Questions with no marked answer

Question

I have been poking around looking at the question answer rates and they don't look too flash at this stage.

At the moment of writing this (the site is one week old) we have 64 questions, of which only 21 have a marked answer. A 33% marked answer rate is not fantastic. This could be the result of several things:

1. question askers are new to the SE way of doing things and are unfamiliar with the practise of marking an answer
2. question askers are not getting answers that satisfy them enough to be marked as the answer and are not elaborating on their questions to get better answers
3. the questions are too generic and no one answer can be picked as the best
4. the question has attracted several good answers and the question author doesn't want to make the decision of which answer should be marked

With #1, it is an often talked about issue on SE sites, and there is no silver bullet to fix it. In fact Jeff has stated in the past that it should not be mandatory to have a marked answer. This is fine, but could potentially be damaging to a beta site if it is one of the metrics being monitored.

With #2, there is probably not a lot of things we can do about this, except for drawing the attention of the questioner at a later time via a comment on their question.

With #3, this is a really dangerous and unproductive type of question to have on the site. There has already been a fair amount of discussion on Meta trying to flesh out what is and isn't on topic. I'd expect this process to continue for a while, but even once it has finished there will still be a bunch of questions that are still too generic.

For #4, questioners have to bite the bullet and choose an answer. Someone's feelings may get hurt but they'll have to learn to live with it - it doesn't mean their answer wasn't good and they will still accumulate up votes over time.

The site has only been going for a week and we shouldn't be getting too alarmed yet, but I think we need to keep a close eye on this - we don't want indecision and inertia to keep the site from thriving.

Has anyone got any comment on this?

(For those who want it, here is a link to the questions that have no marked answer).

Answer 1

As someone who hasn't accepted an answer on the two questions they asked, my reasons are either:

1. The answers given, while interesting, didn't solve the question I posed. (Same as your point 2.)
2. Marking a question 'solved' tends to discourage further answers. The site's too new to do that.

On point one, I asked a fairly detailed question, got many answers, but none that referenced the two points I was looking for. It's possible that no answer exists in the format I'm seeking, but the site's still young!

In reference to point two, I think that having several unsolved questions may be beneficial when the site's opened up for public beta -- a page of solved questions might be less likely to draw new members in to contribute new answers, which is why I left my other question marked unsolved as well.

Answer 2

The major statistic that appears to be used–and I could be wrong on this–is questions that are answered (possibly answered and upvoted), rather than specifically accepted answers. Still, it is worth thinking about solutions to it, as it will provide more of an attraction to answering questions (as on other stack exchange sites).

One thing we can do here is go through and ping the original question askers and see if they have an answer that they like and, if so, could they accept it. That will help address the education aspect of #1 (which also will be addressed somewhat over time) and #4.

I suspect #2 will become an easier hurdle when we get more people, since the domain is sufficiently broad that we may just not have anyone with sufficient expertise to really adequately answer the questions. We can also go through and find questions that do not have accepted answers yet and see if we can provide better answers for them. For this we just need a few people willing to do the review (I seem to recall a few other SE betas have done something along these lines–at least making sure everything had at least one quality answer–at various points to good effect).

I think the best thing we can do with #3 is just continue to discuss topicality on the meta and vote to close questions that are overly broad or that cannot be adequately answered in their current form.

Answer 3

An additional point to consider: If a question has low interest (low views, low votes, etc.) the original poster may be waiting to get more feedback. Essentially, questions are viewed as one of only a handful of types:

1. The OP knows the answer they're looking for and is looking for an answer to support their limited view of the world (these usually get answers quickly).

2. The OP knows nothing of the subject, and is waiting for the community to let them know what they should view as the correct answer (usually by allowing votes to accrue, which may take a significant amount of time)

3. The OP has asked an open ended question that, due to the continued topicality debates, has remained open or contested. The question really deserves no answer, but will get one either based on how much they like the answerer, how much they like the answer, or simply how flighty they are.

4. The OP has done a hit-and-run (asked and not waited around or returned to pick an answer or completely lost interest in the question they asked because it was a flight of fancy in the first place); the question will never be answered.

The acceptance rate is not a significant indicator of community health, but rather the amount of feedback the questions receives. I do recall that Stack Overflow had a rather nice nag about unaccepted answers, and perhaps that will help get people to accept their answers.