Out of idle curiosity, I was looking through some of my (old) answers to questions which have been reasonably upvoted, were often the only answer, but have not been accepted.

Fair enough, you could argue that some of those answers fell short in some way, but on further investigation it seems there are a number of users asking questions who hardly ever accept an answer (and not just by me(!); they have accepted answers for <20% of their questions). For the record - on Physics SE I have accepted answers for 19/23 of my questions that have an upvoted answer, and 4/5 on Astronomy SE (admittedly I accepted a couple of these after review while writing this question!)

Why are some users just not accepting answers? Are the answers unsatisfactory? Are users just forgetting they've asked the question? I think if it is an old question and has upvoted answers, then it might be polite to at least indicate why the answer(s) doesn't meet your requirements, especially if you want answers to any more questions.

• I've noticed this trend a lot recently too. I wonder if some newer users are even aware of what the accepted answer checkmark is for.
– called2voyage Mod
Jun 21 '21 at 13:32
• This post motivated me to check my own questions and accept several more answers!
– Connor Garcia Mod
Jun 21 '21 at 17:04
• I have to confess that I did not accept some of your (and other persons') answers although they answer part of my questions, merely to give other users a motivation to write a second answer, but this strategey does not seem to help much, so questions with a single answer are very unlikely getting a second, complementary answer, I feel. I.a.w. I suspect that there are more unaccepted answers for questions with a single answer than for those with multiple answers. Jun 23 '21 at 11:55
• Aside from "ignorance of how SE is supposed to work", it's possible that the majority of askers are smart enough to realize the the entire SE "rep system" is meaningless fluff. (Note, the "majority of askers" is not the same as the "minority of askers who ask the majority of the questions" - and who seem to also like playing the "meaningless rep" game. Jul 3 '21 at 1:46

There are many reasons.

In favor of our members who ask questions:

• Yes, we have a tour, we have discussions about how to write questions, and we have instructions regarding what to do when one does receive answers to questions.
There's a basic problem with this concept: Nowadays (and probably forever), nobody reads the instruction manual for anything.
• Some new members simply haven't learned the protocol because they haven't read the instruction manual. When a somewhat new member who has asked multiple good questions but hasn't accepted a single answer, it can help, and has helped, to kindly remind that newish member of the site's protocols. The site does issue reminders for comments.
• Sometimes the initial version of an answer is unsatisfactory. I myself have been guilty of writing initially unsatisfactory answers, and then improving them over time.
• There are no reminders. Even worse, there apparently is no way to search for questions asked by a member that do not have an accepted answer. I have been awarded accepted answers years after the fact, and that was done by highly regarded members. Members sometimes happen to come across a question asked years ago and realize they had never accepted an answer.

Editorial comment: I get prompted on a regular basis across the SE to update my cookie selections. I suspect that if I even once clicked on the "Accept all cookies" option, I would ever be prompted to update my cookie selections again. That is not the option I select. It would be nice if SE also occasionally prompted questioners who have not accepted an answer to do so.

In disfavor of some members of our community:

• Some new members are what I call "drive-by questioners." They expect instant answers, and if they don't get an instant answer, they go elsewhere. Instant answers just don't happen at Astronomy.SE, and at other corners of the SE network that aren't highly trafficked. It can take hours, sometimes days, before the first answer is posted.
• Questions on rare occasions are asked in bad faith. A prime example: "Pluto should be called a planet. Do you agree?" Questions asked in bad faith are much more common on non-technical portions of the SE network. By way of analogy, reddit has an AITA subchannel. The questioner is rarely contrite when the answer is an unambiguous "yes, you are TA."
• You're right there's unfortunately no functionality to filter on exclusively one's own questions where none of the answers have been accepted. However, there are a couple helpful tools. First, if you go to your user page and click on "Questions", you can at a glance see which ones do not have an accepted answer because the answer box will not be filled in with green unless it has an accepted answer.
– called2voyage Mod
Jun 22 '21 at 20:09
• The other way is by using the Stack Exchange Data Explorer: data.stackexchange.com/astronomy/query/edit/1425603. If you use this query, all you have to do is replace my user id (6) with your user id (can be found in the URL for your user page). The caveats here are that the Data Explorer updates weekly, so it's not going to have your latest questions, and it is not very friendly for non-technical users.
– called2voyage Mod
Jun 22 '21 at 20:11

update: searching with:

• hasaccepted:no user:me answers:1 has yielded 116 questions
• hasaccepted:yes user:me answers:1 yields 371 questions

(the 1 is interpreted as one or more i.e. nonzero)

This suggests there may be some low hanging fruit for me, and anyone can apply this test to their own questions.

