I originally posted this question on the Philosophy.SE Meta site, but someone recommended that I post it here instead to get a more varied set of responses from experienced users. I'm aware that it's similar to this question, but that one is fairly old and I imagine the system has changed much since then, not to mention policy; plus, I want to generate new discussion and ideas.

I'm still relatively new to this site, and although I used to frequent other SE sites going back several years, I haven't come across anything that suggests to me that there is something in place to prevent a "cementing" of acceptable answers to questions that may have better answers added later on, but because of several factors will never get voted up past the original answers.

Allow me to clarify. Suppose I see a question on the site, and a particular answer given has several upvotes and it's not really a bad answer, sure. But suppose you know the topic even better and are able to write a more thorough, complete answer to the question that really is, in fact, a much better answer. Unfortunately, there are just too many variables related to human performance and system design that prevent your answer from ever reaching the top in terms of votes, not to mention being the new "accepted" answer. I have identified two major factors below:

  1. Operator fatigue

    With questions that have many answers, particularly a large collection of lengthy ones, users will tend to—as a statistical mean—ignore the lower down answers and only read the top ones until they become otherwise bored (move on and do something else) or satisfied with one of the earlier answers (and upvote it). They may accept a mediocre answer simply because they desire to accept something (particularly on a popular answer) but don't want to have to read every single answer in the list.

  2. Question maturity

    Old questions will tend to get re-evaluated less and less, at somewhat of a positive exponential rate based on how old the question is. The great thing about recent questions is that the wealth of users are willing to give it a once over and vote. But people might not want to reevaluate an old question they already looked that, and if the question is particularly old, the user might not even still be active. Thus, even though a newer answer might be better, the first answers will have so many upvotes that, without a constantly growing stream of members or perfect re-evaluation from the users, new (better) answers will be hard pressed to pass old (not as good) answers.

Is there any policy, system design, or other implementation in place to remedy such a drawback? Or is this simply accepted as an inevitable part of dealing with humans (as it is in many systems)? I know you can sort questions by "active" which definitely helps, although the default and perhaps most useful sort is by votes, useful especially when you just want to know the answer to the question and don't really care about contributing. As an engineering psychologist, many potential solutions to this come to mind, both in policy and system design; I was just curious as to what is in place here.

Note that the problem I outlined here applies most to the SE sites where a specific, precise, technical answer is not feasible (i.e. Philosophy). There are only so many ways to write a specific line of code or show how 1 + 1 = 2. But on several of the SE 2.0 sites, there are many ways to answer a question, and answers will tend to have to be weighed against each other as opposed to simply being checked for accuracy. That is, two answers can be correct while also being fundamentally different responses.

  • The existence of the Necromancer and Populist badges tells you that this pattern does exist, but isn't total. There are times when someone writes a great answer long after the question is asked, and times when a non accepted answer is better than the accepted one. Perhaps looking at the relative frequencies of those badges on various sites would be illuminating. Aug 9, 2011 at 10:57
  • 1
    I can't say with any certainty how this will work on Philosophy, as the site is still pretty new and there isn't as much activity as we'd like. But my personal experience on Stack Overflow is that newer, better answers are considered--not only are they frequently upvoted by the community, but sometimes the asker will change his/her accepted answer (if still visit the site occasionally). This is all encouraged behavior (as Kate mentions, we give out badges for it), and the system enables it by "bumping" old questions when new answers are posted. That helps ensure they'll be seen and upvoted. Aug 9, 2011 at 11:09
  • 2
    Are you perhaps overthinking it a little? Maybe you should just answer and see what happens.
    – user1228
    Aug 9, 2011 at 12:13
  • @Won't - This isn't in regards to a particular "answer" that I'm afraid to post because it won't get looked at. It was purely an observation.
    – stoicfury
    Aug 9, 2011 at 13:04
  • 3
    @stoicfury: Sorry, missed that, was staring at my lint hole. My comment was a general answer. When faced with an older question with an accepted inferior answer, don't brood over the sorry state of man and the meaninglessness of his actions compared to the magnitude of the cosmos and the eventual heat death of the universe. Just answer the damn thing and take pride in the fact that you don't care about rep points, but that a question asked was answered well.
    – user1228
    Aug 9, 2011 at 13:12
  • 2
    @won't - I'm an engineering psychologist philosopher — all I do is overthink :P Still, seeing as I enjoy my time contributing here, as many of us do, I figured it can't hurt to try to improve things here and there as opposed to merely "hoping" the system will sort itself out. Maybe some clever person will post a genius solution and then we can have the best of both worlds — faith that the system will work itself out and confidence that the most voted answer to a question you are looking at is actually the best answer. :)
    – stoicfury
    Aug 9, 2011 at 13:48


You must log in to answer this question.