-2

I quite admired the very data-oriented answer to a recent meta question in the ongoing closure war, and now I feel encouraged to ask a data-directed meta question.

Is the probability of a question receiving a closure vote independent of the number of closure votes it has? The question could be answered in an objective way if we had data on the numbers of closure votes each question (removed/closed or not) in the system has.

To make the question clear, suppose that each question is associated with some intrinsic probability p(1) of receiving a first closure vote. If the probability of the question receiving a second, third, and fourth closure vote, is the same, then

p(1) = p(2) = p(3) = p(4)

and there is no increasing or decreasing trend with the number of closure votes already made. This is what would be expected if closure vote probability was independent of the number of closure votes already made.

Whether or not the probabilities are independent, the overall probability of closure would be:

P = p(1) x p(2) x p(3) x p(4)

Obviously, P is a function of p(1), so a question's overall probability of getting closed is clearly dependent on its probability of receiving the first closure vote. This is not what I am asking.

As for the context of the question (what am I getting at?), if it could be shown that

p(1) < p(2) < p(3) < p(4)

then this would mean that it would be misleading to say that a question gets closed with four closure votes, tout court.

11
  • 2
    Ah, the Holy Grail of all data junkies. Not available through the Data Explorer, I'm afraid. See this related question on MSO (which, funnily enough, complains about the opposite thing, too many questions staying unclosed with 3 or 4 votes; but that's a different site, and a different time, too). There was at least one other one that was basically identical to yours, but I can't find it for the life of me.
    – RegDwigнt
    Dec 20, 2012 at 0:09
  • 2
    Given that getting that first close vote is what makes a question show up on the Close Vote review queue, I don't see how further close votes could possibly be statistically independent.
    – Marthaª
    Dec 20, 2012 at 0:49
  • @Marthaª that fact makes a big difference. i think we could accept your conclusion without getting the necessary (but unavailable) data.
    – user31341
    Dec 20, 2012 at 1:20
  • 2
    I take it you mean "Is the probability of a question receiving another closure vote independent of the number of closure votes it already has?". I could always be wrong, but I would say the answer to that is NO. For the simple reason that if a question already has several closevotes, that's probably because it's a poor question - which obviously correlates strongly with its likelihood of getting closevotes (as proven by the fact that it already has some). Dec 20, 2012 at 1:21
  • 1
    @jlovegren: It's at least feasible some people are more likely to closevote things they see in the review queue, but I can say with absolute certainty that doesn't apply to me, so I don't think you should refer to it as a "fact". It's quite possible people with my attitude outnumber whatever type you and Martha are thinking of. Within the review queue, I tend to see my job as to assess the performance of other closevoters, and I consciously look out for cases where they've made what I might consider to be hasty mistakes. Dec 20, 2012 at 1:31
  • ...and as we're always asking here - can you give some more context? What exactly would be the advantage of setting a specific value on the correlation figure you seek? Even if we could isolate and quantify every different factor involved, what would it tell us? Dec 20, 2012 at 1:35
  • @FumbleFingers hopefully my initial phrasing wasn't too misleading about what I was asking. I've clarified the wording and added the context you ask for.
    – user31341
    Dec 20, 2012 at 3:40
  • @FumbleFingers in my response to @ Marthaª, the "fact" I refer to is that there is such a thing as a "Close Vote review queue". Are you saying that no such thing exists?
    – user31341
    Dec 20, 2012 at 3:40
  • 1
    I don't see how this question is actually answerable. For myself, the probability of my casting a close vote is not dependent on any existing votes. I have cast the first on occasion, and decided not to cast the last as well.
    – Andrew Leach Mod
    Dec 20, 2012 at 12:38
  • @AndrewLeach no, the question is not answerable via introspection. You may know your probability of closevoting a question you see, but you'll also need to know your probability of seeing a question. Do you read every question posted in detail? Do all high rank users? Earlier commenters have pointed out that the ideal data set is not available, but have drawn attention to a site mechanism that all but answers the question.
    – user31341
    Dec 20, 2012 at 13:19
  • @jlovegren: Sorry - I misread "that fact" as referencing Martha's statement that further close votes could [not] possibly be statistically independent, with the implication they must therefore be more likely. I don't understand your math/stats (for example, why is it P1-P4, when closing actually takes 5 votes?). But it does seem obvious to me that a question's chances of getting closed are zero if nobody reads it, rising to certainty if an infinite number of people see it. Going into the review queue must increase the number of viewers, so it must increase the chances of closure. Dec 20, 2012 at 16:41

2 Answers 2

2

Is the probability of a question receiving a closure vote independent of the number of closure votes it has?

