7

A lot of our questions (Especially HNQ questions) seem to get a lot of answers that repeat what others have said before, albeit with a different context. A good example is this question.

There are a few answers on there that seem to not help expand out the information further than that has already been posted. Does this matter? Should we do something about these? Do we care? Do we always want to have as many different points of view even if they are treading the same ground?

  • 1
    I don't want to get rid of HNQ, just would like to know if we should do something about these answers. I'm more used to StackOverflow where it's easy to say something is a duplicate or not, but here, it's more opinion based so can be a bit "greyer" – Draken Jan 6 '17 at 13:25
  • 2
    is it really a problem. – Retired Codger Jan 6 '17 at 15:45
  • @RichardU That's why I'm asking, is it a problem? Other stack exchange sites where it's easy to say something is a duplicate e.g. code on stack exchange, they're happy to remove it. Here, it may not be the case so I was asking for guidance in the matter. Do we mind having the same advice several times but worded in a different way? – Draken Jan 6 '17 at 15:57
  • 1
    @Draken I'm tempted to say it's not. Slightly different wording can make a big difference with some people. I also don't see any real harm in this. – Retired Codger Jan 6 '17 at 16:31
  • 1
    @MaskedMan a while ago SE folks made an exception for Workplace allowing 15K users protect questions immediately. At other sites in the network folks indeed need to wait 24 hours and only mods can protect immediately (FWIW at Physics.SE mods tend to preemptively protect questions bumped or about to be bumped into HNQ) – gnat Jan 6 '17 at 18:20
  • 1
    @gnat That's nice to know. It is still a bit of a workaround rather than a clean solution, but I guess beggars cannot be choosers. ;) Nonetheless, I updated my answer now taking that into account (and for a bunch of other things). – Masked Man Jan 7 '17 at 16:05
  • 1
    related at SE.SE meta: Answers quality in hot questions – gnat Jan 9 '17 at 12:20
  • 1
    Your answer on that one @gnat is superb and relates to MaskedMan concerns as well. – Draken Jan 9 '17 at 12:44
9

You have an XY Problem here. The most common cause of this symptom is the HNQ, which has spread across StackExchange but it affects "subjective" sites like Workplace more severely. The problem results from a number of positive feedback loops, as explained below:

  1. HNQ → more views → more answers → HNQ → ...

    When a question gets to HNQ, there is an exponential rise in number of views, which also leads to an increase in the number of answers. This sustains the "hotness" of the question and keeps it in the HNQ longer.

  2. Advice question → HNQ → anyone can answer → more answers → HNQ → ...

    Due to Bike Shed Effect, also known as Parkinson's Law of Triviality, everyone and his/her grandmother can give you advice on something. This is seen more commonly on a "subjective" site like The Workplace, as against ServerFault or StackOverflow.

    When an advice-seeking question gets to HNQ, people find it easy to jump on the answering bandwagon. By contrast, some of the more "difficult" questions, such as those that require a thorough understanding of workplace practices, generally survive this HNQ onslaught.

  3. HNQ → more drive-by voting (with upvote bias) → more upvotes on question → HNQ → ...

    Due to the association bonus, many of the drive-by voters can upvote but not downvote, which heavily skews the vote count. While details of the "hotness" algorithm are not publicly known, we can reasonably guess that vote count is a factor.

  4. HNQ → more drive-by voting (with upvote bias) → more highly upvoted answers → HNQ → ...

    As a corollary of the above point, votes on answers to HNQ question are also heavily skewed towards upvotes. Answer upvotes seems to also contribute to the "hotness" score, thus keeping the question longer in the HNQ.

  5. Subjective question → HNQ → Lengthy answer → Less likely to be read fully → More answers → HNQ → ...

    Answers to subjective questions generally include a few paragraphs of justification. The chances of people reading through and understanding every answer before posting their own answer is inversely proportional to the number and length of the answers.

Currently, a question gets knocked off the HNQ in only a limited number of ways, such as aging or getting superseded by an even "better" question. When a question has multiple positive feedback loops in its favour, that is woefully inadequate.

The example question you posted "benefited" from a number of these loops: highly subjective (that is, capable of receiving multiple "correct" answers), easy to understand (anyone could offer advice), and one which drive-by votes could relate to well.

Downvoting, flagging, etc. can cure the symptoms, but that is action taken after the trouble has already started. A better solution would be a preemptive strike on the disease, before the trouble starts.

Taking down the HNQ feature, or changing the "hotness" algorithm, is not an option because problems that don't affect StackOverflow are of lesser importance to StackExchange employees, and some of them find "entertainment" in them.

Workarounds involve breaking some of those positive feedback loops. Once a question makes it to HNQ:

Alternative workarounds involve forcing the question off the HNQ. When the question makes it to HNQ:

  • Regular users vote to close it as early as possible, which knocks it off the HNQ. Reopen it later once things cool down. Repeat as often as necessary.

