16

After having joined the site and been visiting regularly for a while, I have discovered many, many cases of duplicate questions (i.e. questions having the same answer, per apparent consensus definition for this tag).

I've finally collected enough points to start voting-to-close these old questions, but it appears that at least some users feel that I'm sometimes doing this the "wrong way."

I don't want to cause undue stress on moderators or queue reviewers, nor do I want to abuse the newly-granted ability to VTC. I am aware (I think) of most of the consensus decisions on how duplicates should be handled, and I am in agreement that obviously superior pages should have precedence, even if they are newer. However, in several cases, I have identified a group of 3 or more questions with the same answer, and in such cases, I lean heavily towards the oldest formally accepted (i.e. green check-mark) answer as the target to which other questions should point.

My reasoning is:

  1. "Earliest formally-accepted answer" is an objective feature that will lead to one and only one question. Formally-accepted is better because someone very new to the site can tell at a glance the OP got a satisfactory answer (and therefore it is worth it to keep reading).
  2. "Obviously superior" is a subjective call; if it's not at least 5x better in my opinion, I'd rather stick with the implicit guidance of the duplicate banner's "this question has been asked before" statement.
  3. Sometimes answer quality and question quality point in different directions, further complicating a subjective evaluation.
  4. "Credit where credit is due" seems applicable; newer duplicate answers very often don't point back to the original successful answer (though the recent trend seems to be to do this).
  5. Anyone coming to the site in response to one of the question pages (presumably the one that most closely matched their particular search terms) will be directed to a common "central page" for the group, from which all can be visited via the "Linked" pages section at right (and will thereby itself have aggregated many applicable search terms from the other questions' titles). As it stands, in certain cases, duplicate banners point to pages that are themselves closed as duplicates.

Are there points of consensus (of which I'm not aware) that govern situations like these? Does anyone have a better methodology to suggest? Would it be appropriate to make a feature request to have duplicate banners say the question has been answered "more comprehensively" instead of earlier?

EDIT: Given the significant number of upvotes and the lack of competing answers, it seems Adamant's answer represents consensus. However, I haven't accepted it because of the open questions related to situations where the "best" answer for a group changes (see my comments on it below).

Adamant notes that "[d]uplicates of duplicates [are] definitely an issue" (italics mine), but in actual practice it seems this may be ignored in favor of "better answer" reasoning. (As an example from earlier today, see Main character can "manipulate" his own time , which had a VTC overridden and was set as the duplicate target for Story about main character with regeneration powers despite the fact that the latter already had four other questions pointing to it as a duplicate.)

While I don't disagree with the assessment of the Feb 26 2013 question being the "best" answer of the duplicate group, is it really the recommended procedure that every one of the others be reopened in order to reclose them with a new target? Or is it really the case that there is not a functional consensus that duplicates of duplicates are a significant issue to be managed?

Please note that in this particular example case, the target chosen was an older question which had most likely been missed because of the title Treason being used, instead of A Planet Called Treason like others in the group. (By my own reasoning, being older is a strong preferential factor.) That wouldn't necessarily always be the case, so I don't consider it relevant to the general case discussion I'm trying to evoke.

3
  • Specifically identification questions, at this point? Groups are rare, otherwise
    – user31178
    Oct 7, 2016 at 4:52
  • @CreationEdge - Yes. I've updated the question title to make this clearer.
    – Otis
    Oct 7, 2016 at 4:54
  • Anent conflict between question quality and answer quality: If the answer is bad, we can try to improve it, or else write another answer. If the question is bad, or rather it's a good question but the OP misremembered some major points, there's nothing much we can do about it.
    – user14111
    Oct 8, 2016 at 10:31

1 Answer 1

12

The duplication target should always be the better question1

First, it’s probably worth noting that selecting the better question as a duplication target is both SE consensus (as well as the consensus on this site) and the recommendation of the Community Team.

There are a few reasons to close the newer question as duplicate

This has been touched on before in this question (among others). Basically, we should generally close the worse of the two questions as duplicates. That, said, assuming that two questions are of similar quality, there are two big reasons to close a newer question as duplicate.