Absent of any hard data, I can offer a few reasons why I'm pretty sure the answer is "no."

  • The first close vote puts the question into the review queue.
  • The number of close votes shows up right under the question itself as you read it.
  • If a close-voter left a comment (as is encouraged), you'll see it.

Any of these things could (and almost certainly do in many cases) cause readers to look at the questions differently than they would if there were no close votes. I would venture to say that the whole point of these features is to make subsequent readers pause and say: "Huh, is there something wrong with this question?"

None of this implies that we're all mindless sheep letting our close votes automatically fall like dominoes. I've seen plenty of instances where I've disagreed with a prior close vote.

But pretend for a moment that we could eliminate the subjectivity and measure the quality/appropriateness of a question objectively. Wouldn't we then expect a direct correlation between the quality and number of close votes? If that's the ideal, it seems naive to expect (or even want) the probability to be independent of the numbers of votes.

6

Ongoing war? I thought we were under a Christmas truce...

(Actually, I'd regard this as a debate, or a discussion, not a war, but perhaps that shows which “side” has been more prone to use incendiary language in their proclamations.)

That said, there have been several allegations made during some soul-searching about closed questions of late.

Many questions get closed (both sides agree with that). As for why those questions get closed, there seems to be at least two viewpoints:

  • VIEWPOINT 1: Questions are closed because of an unwelcoming and insular community, which is steeped in snobbery. Some high-rep users have appointed themselves as patrolmen who seem to take more joy in closing questions than they do in answering them, particularly when these questions make them feel insecure in their pseudointellectualism, and they find themselves at a loss to answer the questions. This has led to a trigger-happy close culture that is absent from other Stack Exchange sites. Closures should be reserved for all but the most extreme circumstances (such as spam or indecent content), because quick closures drive away newer users who might eventually become productive members, given enough nurturing guidance.

  • VIEWPOINT 2: Questions are closed for a few different reasons, the most common one being that they are questions that have already been asked, and are therefore closed as duplicates. Many others get closed because are overly vague or general reference. Many questions are not written in a way that would appeal to a “serious language enthusiast.” If left unchecked, this could promote a culture where ELU is inundated with questions that are of poor quality (i.e., not well-presented and/or not well-researched). Moreover, it is possible to close or downvote a question without being “rude” to a new user; such votes can set expectations of quality and serve as feedback mechanisms to promote improvement. If ELU has a higher close rate than other SE sites, that's because ELU is getting more than its fair share of questions that need improvement, or have already been asked.

You've asked for data, though, so I'd like to provide some. I look at those engaged most heavily in this debate war, and this is what I've noticed:

Users Making the Most Impassioned Arguments Against the Current Close Culture:1

  Member Name     Member For     Reputation  
  jlovegren        29 days          1105
  Andrew Lazarus   36 days           240
  spiceyokooko     11 days           711
  Billy             3 months         568
  CUM. TOTALS    ~5.5 months        1624 

Users Who Are Most Active in the Current Close Culture:2

  Member Name     Cl. Votes   Member For     Reputation  
  tchrist            27       2 yrs 1 mos       24532
  FumbleFingers      18       2 yrs 0 mos       51367
  MetaEd             17       1 yr  2 mos       11920
  Mitch              15       1 yr 10 mos       20630
  Robusto            10       2 yrs 1 mos       75281
  CUM. TOTALS                 9 yrs 4 mos      183730

This data reminds me of a joke I once read in Boys’ Life magazine, where a young lad comes home, and his mother asks him, “How was the parade?”, to which the boy replies: “It was great! The whole band was out of step, except for me!”

Why do I bring this up? Maybe this site closes too many questions, and maybe it doesn't. I don't care where you stand in that debate; I think either side could defend their position with valid arguments.

