19

It seems like fairly regularly I'll be running a Google search to assist with a minor problem and the result that I get from a stackexchange.com site is extremely high quality. It answers my question quickly and accurately. Sounds like just what we want from StackExchange, right?

So why are these questions often marked closed or otherwise flagged as low quality? Should Google ignore these results in favor of better quality ones? Should we instead improve these low quality (but commonly viewed) questions?

While my question is very general (hence its appearance on Meta), in order to provide an example, the specific query that brought this to mind was:

Google search for redhat view json

At the time this was written, the top non-redhat.com hit was:

Is there a CLI tool that would prettify a JSON string

While it contains a perfect answer (jq and/or json_pp) the question itself was closed as "too broad." This classification by StackExchange standards seems correct. However, when Google indexes questions like that so highly (for whatever mysterious reasons) should we take the hint and provide a higher quality search result?

One possible way to improve the quality of that specific response would be a redirect to this (very similar) question which also appears (lower) on the Google search results above:

https://stackoverflow.com/questions/352098/how-can-i-pretty-print-json-in-a-shell-script

This whole issue seems to tie in with a semi-common issue on StackExchange-- should we provide information about relatively basic and broad topics, such as what someone might type into Google as a starting point?

https://stackoverflow.blog/2011/02/22/are-some-questions-too-simple/

The answer seems to be no, but we quickly get into a circular situation if we tell users to "Just Google it" and stackexchange.com shows a low quality question as the result of following that advice. This issue was discussed indirectly at people who don't use Google but I'm starting a new post since the specific question of managing responses in terms of overall view volume (implied by Google search rankings) seems to warrant its own discussion.

Since the above specific question is a top Google hit (and presumably sees more view volume as a result), would it be worthwhile to cultivate the answer and improve it over time instead of keeping it closed? Would it be useful to identify other closed/flagged pages which are common Google destinations/high volume views and also improve those?

Alternatively, should we take steps to discourage Google from indexing these closed/flagged answers as top hits? What steps can we reasonably take to do so?

  • 3
    I like this idea. – slm Apr 1 at 17:32
  • 2
    Another category of "problematic Google results" in my experience is questions which are littered with trivial or too specific details, which makes finding solutions for similar issues require more effort (harder to quickly see how similar/relevant the the questions/answers are). – sebasth Apr 1 at 17:39
  • 2
    I don't think the question was too broad in the first place. It seems like a standard software recommendation question with pretty clear requirements already, and it should just be reopened. – Michael Homer Apr 1 at 20:13
  • 1
    @MichaelHomer That may well be the case, however this broader discussion should cover high view count questions regardless of whether they are good as-is or not. Really it comes down to either redirecting users away from these questions (so the view count rate drops) or improving their quality (so those who view the question get what they were looking for.) – Steve Bonds Apr 1 at 21:12
  • 5
    I find a lot of questions with a better written Title closed but with answers and marked duplicate of a question that doesn't show on Google results at all. Managing Titles after OP has gotten answer to make questions more general and searchable seems like a good start to resolving this. – FreeSoftwareServers Apr 2 at 3:18
  • 2
    @FreeSoftwareServers That's a good point that I hadn't considered. In addition to de-optimizing search attractiveness on closed questions, we could work on optimizing their alternatives. – Steve Bonds Apr 2 at 13:46
  • 1
    Beware of "improving" those questions into irrelevance. The question you link to has already the right answer -- use jq to pretty-print json data. I really don't see what "higher quality" answer could be provided to it. – Uncle Billy Apr 2 at 18:11
  • 2
    Interesting, and I agree that the chosen reasons for many closed questions are debatable. But I think the problem space has to be narrowed 1) Closed ≠ low quality: only the "Unclear" close reason sounds quality-related. I think there is no much justification for keeping search engines away from closed Qs. 2) Beware the negativity bias: excluding duplicates, less than 3% of questions with 1000+ views are closed (similar results for 10,000+ views). (And duplicates are supposed to be a feature, not an issue, even in relation to search results). – fra-san Apr 3 at 14:26
  • 1
    @fra-san thanks for the effort to take it to the next level. I was going to check to see how pervasive the issue might be, but it sounds like you already did that. I'm glad to see that it's a small number, which is what I would expect. There's certainly some observer bias going on since I'm bound to notice more of these issues as I started to pay attention to them. – Steve Bonds Apr 3 at 17:11
6

Ideally, closure (apart from duplicates) should be a temporary state. A question should either be open or deleted. If a question is closed, it means that we aren't confident that we can provide quality answers, so we shouldn't expose those answers to the world. And a question that has no answers and no means to add one is not useful either.

So if a question has a useful answer which is on-topic for the site, but the question itself is not fit for the site, then please edit the question to match the answer. In a bit more detail:

  • If the question is off-topic then it's probably not fit for the site. Even if an off-topic question gets a good answer, this site isn't a good place for it because voters on this site can't be expected to be good judges of off-topic answers.
  • If the question is too broad or primarily opinion-based but someone managed to write a good answer, this means that the question isn't so broad and opinion-based. The question may need to be edited to be more focused or less subjective, but the proof of the pudding is in the answering. In particular, having a comprehensive answer disproves “too broad”.
  • If the question is unclear but someone managed to write a useful but perhaps partial answer, then the question should be edited to match this answer.
  • The problematic case is when there are multiple good answers that interpret the question is different ways. In this case the thread is usually not salvageable (this is why it's best to close unclear questions as quickly as possible), but it may be useful to preserve the answers by copying them to two or more new (or old) questions.

Do note that when a question is asked, it belongs to its asker, and other people should always respect the asker — if the question isn't asking what you want, post your own question. But this ownership degrades over time. Once a question gets good answers, it no longer belongs solely to the question: the community as a whole has a share in it, because the question conditions the answers. Preserving the value of a good answer trumps preserving the asker's wishes. For example, we don't allow askers to rescind their question once it's been answered¹. And if the question is closed and answered, then it's even more important to edit it, even in a drastic way, to allow the question to be reopened.

There are too many closed questions to salvage all of them. But if a closed question is a top Google hit for a popular search, it's in the top priority tier for salvaging.

On a related note, if you're working the close review queue, please check whether the question has good answers. If there are good answers, prioritize editing over closing, even if that means putting words in the asker's mouth. If there aren't any good answers and the question can't be saved without a drastic edit then the question should be closed.

¹ With one exception: if there's a single answer that hasn't proven its worth by getting an upvote.

  • 1
    I would add that significant edits (the ones which put words in the asker's mouth) should ideally be done by 2K+ users, otherwise there is a risk that the effort will be wasted when someone (rightfully!) rejects such an edit as conflicting with original intent. – Dmitry Grigoryev Apr 8 at 7:54
  • 1
    @DmitryGrigoryev if the reviewer can do such edits, editors should be able to. – Braiam Apr 8 at 18:24

You must log in to answer this question.

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