  1. To discourage people from asking detailed and well-written questions that they know to be duplicate, in the hopes that these questions will get lots of upvotes and won’t be closed.
  2. Because, all other things equal, a new question without answers isn’t as good as an old question with answers. If the older question is really terrible, it might be worth considering choosing the newer question as the dupe target, but otherwise the overall quality of question plus answers is simply better on the old question.

None of these apply in this case

In the situation mentioned in the question, we have a whole pile of old questions, all of which must already have at least one answer (otherwise we could not close as duplicate, per our policy on closing questions).

Thus:

  1. People aren’t getting any more upvotes by the time the questions are eligible for closure.
  2. All the questions already have answers, and so are of comparable quality in that respect.

So we have to ask ourselves, what do we want?

  • We want people other than the original asker who are looking for the story to find the best-written answer with the most details, so that they have the best chance of identifying it.
  • We want people who are reading answers out of interest, or to find books that they might want to read, to be directed to the most interesting answer, the one that is most likely to help them make an informed assessment.

All of this points toward choosing the question with the best answer, regardless of the age of the question.


That’s the general argument for how to close questions (assuming both have answers that have been confirmed by the OP, of course). But to address the more specific points you made:

  1. The earliest accepted answer is certainly an objective criterion, but that does not make it the best. Following that criterion can lead to a lot of questions with excellent, detailed answers being closed as duplicates of questions with one-line answers that do little more than state the name of the work identified. Subjectivity, in this case, is better for the site.
  2. The same principle, really. Yes, which answer is superior is subjective, and sometimes difficult to determine. That’s fine: in borderline cases it’s perfectly acceptable to use whatever heuristic you like. The point is to make sure that questions with good answers are not being closed as duplicates of those with mediocre answers (or even worse, questions with mediocre answers being closed as duplicates of others with mediocre answers).
  3. Yes, sometimes answer quality and question quality point in opposite directions. Except in cases where the question is significantly more detailed than the answer, I’d stick to answer quality for questions.
  4. Everyone deserves credit, most likely: chances are each answerer identified the story individually. And don’t people also deserve credit for writing a comprehensive, high-quality answer that makes it easy for the OP and future visitors to determine whether something is the correct story, rather than having to do their own research?
  5. Duplicates of duplicates is definitely an issue, but not really so pertinent when it comes to which of a group of open questions to set as the dupe target. If the best question is already closed as a duplicate, and you think it would make a better dupe, casting a reopen vote (and leaving a comment explaining why), asking in chat for help reopening, or asking a moderator are all options that can be pursued (though it is often unnecessary).

1: Usually the question with the best answer.

7
  • 1
    On point 3. We have at least one story-identification question with literally the opposite details, thought I don't know which is set as the dupe-origin.
    – AncientSwordRage Mod
    Oct 7, 2016 at 11:32
  • @AncientSwordRage What question is that?
    – user14111
    Oct 8, 2016 at 4:34
  • @user14111 I can only find this: en.wikipedia.org/wiki/All_Summer_in_a_Day not the posts on here.
    – AncientSwordRage Mod
    Oct 8, 2016 at 19:44
  • @Adamant I agree with your sentiments fully.
    – Möoz
    Oct 10, 2016 at 3:02
  • @Adamant, this seems to be a well-supported answer. Would you mind extending your answer a bit to recommend a procedure for situations where a subset of a dupe group already has banners pointing to answer A, but a newer and "better" answer B has been created? Would it be better to link B back to A or relink everything pointing to A (and A itself) to B?
    – Otis
    Oct 13, 2016 at 15:49
  • @Adamant, since you haven't expanded your answer, I assume that you think this situation is covered by point #5? It seems unlikely that submitting a bunch of reopen votes specifically to reclose with a new duplicate target would be considered desirable. Certainly it is possible that the "best" answer of a dupe group is subject to possible change with every new answer created. (Though, in fairness, it's also possible that the first formally-accepted answer can change, even if it's less likely.)
    – Otis
    Oct 21, 2016 at 16:36
  • @Otis - I wouldn't sweat it. Unless the OP is a precious flower, the fact that their (potentially months- or even years-old) question has been closed as a dupe isn't going to have the slightest impact on anything.,
    – Valorum
    Oct 22, 2016 at 18:34

You must log in to answer this question.

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