16

In my opinion, most new users who post their first question generally fall into one of two categories:

  1. People who post the question, get an answer they like, then ignore the site whatsoever. This is obviously counter-productive to the community, but there's not really much we can do.

  2. People who care about the site's rules, post the question, follow answers and comments, participate in the discussion, upvote, etc.

The problem I see is that most newcomers of the second category accept an answer to their question within minutes of seeing the first one (just my observation). And on a site that hosts questions that are more subjective than these on SO, this is a bad thing (again, IMO). Honestly, I'd say even on SO this is a bad thing, but this is off topic. Many people don't bother answering questions with an accepted answer. And even if they do, the OP is less likely to change the accepted answer. And even if he did, this has stopped some potentially good answers from materializing at all. Furthermore, the answer/question ratio is used as a metric for new SE sites (and a healthy value from what I understand is about 5-6).

So, shouldn't there be something in the FAQ that mentions to new users they should not mark the first posted answer as accepted, just because it makes sense?

I mean, great answers take time to be written. On my questions, I wait until either the question seems to have little new views, no new answers and few comments, or I've seen an answer that covers the question in its entirety flawlessly.

Is this a real issue, or is it just me?

  • @faif: another noticeable side-effect is that once a question has been accepted and becomes the first result, the voting trend very quickly inverts. For instance, my answer on programmers.stackexchange.com/a/152734/3631 was originally voted a lot higher than the accepted answer, so much so that I earned the populist badge for it. But then it took only a few days after the answer being accepted for the balance to be reversed. Accepted answers clearly get a lot more "autopilot" votes where people like a question, like an answer, but won't consider even the second or third ones. – haylem Jul 4 '12 at 5:57
  • @faif: note that I don't mind that the accepted answer appears first (as some seem to think that's unfair, if others have more votes): I agree with the premise that it should be first, as the asker is the one who chose it. I'm just pointing out this unfortunate psychological bias that makes us more likely to choose the first entry, especially if already marked as accepted. Don't think we can do much against it. – haylem Jul 4 '12 at 6:04
10

I don't see this as a real issue.

I guess we could block people from accepting answers for X days or hours, but the immediate downside to this would be that even more 1st time users wouldn't return to select an accepted answer. Or they would just realize once they start asking more questions and people point out their poor accept rate to them (or they might re-create new accounts).

So, I think changes would make things works, and that things are OK as they are, from a tooling and process perspective.

Maybe, as you mention, could it be made clearer to the users in the FAQ, but I'd bet a rather small percentage of first time users read the FAQ, especially for the "I post and want an answer fast and don't even care if it fits the scope of P.SE (or other SE sites)".

The only good alternative I'd see would be to have, when the first answer to be posted appears (as a live refresh or in their inbox), a notification / pop-up pointing them to guidelines about accepting answers to address the common issues (like accepting too fast, upvoting AND accepting, etc...).

But so far I have to say I haven't been too bugged by "fast"-closes. You can't force askers, who may not be here to commit to the SE community and may just be in search of a one-off answer, to do just that: commit to following a long list of guidelines. It's already good enough (or at least not bad) if they at least post decent questions, format them properly without using SMS-style and poor grammar, and give proper feedback and accept an answer.

  • I agree, forcing is definitely not the solution. But I think mentioning in a soft way to the people who may hear might improve the answers. I realize this is just a gut feeling, but I think new users who care for the rules feel obligated to accept an answer fast. And that's just wrong (IMO) – K.Steff Jul 1 '12 at 21:57
  • 2
    @K.Steff: the only way I see to do this is what I suggest at the end: have the site give them some instructions when the first answer is posted, so they are more likely to head over there to read what's next in the process. That's not a guarantee, but that should be rather ergonomic and increase the chances of them following the guidelines. But I don't know if the SE site can do this or if it requires changes. I have no clue about how customizables these SE sites are, and who is in charge of this. I assume you need to persuade mods and they bring it up to SE. – haylem Jul 1 '12 at 22:03
  • I agree with you 100%. I'm not really sure if it's an issue. But I feel even an epsilon improvement is something. And a message could help in my opinion. – K.Steff Jul 1 '12 at 22:14
  • @K.Steff: let's see if that's worth asking a mod on the chat then. – haylem Jul 1 '12 at 22:19
  • 2
    @K.Steff: I don't really know how to get in touch with mods honestly, so I just asked on the chat. Hopefully whenever they connect next they'll see it. I've seen that you can tag questions as "feature-request", which I think brings it to their attention, but I'm not sure it's the proper way (if not, they may not like it if they feel they have to untag it, so I'd rather wait until we get feedback as it's not a pressing matter). – haylem Jul 1 '12 at 22:39
  • @K.Steff: haven't received any answer from the mods on the chat (or if I did, I didn't get a notification. I'm not too familiar with the chat, honestly. I find it rather unfriendly to use). I might try to request moderator attention, but that seems like an odd thing to do on Meta :) – haylem Jul 4 '12 at 5:59
  • This answer has been accepted too fast :) – user2567 Jul 5 '12 at 8:42
  • @Pierre303: indeed, that seems to have caused some confusion :) – haylem Jul 5 '12 at 9:38
  • @K.Steff: see the comments under ChrisF's answer – haylem Jul 5 '12 at 10:56
4

There's nothing much more that I can add to this.

As haylem points out there isn't anything we can do to force people to accept or not accept answers. There is a time limit built in - but it's not very long.

There is nothing to stop you providing another answer if you think the accepted one isn't good enough. If others agree it will get votes and other visitors to the site will see an answer with more votes than the accepted answer and (hopefully) reach their own conclusion as to which is better.

Don't forget that all acceptance means is "this answer helped me (the question asker) the most". It doesn't make the answer more (or less) correct than other answers.

  • 1
    I think there's a confusion because the OP accepted my answer too early (ironic...). I was the one trying to flag it for review by a mod. The question we have (see comments under my answer), is how to file for a feature-request to have intructions shown to an asker when they visit their own question after the first answer has been posted. They would most likely read these (short) instructions and learn how to correctly rate and accept answers. It requires a relatively minor changeM: it doesn't change the way voting works, but takes the asker by the hand to learn. Can we request this? How? – haylem Jul 5 '12 at 9:36
  • @haylem - ah. Post a separate feature-request for this. However, it's likely to get status-declined. – ChrisF Jul 5 '12 at 9:43
  • thanks, that's what we were wondering (or if we could re-tag this one as "feature-request" or if mods were the ones supposed to do this). But I guess writing a better proposal for a feature request would be in order. I'll transmit to OP. – haylem Jul 5 '12 at 10:56
  • @haylem Seemed only fair I demonstrate the issue at hand... (jk of course). I was hoping to generate some discussion as to whether this is deserving of a 'feature-request'. – K.Steff Jul 5 '12 at 14:17
2

Don't worry about accepted answers being chosen early, its just an opportunity for some of the more elusive badges.

  • 1
    Only if this site is a thing to be gamed, which is fundamentaly wrong. It should be about participation, and codification of knowledge. And the point system is just an incentive and a method of determining "authority". – Zoran Pavlovic Jul 5 '12 at 9:43
  • 2
    @ZoranPavlovic: it's a good point, but points and "achievements"-system work both ways: they are meant to be motivational AND to express hierarchy. However, as soon as you introduce a metric, you can expect that some individuals will be overly motivated by this aspect and try to game it. It's a bit of a double-edged sword: you base motivation on ego, but as the same time should try to aim to keep it as ego-less as possible to have impartial contributions (same problem as with games in development teams like continuous integration games) – haylem Jul 6 '12 at 8:37

You must log in to answer this question.

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