7

Situation:

  • You see an answer

  • The answer is generally good. Extensive research, in-detail review of the claim topic, references, etc...

  • However, one small aspect of the answer is awful - doesn't have solid references (but has enough to avoid being edited out by moderators), and/or uses logical fallacy, and/or can be refuted by better references

    • Basically, if that small aspect was the whole answer, it'd be a clear "downvote!" material.

    • moreover, despite the scope of the bad content being small, it is important - either in isolation (e.g. it's denying Holocaust, or defending slavery), or even worse, because it is an important data point for the overall answers' conclusion, and refuting that data point either fully invalidates the conclusion, or makes it less reliable.

  • The poster, upon requesting that the "bad" nugget in the answer be addressed, refuses to do so, explicitly - and they (or moderators) refuse edit suggestions - rightfully - because the edit contradicts the poster's intent.

What should be done in such a situation to address it effectively?

It seems that 100% of actions I'd consider as possible are either wrong or not effective (based on my extensive practical experience on this site and SE overall):

  • Upvoting the answer has a clear downside that it rewards the "bad nugget" content and countra-incentivizes the poster from fixing it.

  • Downvoting the answer is not terribly fair (the good content largely outweighs the bad) - and more importantly, is not effective since a large majority of the users upvote said good content and ignore the bad content

  • Informing the poster, as noted in the preconditions, was not effective. Ditto for trying to edit.

  • Flagging for moderators is a wrong action since the bad content is not moderator-action-awful grade.

  • posting your own counter-answer refuting bad content seems bad/useless - the scope of the "bad nugget" is small enough that merely restricting your answer to that nugget would make your answer "not an answer" to the main claim, or at best, a mediocre answer due to small scope. And anything above that small scope is already 100% answered by the poster in their answer, well.

  • posting your own counter-answer which is a clone of full original answer with bad nugget fixed seems at best awful and at worst plagiarism.

  • I used modertaion in a sense of "community moderation" as opposed to "diamond moderation". Retagging in better ways is welcome. – user5341 Apr 24 '15 at 0:56
  • (BTW, i'm especially curious since I've been on both ends of this situation recently, so I understand the frustration for both sides) – user5341 Apr 24 '15 at 0:57
1

I think the second-last suggestion is best. You can write a small new answer which says something like,

One detail that X didn't mention in their otherwise-excellent answer was Foo.

Foo is referenced at Bar and it's relevant because of Baz.

I disagree with Oddthinking's answer and recommend not reworking X's answer, because:

  • It does readers a disservice to make them read two copies of X's answer.
  • It's not worth trying to take credit (votes) for any mere copy of X's answers.
  • IMO such a supplementary answer is (if the addendum is worthwhile and on-topic) itself sufficiently on-topic that it's not necessarily deleted by moderators
  • IMO such a supplementary answer is worth posting, to help the OP (even if it's a "a mediocre answer due to small scope" instead of being glorious).

Note that Foo must be relevant to the question. If it's relevant to (i.e. some criticism of) something in X's answer but not directly very relevant to the OP's question, then perhaps post your opposition as a comment to X's answer instead of as a competing answer to the question.

This type of supplementary/auxiliary answer is very common (IMO) on the software programming forums.

  • We've had occasional trouble where the original question is deleted, leaving an orphan supplementary answer. – Oddthinking♦ Apr 24 '15 at 23:16
  • Yes. Here the OP suggests that the other answer is "generally good" and that its author doesn't want to change it (therefore perhaps likes it and wouldn't want to delete it). Also if it were deleted then the balance changes, and you could 'save' the 'orphan' answer by 'cannibalizing' from the deleted answer (which would still visible and therefore available as a source to a high-rep user like DVK). – ChrisW Apr 24 '15 at 23:21
1

The problem here is balancing out what a user think is wrong, and another thinks is right. If there are no references, moderators can intervene.

If it's a grey area, I'd say leave a comment. Yeah, it won't work, but it doesn't matter. Your correction is there with the answer now.

Comments do influence votes, and sometimes authors actually... correct their answers. Frank but courteous feedback, focused on the matter at hand seems to work best: never say "you are wrong", but "this sentence is wrong" (or unsupported). Beyond that, we can't really "fix" all answers because, in many cases, there are tons of things on which we disagree, and no one is the wiser -- even though we all think to be honestly right!