  • Regular users downvote it heavily. Once the question gets knocked off the HNQ, edit the question and change your vote. I don't like this one bit, as it is extremely unfair to the asker to lose so much reputation because his question was picked up by a flawed feature called HNQ. Nonetheless, I leave it here for completeness.

These are extreme measures, but extreme problems require extreme measures. The sooner you take measures, the more likely you are to succeed. :)

  • 2
    I tend to agree... Hot Network Questions has a bad definition of "hot", and while it's sometines an interesting snapshot of why you might want to look at areas you don't normally follow it seems to have unintended and undesirable consequences. I'd almost be inclined to ask that those links set a flag which automatically disables voting and flags answers for review. – keshlam Jan 7 '17 at 19:03
  • 1
    I think Gnat's answer here might also help "solve" our issue Answers quality in hot questions – Draken Jan 9 '17 at 12:44
  • One of the most important things for HNQ is to not let questions get there that are really poll-like. Once a question has a lot of answers that are pretty bikeshedy it's almost impossible to prevent a mess. Either edits or closing. Given that you have to click through and acknowledge that your answer is adding something new to post an additional answer (I think every question with more than 2 answers here), I have a lot less sympathy for answers adding minimal new content, too. – enderland Jan 9 '17 at 15:08
  • 1
    @enderland While that may certainly be true for well established users, the temptation to post what the user thinks is a useful answer is often too great for a "nuisance" dialog box to offer any resistance. With 3 answers that are 10 sentences each, not reading them may be inexcusable, but with 10 answers that are 25 sentences each, a case could be made for more leniency. Consider this analogy: "Post one English word not found in this answer." as against "Post one English word not found in any of Harry Potter novels." – Masked Man Jan 9 '17 at 15:58
  • 1
    In addition, the reward to risk ratio is too high for that dialog box to be meaningful. To lose reputation, you have to get 6 downvotes for every upvote, and due to association bonus and HNQ, you already know the audience is heavily skewed towards upvotes or no vote. Moreover, if you do lose reputation, you can always delete your answer and get it back. There is so little to lose, closing an annoying dialog box is a no-brainer. – Masked Man Jan 9 '17 at 16:04
  • 1
    almost 3 years ago I suggested a feature to help manage questions with multiple answers: When there are many answers already, help me check that mine won't repeat others. It hangs in there ignored (not surprisingly) – gnat Jan 9 '17 at 19:38
  • 2
    If i were a conspiricay theorist(and I am), I might think that the powers that make the decisions have decided that this "Bug" is a feature that adds value to their bottom line... But thats just crazy talk from someone who knows that a tinfoil hat is not nearly good enough protection. – IDrinkandIKnowThings Jan 13 '17 at 20:39
2

I see this a lot as well, usually from new users and very often on old questions. When posting, the user should see a prompt reminding them not to repeat any of the many previous answers. My standard response is to downvote, flag for deletion, and add the following comment, with the optional bit in brackets:

This answer does not add anything new to the previously posted answers. Please remember to not repeat others, [particularly when resurrecting a question that is 2 years old.]

On a related note, would there be support for adding a custom deletion reason for repeating existing answers?

  • It would have my support, which flag do you use at the moment for the deletion so I can help get the ball rolling (Or is that a 3K only flag)? – Draken Jan 6 '17 at 13:01
  • @Draken I use "not an answer". The only options are that, "spam", "rude or abusive", and "in need of moderator attention". "spam" and "rude" should definitely not be used, as they get logged and affect other metrics. – David K Jan 6 '17 at 13:10
  • Not an answer always feels a bit wrong to me, as it is an answer and often very legitimate in it's own right. It is however a duplicate, so taken as part of a whole it should possibly be removed. – Draken Jan 6 '17 at 13:26
  • @Draken I agree that 'not an answer' is inaccurate, which is why I think a new delete reason should be considered – David K Jan 6 '17 at 13:33
  • 3
    David, flagging troublesome answers of that kind has been discussed in the past here and here and my understanding is for repetitive answers moderators recommend custom flagging with detailed message helping them easier see the problem – gnat Jan 6 '17 at 21:00
  • Sometimes this happens with established users and/or with newer questions. workplace.stackexchange.com/questions/114678/… – Lumberjack Jun 25 '18 at 14:56
2

It looks like this question generated some interesting discussion, particularly around the HNQ list. Unfortunately, it doesn't look like any of the answers directly address the heart of the question. Namely, what should we do when a user posts a duplicate answer?

We have talked a lot about what we believe are the causes, but there isn't a clear answer to "what action should we take?"

I answered a question earlier today, and about 30 minutes later an established user with a great deal of reputation posted his/her own answer that is an almost word for word duplicate.

Reading the comments from David's answer, one might conclude that flagging is the appropriate course of action, but is there community consensus around such a notion?

0

Management science is not a field where there is one right answer. Because of that, a lot of instruction takes the form of cases.

That would make me say that duplicate answers with different contexts are useful because they gauge consensus and illuminate different facets of an issue.

Allow duplicates and let the best ones float to the top.

You must log in to answer this question.

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