19
$\begingroup$

I have now experienced it too often to stay silent. People vote to close a question they don't know the answer to. The question itself is clear and easy to answer, but some people prefer to put others off by arbitrarily closing questions because they don't understand them.

Case in point: Bell Lift distribution

We are asked to make the site more inviting to newbies. Then why are we happy with some people closing valid questions because they have the power to do so?

$\endgroup$
  • 5
    $\begingroup$ Why not vote to reopen it? You can do it with a single vote since you have the gold [aircraft-design] badge and you know that it is a clear question. $\endgroup$ – ymb1 Apr 22 '19 at 11:32
  • 7
    $\begingroup$ @ymb1: Yes, you are right and I should do so. So far, I try to avoid using my superpowers. I would prefer if we treat newbies less harshly. $\endgroup$ – Peter Kämpf Apr 22 '19 at 15:02
  • $\begingroup$ Up till today I am member for 39 days, and I got a reputation score of 1120; unhappily here I am ending my participation for the following reasons: indeed I found very interesting questions to which I have answered getting closed arbitrarily, but the worst is the case of answers being arbitrarily suppressed depriving the community the benefits of the answer content. These are the reasons for which these lines put an end to my contribution. Good luck to every body. $\endgroup$ – user40476 Jun 25 '19 at 18:24
  • 1
    $\begingroup$ @user40476: Come on, I had the same happen to me. But you have had well received answers here, should that not be encouragement enough? Please don't give up so easily. $\endgroup$ – Peter Kämpf Jun 25 '19 at 18:33
9
$\begingroup$

I'm generally a little on the users-over-rules side, so it'll be hard for me to give an unbiased answer, but I'll present some research I did.

Data collection

I put together a Data Explorer query to look for questions that were closed then reopened, filter only ones reopened under suspicious circumstances like reopens without edits, and print the list of close votes for each question.

The query can be found here: https://data.stackexchange.com/aviation/query/edit/1037498 . There's a quick summary of count per user here too, but I'd advise you to look at the questions first before jumping to conclusions based on user statistics.

It's not very pretty but without some serious string magic that's the best I can get.

I realize there's some grey area and debate about these criteria for a "suspicious" or "overturned" close vote. We need some debate on both sides about the quality standards on this site, so I'll avoid calling these overturned closure votes "bad" or "incorrect".

These closures merit attention, but there aren't many of them

First, while every user is important, it's clear this isn't a huge problem in the scope of the site. I could only track down 32 cases of overturned question closures where the OP was a new user and the question was marked off-topic or unclear. Second, there don't seem to be any users whose close votes are usually overturned, as almost all voters have only a 3-10% rate of suspicious close-votes. The number of overturned close votes is only high (as in, greater than fifteen) for a handful of users who are voting to close and reopen hundreds of questions already.

Is the problem small frequency in large numbers, or large frequencies in small numbers?

We could focus on the power-users, because about half the overturned close-votes come from about a dozen users. However, with a few outliers, most power-users average about a 7% suspicious close vote rate- and that's including duplicates and other closure reasons not brought up by the original question. Most of these users are already voting to reopen hundreds of questions as well. Definitely room for improvement, but hardly a case of toxic users.

We could also focus on the ~20% of suspicious close votes that come from users who don't have many other closure votes. There are some definite outliers here, but I don't know any good way to address them. The low percentage coming from these infrequent users also suggest the already-high reputation threshold is sufficient.

Graph of closure votes Pareto graph of closure votes by user

My Two Cents

My advice for voters: in my experience obscure questions are actually less annoying and less likely to distract from better questions than questions on common topics. Vote based on whether an expert could answer the question, not just a generalist. Also, remember to keep an eye on questions you've voted to close and if the question is edited, remove your vote. Several of these questions were closed just minutes after being edited to fix the issues.

My advice for askers: I know you're likely not reading this, but if you are, please avoid using acronyms and explain things so that even the new guy on your job would understand.

Best way to address the situation

Unfortunately there are limited ways to address this. This thread is a great way to start discussion. However, most of the posts here come from users with low rates of suspicious closure like Peter Kämpf and ymb1. Mods could PM the outliers, but that seems to be standoffish and I know directed guidance on how to vote is normally reserved for extreme situations.

It'd be helpful if the language for "unclear what you're asking" made it clear that we're talking about clarity to a subject matter expert, not just to a generalist, but I can't think of a clearer syntax off the top of my head and I believe changing it would involve a change to all of Stack Exchange.

Appendix: examples

Here's question closures I cherry-picked where obscure subject matter is a factor

$\endgroup$
  • 3
    $\begingroup$ Hi Cody, great analysis and I agree that in general the close votes are justified. Thank you very much for your insights! $\endgroup$ – Peter Kämpf Apr 25 '19 at 1:31
  • $\begingroup$ Can I ask you if you could share how did you derive the orange line in that plot? Because, for example, I see that 61 of my close votes got overturned, but I have closed 1407 posts, so I am not sure how that leads to a 40% ratio. $\endgroup$ – Federico Apr 30 '19 at 13:56
  • $\begingroup$ @Federico sorry, I should have labelled the axis. This is a Pareto graph with a cumulative % sum in orange. So it isn't that Prob(Frederico's vote overturned)=40%, it's Prob(vote overtuned is from Other, fooot, or Federico)= 40%. The intention of the orange line is to show how much of these close votes are coming from a few users, and how much is coming from low-frequency "other" users. $\endgroup$ – Cody P Apr 30 '19 at 16:58
  • $\begingroup$ ah, ok, but then without a grid of some sort that's really difficult to read, IMHO (and in general I find cumulative graphs without the corresponding non-cumulative one not really informative, e.g. I can't discern well where the highest % of votes overturned is) $\endgroup$ – Federico May 1 '19 at 13:55
21
$\begingroup$

Please Skip

I have said it before, there is an option to skip. In the review queue the reviewer is presented with leave open, close, and skip.

