37

Stack Overflow was the first SE site, and many mechanisms were designed for SO and not necessarily for all the different sites that came after it. The "accepted answer" concept is one of those mechanisms that exists pretty much unchanged for years now.

On a site about problem solving, which SO is, the concept makes sense. The asker gets a few answers, tries them out and accepts an answer that worked for him. When you're dealing with a programming problem, the asker can in almost all cases verify if the answer worked for them, in many cases you can just copy the code, run it and check the result.

The situation changes dramatically on sites that are not really about problem-solving. I'll use two sites I'm familiar with as examples, but there are probably more. Skeptics and the new Health site are sites where many questions are asked by users that do not necessarily have knowledge needed to evaluate which answer is better. On Skeptics there is the additional problem that for some controversial subjects, users tend to accept the answer that fits better to their own world view, not the better answer overall. This leads to worse answers being shown first, instead of the highly upvoted but not accepted answers.

Having an accepted answer doesn't really provide any value on such sites, the askers don't have an easy way of verifying answers in most cases. Voting alone should put the best answer at the top, and letting the asker determine the sorting order can be problematic.

The accepted answer concept has too much effect on reputation gain for it to be removed entirely without causing a lot of issues. But the changed sorting order for accepted answers is something that could be changed without significant side-effects. I know this has been declined before, but I think the situation for some sites that differ fundamentally from SO is different enough to warrant revisiting the decision. The solution might be something like a per-site toggle, or a global change. I'd prefer changing this globally to minimize confusion, but there are certainly other options possible.

3
  • 9
    I would recommend that, instead of removing the effect of accepting on sort order, you limit it to one-and-a-half times the effect of a regular upvote, giving a one-rep asker somewhat more control over the top answer slot than any one voter, but not much more. Coincidentally, this is the same ratio that acceptance rep gain has to upvote rep gain. Commented Apr 23, 2015 at 16:24
  • 3
    Code Golf questions, with their "indisputable" requirements, also make one wonder what's the point of accepted answer
    – gnat
    Commented Apr 23, 2015 at 17:47
  • I would be keen to see your thoughts as an answer or comment on my proposal for an alternative/complementary way we could deal with outscored but accepted answers being always first in the sort order forever: meta.stackexchange.com/questions/268666/…
    – PolyGeo
    Commented Nov 1, 2015 at 8:54

1 Answer 1

10
+25

As noted in the question, on a lot of sites there's not a lot of signal in knowing which answer is an asker's favorite. But there is some signal, so we still want to make it easy to see this. So we should order answers on the "votes" view by, well, votes, and also provide a way to jump to the accepted answer easily. Here are a couple ways we could do that.

Add a link under the question to jump to the accepted answer, wherever it is on the page:

add link under question

Alternatively, add a fourth view to the answer list:

add a view

The "accepted" view would have the current behavior -- accepted answer first and the rest by vote order. The "votes" view, which would remain the default, would place the accepted answer by its votes, like for self-answers now.

(The new links aren't really red, of course; that's just to highlight the interface change. They'd also, you know, match font and size and be properly lined up.)


Discussion

Either of these changes has the feature that the casual visitor to Stack Exchange will see the top-voted answer first and will probably be satisfied with that. But for anybody who wants to know, the answer that best satisfied the asker is but one click away (or some unknown amount of scrolling).

I would suggest trying this out on sites that are most prone to this problem and then, later, deciding whether to apply the change on all sites.

As an asker, this change would make me feel better about accepting answers on questions where answers can't come with proof. I can't prove which approach to a workplace problem or a writing quandry or a matter of biblical interpretation is globally correct; I can only signal what satisfied me. I'd like to be able to do that without interfering with the placement of other answers.

You must log in to answer this question.

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