I can speak for myself, I'm not sure if I'm an outlier and I'm not sure if I'm an "offender".

Question and answer posts show up in people's searches, but they also receive eyeballs when they are in the active queue, and those eyeballs can result in new/different answers or helpful comments and votings.

I like to maximize eyeballs by spacing out events, and one way to do that is to not accept an answer for at least several weeks.

But because I am ephemeral in nature, I sometimes loose track.

Because of that, I sometimes go back and look for my answered questions where acceptance is indicated.

If I don't agree with an answer say at least 80%-ish, I won't accept it.

If I don't understand an answer then even if it's highly upvoted I won't accept.

Sometimes I'll accept an answer years later, and it's half-and-half forgetfulness and finally understanding the answer.

If there are two or more equally good answers in that both helped me equally and if one person is new low/rep, I will sometimes accept that one. If they are both high rep people and I can't choose, I don't choose.

mea culpa

I may have some answer-accepting to do here. I'll start looking. There's probably some booleans I can put in the search bar too look for my questions that have answers but no accepted answer.

I will get that round tuit and get on with it!

If there are some cases of unaccepted answers that need my attention/consideration please feel free to let me know!

• You very recently accepted an answer that I supplied years ago. That late acceptance was not your fault. Given the emphasis on accepted answers, it should be very easy for questioners to find questions that they have asked, that do have answers, but that do not have an accepted answer. Even better, it would be nice if the system prompted questioners that an answer was supplied or a previously supplied answer has been updated to a question asked by the OP. Jun 22 '21 at 21:32
• I know that SE is quite capable of occasionally prompting its users because the system regularly re-asks me to re-select what cookies I will allow. I happen to choose the one option the SE network apparently does not like (necessary cookies only). Once a week or once a month should be good enough to remind questioners that they have questions for which answers exist but no answer has been accepted. Jun 22 '21 at 21:41
• @DavidHammen I'd love for there to be an answered/unaccepted question explorer utility, or for a per-post alarm clock I could set that reminds me a few weeks later to revisit it.
– uhoh
Jun 23 '21 at 0:43

There might be one more reason. StackExchange Astronomy is one of the few places where astronomy-dummies can get answers to our questions. So those of us (myself included) who are learning about astronomy via the web or books rather than a classroom find this forum a godsend.

When the answers appear conflicting or is accompanied by a great deal of comments that indicate disagreement, I feel unqualified to check that one answer is more accurate than another.

There should be a way for moderators to check that a question has been answered, or over-ride my green check if I have accepted an inaccurate answer.

• +1 I appreciate that predicament, but the "There should be a way..." will not happen. Voting is generally the best way a future reader can judge answers, even though it's not accurate, and comments offer further guidance though they may be confusing. Also check the dates, a late answer many years newer may have far fewer votes because the flurry of voting in the very beginning (first few days) never happens again. That's all there is, there ain't no more! The accept check mark is always strictly the view of the question's author (the "OP").
– uhoh
Jun 28 '21 at 23:41
• It's always been that way and changing it now after millions of SE posts would be confusing. Moderators have moderation skills but are not necessarily any more of an expert for any given question than anyone else. They are users like us but are generally a lot more skilled at (or at least committed to) moderation than the average person. We shouldn't assume they have any more expertise than any other commenter or voter, and when there's debate then there simply may not be any clear or 100% correct answer on the page. The way SE works as it is is it's imperfect yet captivating magic.
– uhoh
Jun 28 '21 at 23:44
• Maybe have a flag that indicates the question was answered without having to select a specific answerer? That way the question doesn't keep climbing to the top as unanswered. And as far as expertise goes, 99% of the moderators have forgotten more about astronomy than I will ever learn. I am quite impressed by the results I get here. Jun 28 '21 at 23:56
• SE already displays has/doesn't have an answer, number of answers, and accepted status on all questions i.stack.imgur.com/LeZoV.png and offers may ways to sort based on answer status i.stack.imgur.com/61MTb.png so I'm not yet sure what it is that you think should be done differently.
– uhoh
Jun 29 '21 at 0:11