17

Browsing through my old answers (looking for a bit of code I wrote) I notice that many of them are on questions for which an answer has never been accepted, even when there are multiple upvoted answers. In some cases the OP has come back and thanked someone in a comment but never accepted. Sometimes I've upvoted and provided an alternative, for example; so it's not about rep for me.

Elsewhere within the network questions with decent answers tend to get an acceptance pretty quickly; at least it seems so to me. Is something different here?

  • 11
    In my impression we have many careless users (often the unregistrated ones) that come just for one question, it is solved and they never appear. And we have also good answers with just one crummy upvote but deserve definitely more. It is not about the rep count, but it is the appreciation of work done by real caring users! – user31729 Sep 2 '16 at 16:35
  • 1
    @Christian that's my view too, but when searching because I'm stuck, I rate questions with accepted answers highly and I'm sure I'm not alone. – Chris H Sep 2 '16 at 16:52
  • You can exploit the data site and submit a MySQL query about acceptance / answer rates (I am not very familiar with SQL, however). Perhaps there are similar queries already – user31729 Sep 2 '16 at 17:00
  • @Christian I'll see if someone's done something similar. I'm very unfamiliar with SQL. – Chris H Sep 2 '16 at 17:18
  • 2
    Ironic that the one answer has been upvoted six times, given a comment saying thank you by the original poster - and not accepted. =) – heather Sep 2 '16 at 20:57
  • 1
    Heather :) yet! – Chris H Sep 2 '16 at 20:59
  • 5
    @heather: Waiting some period until an answer is accepted is perfectly acceptable ;-) – user31729 Sep 3 '16 at 11:04
  • @heather Encouraged, even. – cfr Sep 3 '16 at 23:26
21

Here is a timeline of how things changed on this site since 2011 (I used a year-specific extraction of a data query). More specifically, it shows the cumulative questions and accepted answers as well as this ratio (or percentage) as the site grows:

enter image description here

                    With accepted 
         Questions     answers     Percentage  Growth
2011        13108       10095        77.01
2012        30159       22006        72.97      -5.26
2013        52625       35886        68.19      -6.54
2014        77097       50031        64.89      -4.84
2015       102148       63772        62.43      -3.79
2016 YTD   119297       72239        60.55      -3.01

It is more likely that older questions will have accepted answers as they've been around longer (and there's no limit on when one can accept an answer), hence taking a cumulative view on the site over time.

There is a decline in the accept percentage, but other sites show similar declines. Here is Stack Overflow:

enter image description here

                    With accepted 
         Questions     answers     Percentage  Growth
2011       2319972     1617512       69.72
2012       3983113     2678124       67.24      -3.56      
2013       6057454     3843306       63.45      -5.64
2014       8234853     4985830       60.55      -4.57
2015      10563270     6067815       57.44      -5.12
2016 YTD  12311318     6745451       54.79      -4.62

I think it's reasonable to assume that one may lose answer acceptance over time, with the decline of year-over-year acceptance slowing (perhaps plateauing) as the site grows. This may be as a result of site dilution (through increased scope) and an increased user base (which results in more questions but not necessarily more answers being accepted), amongst other things.

Related queries on Data.SE:

  • 3
    In fact a remarkably similar pattern. Thank you – Chris H Sep 2 '16 at 20:09
  • 1
    i'm not at all good at sql, so i can't tell this: have closed questions been excluded from the count? unanswered questions are likely to be closed after a certain period if the o.p. hasn't responded to comments, and also hasn't been seen on the site for a long time; this may also happen, though not nearly as often, to questions that do have a (not particularly satisfying) answer that hasn't been accepted. (very nice answer anyhow.) – barbara beeton Sep 2 '16 at 20:52
  • @barbarabeeton: No, closed questions have not been excluded. A quick query on the main site shows that there are some closed questions that have accepted answers (mostly duplicates). So, one would have to write a more complex query that excludes questions that are closed without an accepted answer. It's actually quite a number (~12K as of today)... – Werner Sep 2 '16 at 20:58
  • 1
    @barbarabeeton: ...at least the comparison to Stack Overflow is done using the same approach, comparing apples to apples, not apples to ... ducks? – Werner Sep 2 '16 at 20:59
  • oh, it's quite obvious that the two sites were analyzed in the same way. the possibility that they weren't didn't even cross my mind. participants on this site may be weird (self included), but i don't know any of them to be less than honorable. – barbara beeton Sep 2 '16 at 21:03
  • 3
    Decline in acceptance might be related to the phenomena described in dx.doi.org/10.1109/MS.2016.34 – user36296 Sep 3 '16 at 12:37
  • Can you graph “with answers, but none of them accepted” into both, for comparison? Thanks! – mirabilos Sep 4 '16 at 19:22
  • @mirabilos: Those figures are roughly the same as (questions - with accepted answer), or at least grow at the same rate. – Werner Sep 7 '16 at 4:05
  • @Werner so we have a roughly 50\% acceptance rate, then? – mirabilos Sep 7 '16 at 15:45
12

Acceptance is something that new users genuinely don't know about. Contrary to the common belief that Stackoverflow is super well-known it is basically not known outside coding circles. As these circles include more esoteric subjects some people drawn to it and ask/answer questions.

This is part of a general discussion of why we should or should not bother with new users being lectured about a wall length of text to ask a simple question. Most of the users only do accept when they are told so because they don't really bother with the site itself but only their question and don't even notice the tick mark next to the answer (based on first hand experience too).

Thus, acknowledging the user base behavior is more important than keeping an utopia statistics.

You must log in to answer this question.

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