I use skip a lot when I don't know about the subject, so this answer is merely a reminder to others that it is okay to skip.

I also find comments that explain the down-vote/vote-to-close because the asker didn't expand an acronym.

This is also not a good excuse. Does everyone understand every question on Stack Overflow? Of course not. There are fields, and a term or acronym that is not clear to someone, does not mean it is an unclear question.

The Help Center says:

When should I vote to close a question?

Questions should be closed by casting close votes if:

  • they are sufficiently similar to existing questions and would be answered identically to them.
  • they are unclear, too broad, or otherwise problematic to identifying the problem in a way that can be properly addressed by answerers
  • they are sufficiently off-topic, as outlined in the help center.

The Bell Lift distribution question is closed as unclear. Unclear means a question was not asked (problem was not identified) – but here a question was asked.

Do I know anything about that subject? No.
Would that make it [objectively] unclear? No.

$\endgroup$
  • 3
    $\begingroup$ Unclear doesn't have to mean there was no question, just that it's not clear. But I agree with your points, I'll try to keep that in mind on future reviews. $\endgroup$ – fooot Apr 22 '19 at 16:34
  • 2
    $\begingroup$ "Unclear means a question was not asked" No. "Unclear" means the question was unclear. Maybe there was no question; maybe there was a question but it wasn't understandable. $\endgroup$ – David Richerby Apr 28 '19 at 21:00
  • $\begingroup$ @DavidRicherby: Grammar/spelling notwithstanding (which is not being contested), if it still remains unclear to some, does not make it objectively unclear. Hopefully this clarifies it. $\endgroup$ – ymb1 Apr 28 '19 at 22:18
  • 2
    $\begingroup$ @ymb1 It says "unclear", not "objectively unclear". Rational people can disagree about whether something is unclear. That's why it takes more than one person's vote to close a question, and why there's a procedure for undoing that. $\endgroup$ – David Richerby Apr 28 '19 at 23:38
  • $\begingroup$ @DavidRicherby: If your unclear is different to my unclear, please, post an answer. As what you're suggesting is the question must be made clear to all walks of life, which certainly is not the case on all of this network, given by the SO example. $\endgroup$ – ymb1 Apr 29 '19 at 9:57
  • $\begingroup$ @ymb1 I am saying that something that is clear to you can be unclear to me, and vice-versa. It is perfectly appropriate for one person to vote to close as unclear while another person votes to keep open. If a question is unclear to enough people that it gets closed as unclear then, yes, it probably does need to be clarified, even if some people already understood it. $\endgroup$ – David Richerby Apr 29 '19 at 11:00
  • $\begingroup$ @DavidRicherby: The problem with that rationale is those to whom the question is clear, need to be in a race against the masses in choosing leave open, and if they are away for a day or so, then a valid on-topic question would get closed. Hence, it is okay to skip. $\endgroup$ – ymb1 Apr 29 '19 at 11:07
5
$\begingroup$

Notwithstanding ymb1 very well put response I just wanted to add that even though I personally do not know the answer to that question it just feels very broad.

It does not specify what the actual problem trying to be solved is, which is a core tenet of questions in a QA format.

If you ask a vague question, you’ll get a vague answer. But if you give us details and context, we can provide a useful answer.

source: https://aviation.stackexchange.com/help/how-to-ask

To me, that question seemed vague. This might be my lack of understanding of the topic.

$\endgroup$
  • $\begingroup$ I agree that "why not" questions, while popular and useful, can be tricky to answer succinctly because good answers must address pros and cons, where usage does and doesn't make sense, and what's used instead. For this question that isn't too long, but for many questions that can lead to a whole book worth of answers. $\endgroup$ – Cody P Apr 30 '19 at 17:16
3
$\begingroup$

This question was asked 8 months ago, and nothing has changed. It is beyond me why questions should be closed when they are not understood by the voter. @ymb1’s advise is sound.

$\endgroup$
-5
$\begingroup$

Yes I would like to see more time to give others in the community a chance to also answer. Closing, deleting and down voting questions can drive novice users in Aviation into a question ban that lasts for 6 months and years. I almost think that users can be filtered and driven out this way by a handful of people even though their questions are not so bad?

$\endgroup$
  • 6
    $\begingroup$ If somebody is getting so many downvotes and close votes that then get question-banned, that's a pretty big hint that they're doing something wrong. Questions that are "not so bad" aren't going to get that kind of response. $\endgroup$ – David Richerby Apr 29 '19 at 11:02

You must log in to answer this question.

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