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.

Looking at Greg Hewgill's stats graphs the number of answers per question is slowly trending downwards:

graph of answers per question by day

Is it going to stabilise at some point so that (on average) questions receive one answer or will it drop below 1 so that questions are no longer statistically likely to receive at least one answer? How would that affect the site and is there anything that can be done about it?

share|improve this question

7 Answers 7

I think what people don't take into account when looking at questions and answers per question may be trends in terms of industry as far as what new technologies are being released at what times.

Example: When MVC is first released, it is likely that there is a lot more activity in questions and answers due to its newness. But when there is nothing new being put out at the time, you are going to undoubtedly have a lot more esoteric or very domain specific questions which the potential audience for is limited, which will cut down on the number of answers.

I would not be surprised if when Windows 7, Visual Studio 2010, and other new things become full blown releases you will start to see an uptick in questions which the audience may be greater, leading to more answers per question.

share|improve this answer
1  
Yeah, the easier question are answered and archived first. Over time, the only non-duplicates that show up are going to be harder to answer. –  hyperslug Sep 26 '09 at 15:03

For it to drop (significantly) below one would imply that most questions only receive one answer and some large fraction are considered unanswerable. There are a number of truly unanswerable questions -- questions without enough data to make even a guess as to the problem, questions on esoteric technologies with few or no other users, etc., however I don't foresee this becoming a problem in the general case. Given the drive for reputation, I expect that the majority of questions will always have at least one answer, the fraction of unanswered questions will remain low, and there will be enough questions with multiple answers to keep the average number of answers well above one. In the grand scheme of things, we only need one answer -- the "right" answer -- on each question for the site to work.

share|improve this answer

I notice this a lot on Superuser, where often only one good post is needed to either answer the question or let it be known it's not possible.

"Luckily" there are still topics where there are different points of view or alternative solutions, so there will always be some distribution.

BTW: you should check how many of those questions are also accepted as the correct answer, in which case it's not necessarily bad if they are answered within one answer

share|improve this answer

It doesn't look to me like it's steadily going downward. It looks to me like it's converging to 2.5.

share|improve this answer

<hangs head in shame>

I'll just have to try harder.

share|improve this answer

The problem is that it is hard to tell (looking at the data en-masse) if 1 answer is enough. If every question was solved with just one answer, I'd consider that good (although a second/third perspective is obviously very welcome). So perhaps to understand the data we'd need to separate out the questions which are ticked, or have no tick but a highly-rated answer (most likely: the right answer, but to a drive-by user who didn't mark it complete and is now deleted through age).

You might also need to exclude CW questions, as these tend to get more than average replies; and we used to get a lot more of the "what is your favorite *" questions. I wonder if this is also skewing the answers? Most "real" questions get (from what I see) < 5 replies, in part because we've actually got some clever people on the site who have given the right answer already.

So an average of, what, 2.5 answers/question doesn't concern me.

share|improve this answer
  • Low-hanging fruit got picked up
  • More controversial/discussion-y questions got asked a year ago

I suspect TXI is correct, in that after a while, spikes of usage will show up when new ponies software versions are released into the wild.

share|improve this answer

You must log in to answer this question.

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