What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 133 Stack Exchange communities.

I'm kind of curious what kinds of criteria other people go through when selecting the correct answer to their questions. Here's what I consider when selecting the answer to my question:

  1. Is the answer objective? It's easy to pick out biased answers. Usually they involve a statement along the lines of "x sucks, you should ALWAYS use y."
  2. Do other members seem to support the answer? Has it gotten a significant number of +1 posts or has it been voted up a number of times?
  3. Is the answer the most thorough? Does it have an example?

With that said, there are a couple of BAD reasons to choose an answer in my opinion:

  1. The answer is what you want to hear. My biggest pet peeve is when somebody asks what ostensibly appears to be a valid question, but is really just a thinly veiled attempt to get someone to agree with them. Thus, the question usually ends up showing a certain bias. And guess which answer gets selected? The one that confirms that bias.
  2. Nobody posted a better answer. If you don't see a good enough answer, don't select one.

I'm sure there are a few things I'm leaving out, plus some of what I've said may be arguable. What is everyone else's take on this?

share|improve this question

migrated from stackoverflow.com Sep 1 '09 at 19:33

This question came from our site for professional and enthusiast programmers.

8 Answers 8

up vote 19 down vote accepted

Here are some guidelines:

  • Does it answer your question precisely?: The selected answer should answer your exact question correctly, and not some tangent topic created from your question.

  • Do others agree?: Make sure at least some other stack overflow users agree with you. You don't need to select the highest voted up answer, but try to avoid selecting an answer that has net down votes just because you agree with it. You may be wrong in that case.

  • Don't settle: Don't select an answer just because it is the best one available. Only select an answer if it answers your question.

  • Don't pick too quickly: Try not to select an answer minutes after posting it. Sometimes a better answer will come hours later, and you might miss out on it if you select your answer too fast. The other problem with this is that other stack overflow users will see the accepted answer, only read that one and not the others, and vote it up. You will have an inflated vote answer for possibly not the best answer.

  • Multiple answers: If multiple people answered only part of the answer, comment on each and ask them to expand it including @username's answer. Upvote each of the partial answers and finally accept the complete answer once someone posts the answer completely. The person adding to their answer should give credit to the person they borrowed ideas from. It is better to modify someone's post with a small missing detail and accept it, then to accept a partial answer. If someone posts a summary answer, even if they contributed no new information, and if it answers your question completely, then you should accept it.

  • Changing accepted answers later: It is a good idea to review your responses tab on your profile page. Sometimes people will post an answer even know you have already accepted an answer. The new answer might be far superior than the one you accepted.

  • Examples and code: A good answer often has an example and code in the answer. But sometimes the answer is purely theoretical. If you have a practical question, hold out for a more practical answer that has examples and code.

  • First answer vs similar subsequent: You should not select the first answer just because it was the first to have a major point. Instead select the most correct.

  • First answer vs exact subsequent: If 2 answers are identical accept the oldest.

  • Periodically review your unanswered questions: If you don't religiously watch your Responses tab in your profile, you should periodically review your unaccepted answers and look to see if there is an accepted answer candidate.

  • Subjective questions: In my opinion, questions tagged with subjective should not be able to be accepted. I think it is better not to accept an answer if your question is subjective. By the very definition of subjective, there is no best answer anyway.

Contribution of ideas in this answer: Myself, @DannySmurf, @Sergio Acosta, @Zooba, @jjnguy, @Daniel Auger, @Jason Baker, @kristof

share|improve this answer
    
Questions tagged "subjective". You could include questions tagged "discussion" under that point too. –  codewaggle Jun 13 '12 at 23:52

In my opinion the main criteria should be a positive answer to the question "Does it solve my problem?"

share|improve this answer

@kristof: I agree, but it's not always that simple. Sometimes there are multiple answers that all solve my problem. Or sometimes I may not really know if it solves my problem until I've put a week's worth of work into it.

And sometimes I see people making choices that I don't personally agree with. There's nothing wrong with that, I was just curious if anyone else had any radically different ways of thinking about this.

If I'm overcomplicating things, then so be it. :)

share|improve this answer

One thing that I've noticed is that people tend to accept answers too quickly before it plays out in responses. There is nothing wrong with letting your question linger a few days without an accepted answer IMHO.

share|improve this answer

Whatever answer lends to my code working correctly gets the accept by me.

If a later answer might be more thorough but give mostly the same answer, I give the credit to the person who helped me first.

share|improve this answer

I posted on UserVoice a suggestion to not allow questioners to select an answer for some period of time after asking the question (I suggested 24 hours). This would give time for a few decent answers to appear.

The suggestion was rejected as "we already have problems with people not accepting answers."

share|improve this answer

I think selecting an answer to accept is not always easy.

If the question is related to debug or accomplish something, the first answer that helps solve the problem/accomplish the goal is the correct answer. Or at least it is the answer you were looking for. Then it is easy.

But every other type of question, like 'what tool do you recommend for the job' kind of question is not as easy. There can be several equally helpful answers.

In other words, I don't think every question should necessarily have one and only one accepted answer.

share|improve this answer

@Sergio Acosta: Like in any multiple-choice situation, there may be several close (ie, helpful) answers. But if you can only select one, you need to decide which is the best answer. That is, if you're actually asking a question in the first place, which not everyone is. Just because we call every entry here a "question" doesn't mean that every entry actually is a question. A discussion about recommendations about a tool for a particular job certainly isn't a question, so why should the original poster have to select an answer at all? That's where voting becomes more useful.

That said, I would love to see a discussion area to this site, specifically for those situations.

@Jason Baker: I have no problem coming back days or weeks after the fact to mark an answer as right. I have an outstanding question now, with several helpful suggestions that it's probably going to take me two weeks to work through. When I figure out which (if any) solved the problem I was asking about, I will come back and give someone the correct answer. If we're using this site as an archive, which Jeff said was one of the main purposes, there is absolutely nothing wrong with that.

share|improve this answer

You must log in to answer this question.

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