What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 134 Stack Exchange communities.

It's taking longer and longer to close questions. Should it be easier to do so and how would this work?

I think this is related to Are unanswered questions a problem yet?, as the rate of increase of the number of unanswered questions keeps increasing something needs to be done. It's also semi-related to this (shameless plug) of mine: Keep posts flagged spam or offensive in the review queue. In other words it falls into the group of questions that should be tagged

Here's the standard data.se query. ACT stands for Average Close Time and is in seconds. I've split this into two, ACT - 7 days takes into account only questions that were closed within 7 days of being asked and ACT - 1 day only questions that were closed within 1 day of being asked. If this isn't done the average close time for the questions from 2008-9 is massive as they sometimes get closed years later. The % Closed column is the percentage of all questions closed for that quarter and the Answered column is a count on the number of closed questions with at least 1 answer. I think the results for 2008 can be safely ignored as being completely atypical.

Year Quarter Total  Closed ACT - 7 days ACT - 1 day % Closed Answered 
---- ------- ------ ------ ------------ ----------- -------- -------- 
2008 3       19783  994    110958       12769       5.02     982      
2008 4       41855  1180   39637        11037       2.82     1155     
2009 1       56579  1173   18651        8679        2.07     1128     
2009 2       78406  1209   27915        14816       1.54     1140     
2009 3       101800 1220   34076        18061       1.20     1146     
2009 4       116450 1116   32597        22367       0.96     1031     
2010 1       146730 1452   34505        20120       0.99     1323     
2010 2       163813 2160   43525        24787       1.32     1904     
2010 3       193548 3835   40818        21410       1.98     3376     
2010 4       212085 4404   38673        20001       2.08     3719     
2011 1       276111 6276   33815        19074       2.27     5028     
2011 2       311008 9681   45556        21554       3.11     7690     
2011 3       348000 12822  43514        22824       3.68     9738     
2011 4       352552 12905  55807        29879       3.66     9880     
2012 1       423268 13701  71005        36510       3.24     10643    
2012 2       437245 17779  77051        37340       4.07     11259        

As you can see there's good news and bad news. The good news is that despite everything a greater percentage of questions are getting closed. This may be slightly skewed by the auto-deletion policy but I don't think by much.

What's worrying is that the average time to get a question closed is increasing, quickly. It now takes on average 10 hours 22 minutes to close something up from 6 hours (5:59:14) only a year ago. That's an additional 4 hours and 22 minutes that questions remain open when they shouldn't be. To tie this back into my answer on the linked post; it also takes up valuable answerer's time, which is in shorter and shorter supply. The vast majority of closed questions have been answered.

It's a similar story on Super User, though they seem to have stepped up their game in the latest quarter.

Year Quarter Total Closed ACT - 7 days ACT - 1 day % Closed Answered 
---- ------- ----- ------ ------------ ----------- -------- -------- 
2008 3       70    7      null         null        10       7        
2008 4       55    3      null         null        5.45     3        
2009 1       55    1      null         null        1.82     1        
2009 2       152   8      null         null        5.26     8        
2009 3       9112  472    42169        19083       5.18     466      
2009 4       9562  435    46618        27828       4.55     423      
2010 1       9299  601    42405        24631       6.46     541      
2010 2       8993  499    38234        23670       5.55     413      
2010 3       9945  617    39970        25018       6.2      516      
2010 4       9302  605    37107        26661       6.5      497      
2011 1       10770 696    42791        26440       6.46     537      
2011 2       11965 741    49373        22945       6.19     552      
2011 3       12116 1014   54805        24796       8.37     666      
2011 4       11182 921    78977        26734       8.24     607      
2012 1       11562 849    81987        25369       7.34     523      
2012 2       12608 1140   63333        21298       9.04     520      

I can't do a complete analysis as I don't know who did the close-voting but I would hazard a guess that a possible reason is that the people who vote to close are, in the main, answerers not questioners as the number of questions has increased out of all proportion to the number of people answering the ability of the answerers to get rid of them.

Are these numbers acceptable and should it be easier to close a question and if so how can this happen?

There's some hope among the mess. The fastest ever time to close a question by a non-mod is 62 seconds, so SBs comment is almost right! casperOne♦ wins the mod prize at 28 seconds :-).

share|improve this question
1  
Keep in mind that SEDE doesn't include deleted posts. This may skew the numbers as older closed questions may have been deleted. I don't think the percentage of closed question is particularly relevant anyway, not without a reference to see how many questions should have been closed. My informal feeling is that closing is difficult in rare tags, but the review interface (old or new) helps this (however there is a huge backlog to go through). –  Gilles Aug 25 '12 at 13:34
    
Yes @Giles, but it could make it worse. I wish I could agree with you about the new Review but I think that the Close Votes section has been steadily increasing. –  ben is uǝq backwards Aug 25 '12 at 13:41
    
How do you calculate the average time? Do you filter out old closed questions? Or do you use a median? There were several really old questions closed recently, which might skew your results. –  CodesInChaos Aug 25 '12 at 13:43
    
@CodesInChaos, I only count questions that were closed within 1 or 7 days of their creation to avoid that. –  ben is uǝq backwards Aug 25 '12 at 13:44
    
The problem's worse on smaller sites, it's extremely hard to gather non-mod close votes on graduated but small sites –  Ben Brocka Aug 25 '12 at 16:20
    
Holsters the still smoking gun... –  casperOne Aug 26 '12 at 3:22

2 Answers 2

up vote 7 down vote accepted

Question don't get closed for one principle reason: they don't get enough close votes.

That seems obvious, but the reasons why are not so obvious:

  • Some questions are so uninteresting that they don't get enough views to attract the needed close votes.

  • Borderline questions are seldom closed by the community. People don't want the hassle of arguing about them.

  • There is a trend towards downvoting and flagging questions instead of voting to close them. I suspect this is because it is easier than voting to close, where you have to figure out what the close reason is. More community members have the ability to flag and downvote than they do to vote to close.

Moderators wind up closing most of these questions, creating the perception that it is the mods, and not the community, that is in control. Also, "I'll just let the mods take care of this, since they're apparently the ones doing all the work."

We don't need easier ways to close questions. We need better participation and a greater sense of ownership from the community at large.

share|improve this answer
    
Yes, that's the obvious corollary, if you don't make it easier you need more people to be doing so. –  ben is uǝq backwards Aug 27 '12 at 10:45

Yes, closing questions should be easier.

In order to do this you could allow people greater "Close Vote Weight", i.e. allow some close votes to count as more than one based on a logarithmic progression of reputation. At 3k your close vote counts as one, at 10k as 2 and at 100k as 3. Assuming Jon Skeet reaches the 1m mark he'd get 4.

There will never be a situation when only one "ordinary" (non-mod) person closes a question.

share|improve this answer
1  
No, I still want 5 pairs of eyes to close a question. The tools to get the questions to the right eyes could certainly use improvement. –  Gilles Aug 25 '12 at 13:34
2  
Nope, nobody's vote should be stronger than that of anybody else. The only exception being that of a moderator. I can't agree with this. (Btw, why make this community wiki?) –  Bart Aug 25 '12 at 13:36
    
I suggested as an alternate to the close weight to allow multiple votes an a single post. Though my current opinion is to reduce friction for those who can vote to cast them easily, like the review tool. –  yhw42 Aug 25 '12 at 14:08
2  
@yhw42 - Allowing multiple votes is essentially the same as giving a user more close vote weight. The point is we don't want just one or two people making these decisions, we want 5 separate, distinct people to have to make them. –  jmort253 Aug 25 '12 at 16:51

You must log in to answer this question.

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