing and what changes are realistic.

| |
$\endgroup$
  • 2
    $\begingroup$ I found this discussion on meta.SO interesting. meta.stackoverflow.com/questions/394048/… I don't know if this has gained traction with corporate SE but it seems necessary to incentivize greater participation in the queues. $\endgroup$ – Sycorax Jun 25 at 21:01
  • 1
    $\begingroup$ Math.SE has this problem much worse, largely due to far lower question quality combined with far higher question volumes. The Close queue has been near 400 for months, now, and it's staggering how many close votes are cast every day - well into the hundreds. Of interest to me would be the mean time to leave the queue, either closed or kept open or aged out. $\endgroup$ – Adrian Keister Jun 26 at 17:55
  • 2
    $\begingroup$ I'd like to know that myself, @AdrianKeister, but I'm not sure how to get it from the SE data dump. I don't see an existing query for it. I tried to see if I could fork a related query, but I couldn't figure out where that information is stored in the database exactly. My SQL-fu is somewhat sub-par, though, maybe someone else can figure it out. $\endgroup$ – gung - Reinstate Monica Jun 26 at 18:02
  • 2
    $\begingroup$ Aging out after only four days is too fast. That must be increased, in addition to whatever else is done. $\endgroup$ – kjetil b halvorsen Jun 26 at 18:59
  • 1
    $\begingroup$ To me the boundary for the vote-to-close privilege also seems on the high side. Perhaps we should consider lowering that a bit? Admittedly I am only an occasional user and can usually only answer simple questions so gathering rep is also not all that fast. If the bar for privileges is so high I can never help. Then again as I said I may be biased. $\endgroup$ – Maarten Punt Jun 27 at 15:26
  • $\begingroup$ Linking related discussions on SO; the following was the final result; previous discussion were closed as duplicates from this one: meta.stackoverflow.com/questions/391832/… $\endgroup$ – Andre Silva Jun 27 at 22:18
  • $\begingroup$ Maybe your problem is, at least in part, due to your (plural!) being too eager to cast "close" votes? $\endgroup$ – Igor F. Jun 30 at 16:14
  • 5
    $\begingroup$ @Igor That's a possible explanation, but it doesn't stand up to the record, which shows that the community (with 5 votes) or the mods have supported the vast majority of the proposed closures. Reviewing 100 threads to keep 15 open is a common experience. $\endgroup$ – whuber Jul 1 at 19:23
  • 1
    $\begingroup$ @whuber: Two points: 1) Are these community votes independent events? As a statistician, you are certainly aware of the repercussions... 2) If there is such a high agreement, where does the problem stem from? Too few users with vote-to-close privilege? If so, why not lower the threshold reputation and have more such users? $\endgroup$ – Igor F. Jul 2 at 7:04
  • $\begingroup$ @IgorF., lower the threshold to what value? And why 'that' value? Voting events are not independent, but why do they should? For example, do you think an ignored comment by OP from someone requesting improvements in Q who cast a closure vote should be ignored by the following reviewer? $\endgroup$ – Andre Silva Jul 2 at 16:13
  • 3
    $\begingroup$ @IgorF., bear in mind that this is an option for discussion because this is something SE has decided it will do. There many be many other changes we could dream up that might work better for some reason, but they aren't going to happen. There are plenty of people with the close vote privilege, but many don't work through the queue regularly. FWIW, our close rates are not out of line with other top SE sites. If you want to make a positive case that our close rate should be some other amount, you should start a new meta thread to discuss the issue. $\endgroup$ – gung - Reinstate Monica Jul 2 at 16:38
  • 2
    $\begingroup$ I mainly made the earlier comment because the threshold for the privilege is actually lower at economics.se so I figured it would be possible here as well. Then again economics is in beta and has fewer users so perhaps that's the reason. $\endgroup$ – Maarten Punt Jul 2 at 18:21
  • $\begingroup$ There must be the time now to close this and implement the proposed change! I have seen little disagreement here. And the queues are growing ... $\endgroup$ – kjetil b halvorsen Jul 24 at 16:22
  • 2
    $\begingroup$ There's no need to close this Q, @kjetilbhalvorsen. The SE community mods have been notified that CV has consensus on the issue & that we should be added to the list of sites where this should be implemented. From here, it's however long it takes the developers to get to it. Unfortunately, there's no telling when it will happen. $\endgroup$ – gung - Reinstate Monica Jul 24 at 16:26
25
$\begingroup$

I'm all in favor of option 3: ask SE to lower the threshold for closing questions from 5 to 3 votes. (Importantly, the same should apply to the threshold for reopening a question. This is also the case at SO.)

In addition, at SO, qualified users of 3000 rep minimum can cast up to 50 close or reopen votes per day. Here, we only get 24 close (and presumably reopen) votes per day. So I would also propose to raise these limits to 50 per day.

My personal impression is that a large majority of questions that need closing are really clear-cut cases. We don't need a consensus of five sufficiently experienced users. Three is plenty. And if something gets closed that shouldn't, then the same logic applies. I can't recall a question where a larger consensus would have been useful, and if so, we can always open a question on Meta.

On the other hand, I don't see a good way to increase participation in the review queues, although that would certainly be a worthwhile goal. I'm as guilty as the next user of not looking at them enough, simply because it's a thankless job. By definition, you don't see the kind of interesting and fun questions that draw you to CV, but the exact opposite.

So, as long as it's a thankless and boring job, let's at least increase our effectiveness by 67% (if 3 votes can close a question instead of requiring 5, then each vote is $\frac{5}{3}$ as effective).

| |
$\endgroup$
  • 10
    $\begingroup$ I agree strongly. It's important detail that (1) we are down one moderator as compared with the period before recent troubles. My further guesses include (2) some measure of dissatisfaction and so lesser engagement with CV among several experienced users (3) many of us tend to vote to close directly on questions of interest to us, and looking in this queue for more poor questions is not enticing. I sense beyond SE an increasing frequency of questions assuming that the poster is entitled to an answer, regardless of how little effort they put into a question -- or in working it out themselves. $\endgroup$ – Nick Cox Jun 26 at 12:54
  • 5
    $\begingroup$ These are good points. I hadn't thought of the number of possible votes as a factor. Something that occurs to me is that I think an ordinary user can only review 20 threads per day in the review queue, even if they get 24 votes. So it might be worth pairing increasing the number of reviews you can do in step with increasing the allowed number of votes to 50. $\endgroup$ – gung - Reinstate Monica Jun 26 at 12:56
  • 2
    $\begingroup$ I agree with this but just one extra point. Although I often hit the 20 close votes barrier there are seldom many left for me to do at that point so increasing my number would have less effect than reducing the threshold. Mind you I hardly ever see one in the queue with more than one vote already but that may reflect my time zone (currently UTC+1) versus that of other users who close a lot. $\endgroup$ – mdewey Jun 28 at 12:45
  • $\begingroup$ Anything new on the request to SE about lowering the threshold to 3 votes? $\endgroup$ – Stephan Kolassa Oct 8 at 7:29

You must log in to answer this question.

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