Of course if it is true that a small particular invalidates the whole answer, then point out the mistake, downvote, flag for mod attention and move on. I think it's best if a diamond mod or a community manager takes the blame if we have to edit out something a bit atrocious that the OP will possibly resent or delete the whole answer.

  • 1
    The question was theoretical, but the practical impetus was based on 2 specific posts, in both of which the moderators explicitly expressed that there's "not enough badness" in content for the moderator to intervene to fix. Also, while I agree that the comment is the best approach in a general case, my question was very specific to a situation where the poster is actively opposing the commented suggestion (due to honest intellectual disagreement with the commentor, not out of nastiness) – user5341 Apr 24 '15 at 15:36
1

Write your own answer. If it happens to be a copy in part or a derivative of the lower quality answer (and it doesn't necessarily have to; you can do your own work de novo) give credit, but that is all the creative commons licence requires.

Writing a good answer is never a disservice to readers. In the long run, it will get up voted so that readers that don't have the time or interest to read two similar answers won't. Those that are interested will. Your answer should stand on its own as a response to the question and not be a response to another answer (because the other answer might change or be deleted, among other reasons).

Here is an excellent example of the use of an alternative answer to address issues with the original answer: Were Mermaid passengers shipwrecked by 5 different vessels before being successfully rescued?

Both answers are interesting, well referenced, and stand on their own. Reading both gives insight into how the original answer perhaps got the wrong conclusion, and even into how the myth originated and was perpetuated.

Also, flag the low quality answer for the moderator to add the "some parts of this answer need additional references" banner to warn readers that not all of its assertions are well supported by evidence. Eventually, the mod can delete un/under-referenced material. (This is assuming that the material and its references don't meet this site's standards. If they do, then ignore this paragraph, but then I don't understand your question's premise.)

Be bold, create good content, and trust authors to have released their work under a creative commons licence willingly so that you can leverage their hard work to make better things.

  • Which part of the question's assertion that "doesn't have solid references (but has enough to avoid being edited out by moderators)" leads you to believe your last paragraph is ontopic to this question at all, please? – user5341 Apr 24 '15 at 15:03
  • Also, just because stealing (*cough*copying*cough*) 95% of someone's work is legal under CC, i am fairly confident there are ethical angles besides legality to consider here. -1 for not considering that. – user5341 Apr 24 '15 at 15:05
  • @DVK Maybe we have different mental models of the reader. Perhaps Odd and nomen think that the reader will only read one answer (the most upvoted and accepted one), therefore that answer must be perfect and contain everything. Whereas I think that an interested reader will read many upvoted answers and it's easier/better for the reader if the answers don't contain duplicate information. Posting semi-duplicate answers is I think unconventional (people avoid doing it) and discouraged (often someone will post comment saying, "your answer repeats this other answer, which was posted before yours"). – ChrisW Apr 24 '15 at 15:12
  • @ChrisW - this is simply tested. Go to a moderator's incredibly popular (and recent) answer on a hot topic question. Steal 95% of it and say that you disagree with 5% remaining and here's the references for disagreement. See how long your "improved" answer lasts (preferably, do it outside Skeptics since the pool here is contaminated by this post to be a clean experiment) – user5341 Apr 24 '15 at 15:25
  • @nomenagentis - you don't see any ethical implications in taking someone's work that took several hours of research and editing, and using 95% of it unbidden? – user5341 Apr 24 '15 at 15:30
  • +1 because I agree with your concrete example i.e. the Mermaid answers: they worked well together. They both/each made strong contributions, and sfaik the second did not copy 95% of the first (was not, as described in the OP, "clone of full original answer"). – ChrisW Apr 24 '15 at 17:03
  • @DVK No (I won't attempt that test), I agree with you that stealing 95% of an existing answer is abnormal and frowned upon (in my experience) and not the best way to handle this situation (in my opinion). Therefore I was inclined to recommend a "supplementary answer". – ChrisW Apr 24 '15 at 23:47
  • @ChrisW: One of the great advantages of Stack Overflow model is that I don't have to read each and every post in the thread to get an answer -- each answer should be self-sufficient -- all moral rights are gone as soon as the author of the original answer has refused to fix the 5%. – jfs Apr 26 '15 at 21:22
  • 1
    @J.F.Sebastian If I'm interested in a topic I will read each and every post in the thread. Even on Stack Overflow I'll read other posts, before I try to use/implement the accepted answer. – ChrisW Apr 26 '15 at 21:25
  • @ChrisW: you can read whatever you like. It doesn't mean that it is ok to make signal/noise ratio worse for the rest of us. – jfs Apr 26 '15 at 21:37
  • 1
    @J.F.Sebastian Copy-and-paste would make the S/N worse for me; and having separate answers in an example like this one is ok. – ChrisW Apr 26 '15 at 21:42
1