However, I don't particularly care for the direct allegations of snobbery that have been leveled, particularly by four members whose combined time on ELU is less than half of mine. You want to encourage us to close fewer questions? Fine, but do so productively and constructively.

Four users came in here, took a look around, didn't like what they saw, presumed they had us all figured out, and painted us with a very broad brush. To those four users: quite frankly, I don't care if this site closes more questions than other SE sites, and I don't care if that bothers you. I was a new user here less than a year ago; I liked what I found, I contributed in a positive way, and I earned the right to cast my own votes. I wouldn't wander into someone's parlor for the first time and tell them how to decorate their Christmas tree; I wouldn't run for mayor of a town where I had only lived for three months. Stop bashing the people who have built this place with their own hard work. Not every downvote is a knee-jerk reaction; not every close vote is cast by an insecure member wanting to defend a fiefdom; not every allegation of poor quality is proof of snobbery. Stop railing like it is.

Questions will continue to be closed, and, unfortunately, people will read that how they want to read it. I think there's a third viewpoint, one that may not be mentioned much during the “war,” but may be behind the spirit of the “Christmas truce”:

  • VIEWPOINT 3: Contrary to some initial impressions, ELU is made up of hard-working community members who do their best to promote and maintain the health of the site. Perhaps some folks are more prone to play the role of the “good cop,” and others the “bad cop,” but this is because everyone has their own ideas about what ought to be improved, and about the best way to go about promoting that improvement. Perhaps it wouldn't hurt if a few high-rep users became a bit more introspective every now and again, and asked themselves if they've become too jaded, and maybe need to make a conscientious effort to edit more, and close less. Perhaps it wouldn't hurt if a few lower-rep users viewed the feedback and comments left by those users as helpful exhortations, rather than personal attacks, and didn't jump to the conclusion that a downvote or a terse comment only proved that they were not “part of the clique.”

And with that, I give you Moore:

“Happy Christmas to all, and to all a good night.”


1Based on comments found among these five meta questions: #3384, #3422, #3425, #3438, #3451.
2Taken from the data found in this answer.

6
  • I appreciate the data, although I might have drawn different conclusions. Dec 25, 2012 at 1:48
  • i've been somewhat careful about not making ad hominem arguments on this topic. i'd even like to think that i'm snobbier than the average high rep user. if it's that you'd like me to wait until my rep is at 10k or so to try to change the site policy, i don't mind.
    – user31341
    Dec 25, 2012 at 3:37
  • @jlovegren: Not at all. Every user gets a say, that's why it's a community. Plus, you've been quick to temper your rhetoric, and ask some productive questions on meta. If many of the comments over the past couple of weeks had said, "There seems to be a few users who are too quick to close," or, "Maybe we should try to give more helpful guidance than we're giving," you'd find me in agreement. Instead, though, I've seen the whole community indicted: "This is rudest community on the whole Stack Exchange;" "Math has world-class mathemeticians; you guys are just a bunch of linguistic wannabes."
    – J.R.
    Dec 25, 2012 at 7:22
  • 1
    (cont.) I think such rhetoric is neither accurate, nor productive, and I also think there's some irony in relative newcomers telling the rest of the community that they have us all figured out. I have no problem with new users asking us to do some introspective soul-searching, but comments like "It would be nice if you and your fellow bullies would find your own website and let the rest of us decide which questions are trivial" seem to cross a line in my mind. That's not how I would phrase it, particularly if I were the new kid on the block.
    – J.R.
    Dec 25, 2012 at 7:37
  • 1
    @J.R. the point is well taken but the ones who would find the lessons on urbane interaction instructive are not reading. i'm finding that legitimate queries might be effectively brushed aside so long as boorish types have the same query. i suggested resuming the issue when my points are high enough so that issues i raise won't be lumped together with the same issues raised by those who don't observe decorum.
    – user31341
    Dec 27, 2012 at 23:55
  • @jlovegren: I see no need for you to wait until you reach a certain threshold; so long as questions are worded constructively, inquiries are welcome. (Even if they're not worded constructively, they're still welcome – but that often invites more pushback.) I apologize if my answer here erroneously lumped you in with the wrong crowd.
    – J.R.
    Dec 28, 2012 at 11:09

You must log in to answer this question.