Your question eliminates all the possible situations which would deal with a good answer that has some flaws. Here is what one should do if they find a good answer which has some flaws.

Let's start with:

Upvoting the answer has a clear downside. Downvoting the answer is not terribly fair.

Ideally, a good question (even if it flaws) should be upvoted and a bad answer should be downvoted. Simple. Upvotes and downvotes are irrelevant when you have flaws in a good answer. You can hold your vote until the problem has been resolved, then decide later. At worst, you can not vote at all.

You go on saying:

Informing the poster, as noted in the preconditions, was not effective.

This isn't true. Notifying the answerer via comments and telling him what is the weak point in the answer works. If the comment was effective, it will be upvoted and there you have made a successful notice.

Informing the poster politely is the most effective way to deal with a good answer having flaws.

Example:

I wrote an answer on this question, it was okay but had a vague part, so this is what the moderator did, and in few minutes the weak part of the question was fixed.

enter image description here

If the answerer didn't reply to your comment, attempt to edit the answer yourself if possible. If the answerer doesn't like it, he will rollback.

My question was very specific to a situation where the poster is actively opposing the commented suggestion

If answerer opposed the commented suggestion due to honest intellectual disagreement or due to nastiness, then you can downvote the answer and walk away. It is not your answer and you have no right to force your views on the answerer.

For example:

Was a pair of handcuffed hands found near the World Trade Center after 9/11?

I referenced the one single source and the official report of the events leading up to the September 11, 2001 attacks; it referenced 900 articles inside. So, I concluded inconclusive. Members disagreed with the conclusion since I am examining only on source, I opposed their commented suggestion since I didn't agree with them and I told them to downvote my answer and walk away if they don't agree with my conclusion.

Flagging for moderators is a wrong action since the bad content is not moderator-action-awful grade.

Irrelevant to post this point. Flags should not be used to indicate technical inaccuracies, or an altogether wrong answer. I've done that once, and received a decline:

enter image description here

Posting your own counter-answer refuting bad content seems bad/useless

Agree.

Posting your own counter-answer which is a clone of full original answer with bad nugget fixed seems at best awful and at worst plagiarism.

Agree. You shouldn't copy the answer's answer and post it as your own or post a similar one that show similar sources, main points; it would be unoriginal. I rarely see that anyway on skeptics.se.

If you want to post a follow to up the answer, and add new value because you disagree with the conclusion or some point of the answer, then this is fine; since you're not copying the work but adding an additional value to it.

Example:

Did Pope Gregory IX call the Holy Fire a fraud?

I have digged many books to find the original source of Pope Gregory IX letter which I posted in my answer.

I have concluded that it is true and my answer was ticked. But later ChrisW found the letters, translated them and added a new value to the answer:

ChrisW's answer: enter image description here

So, I changed my conclusion:

enter image description here

Takeaway: If you notice a good answer with flaws, do the following:

  1. Comment so that the answerer would be notified.
  2. If the answerer ignores, you try your fix the question yourself.
  3. If the answerer agrees and fix, you win.
  4. If the answerer disagrees, downvote and move on.
  5. If you have some points to add, you may add a follow-up answer which has new points.
-3

I personally think the last suggestion is the best.

Post your own answer, based on the original. Openly, proudly, acknowledge your sources. Use quote marks so you are not claiming the work of others.

When Stack Overflow/Stack Exchange started, it was suggested that you should consolidate several answers to make a better one. [citation-desirable]

The acceptability of copying answers is discussed here.

You must log in to answer this question.

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