22
$\begingroup$

It is not difficult to notice that sometimes an answer is deleted (by users other than the OP)1 simply on the grounds that it is considered incorrect by the users who voted to delete. (Probably if you follow meta, you have seen some questions sparked by such deletions. If you have access to moderator tools you might find some such deletions there - although I do not see an easy way to distinguish between deleted answer and deleted questions there, so this might be not easy. Or you might have even seen discussion of such deletions in chat.)

EDIT: Based on some comments (some of them now deleted) it seems that I did not make this sufficiently clear, so I will stress once again that I am asking about deleting answers by users other than the OP.1 Self-deletions probably deserve to be discussed separately and I believe there already have been a few such discussions.2

On the other hand, on meta.SE I read that wrong answers should actually be downvoted, not deleted. At least in the case that the answer is actually an attempt to answer the question, but it is incorrect. (As opposed to "not an answer".) Some relevant posts from the main meta: Why shouldn't I delete wrong answers?, What can be done with a blatantly wrong answer to a popular question? and When should I delete an answer? To make the reference to the network-wide policy bit more specific I will add an exact quote from FAQ item: How does deleting work? What can cause a post to be deleted, and what does that actually mean? What are the criteria for deletion? from the part "What are the criteria for deletion?"

For answers, any post that is not an answer (should be a comment, doesn't answer the question, etc.) should be deleted. Answers that are wrong or that dispense poor advice should be downvoted, not deleted.

Of course, there is wide range of what could the phrase "wrong answer" mean. Obviously, if an answer is posted which has altogether nothing to do with the question, such answer should be deleted. (The most extreme cases are spam posts, but there might be even answer which seems closer to an answer and is at least partially related to the question, but the main purpose of the answer is to promote answerer's blog or answerer's own results on theories.)

Here I mean an answer which clearly is an attempt to answer the question. It might be possibly wrong. But the OP decided not to delete the answer or simply abandoned it. (For example, the OP might disagree with the assessment of other users and still think that the answer is correct. Or the OP thinks that even if the answer is incorrect it is still useful.) In situations like this, is it ok if other users vote to delete this answer? Or is a better simply to downvote the answer (and possibly comment) but without delete vote?

Clearly, there is wide range of opinions on this. Some users argue that downvotes and comments are not strong enough indication that something is incorrect. (Especially in cases where downvotes are compensated by pity upvotes or upvotes by users who did not spot the mistake or do not have sufficient understanding of the topic.) On the other hand, sometimes there might be disagreement whether or not an answer is actually wrong. (So, for example, an answer might be deleted by a few users who consider it incorrect. And the fact that there are some other users who consider the answer satisfactory, does not prevent deleting the answer. At most, if the answer has some upvotes, this makes deletion a bit more difficult since the answer has to reach negative score first.)

Clearly, there were some related discussions in the past. For example, Wrong answer is not corrected/deleted or Deleting Wrong Answers in "Low Quality" Queue. (Maybe also this one; although the discussion there seems to be mainly about questions: Are we too trigger-happy about deleting (relatively) new posts?) But probably it is ok to have a new question discussion about this. After all, the site evolves. And, additionally, some activity on meta in the past few months seems to indicate that this is now a topic which several users have on mind.

Since from past discussions seems that users of this site has very differing opinions on this, it would probably naive to expect that some consensus is reached. But even if we at least see clearly formulated positions of some factions and if each side at least has opportunity to hear arguments for different viewpoints, I suppose this might be useful.

TL:DR; Suppose that an answer is not blatantly out of place but it is genuine attempt to answer the question - although perhaps incorrect. Is it ok to cast delete votes on such answer? Or should incorrectness be pointed out merely by comments and downvotes?


1Users which have at least 20k reputation can vote to delete answers. See Privileges - Trusted User for some basic info and FAQ on deletion for more details. An answer can be also removed if it is deleted together with a question, users with 10k+ rep can delete questions. But here I am asking specifically about voting to delete single answer, not about deleting question together with all answers.

2Here are some posts related to self-deleting incorrect answers: Protocol for deleting wrong answers., Deleting a downvoted answer?.

$\endgroup$
  • 5
    $\begingroup$ "Not all answers are created equal": I really think there are qualitatively different kinds of "wrongness". I would hope than an answer with a typo, or a miscalculation, would receive a comment to this effect, and (most usually), the answerer quickly edits to correct the minor error, and politely acknowledges the mistake in a comment. On the other extreme, there are occasional very very "wrong" answers, (even flat out contradictions) that can utterly confuse an asker, and perhaps complicate their education,. Those are the answers I may very well vote to delete. $\endgroup$ – Namaste Nov 25 '17 at 20:48
  • 1
    $\begingroup$ I just say this, @Martin, to note that I doubt there is any "overall answer" to this question that is appropriate in every context. (Not implying in any way that you asked for that!) $\endgroup$ – Namaste Nov 25 '17 at 21:02
  • 2
    $\begingroup$ @amWhy I agree. I think that deletion guidelines should specify that it should only be used for blatantly wrong or fundamentally flawed in a way that cannot be salvaged. $\endgroup$ – Stella Biderman Nov 26 '17 at 2:14
  • $\begingroup$ On more related older post: Deletion of on-topic answers only because they contain mathematical errors. $\endgroup$ – Martin Sleziak Jul 3 '18 at 13:49
  • $\begingroup$ I have an incident before where someone posted an answer which was accepted as correct and voted up several times with no other answers posted. I explained again and again that the answer could not be correct, the poster insisted there was a typo in the exercise and he would not delete his answer even when I got confirmation that there was no typo, so now the incorrect answer has to stay there and continue to confuse people in the future. $\endgroup$ – Tom Feb 10 at 14:12
  • $\begingroup$ @Tom I think that if you clearly said in the answer that it is incorrect, you have done what you could. In case you still feel that the answer needs to be deleted, you can ask moderators for help. (Of course, there is no guarantee the the mods indeed deleted the answer - depending on circumstances.) Some related previous discussions: How to delete an accepted answer? and Deleting my accepted answer and contacting the mod. $\endgroup$ – Martin Sleziak Feb 10 at 15:55
24
$\begingroup$

In my opinion, the author of incorrect answers should, when an error is pointed out, delete their answer unless (or until) it is amended. If the author feels that the error is of value to the site, it is acceptable to edit the post to include a prominent notice right at the top that it is erroneous, and explain where the error is, correctly, and preferably explain how to fix or circumvent the error.

I even go so far as to say that authors that sincerely attempt to answer the question will never leave an incorrect answer lying around without fixing or deleting it, once it has been pointed out. Hence when the error has been pointed out and yet the author refuses to fix it, then I am totally in favour of deleting it, since sincerity is indeed lacking.

$\endgroup$
  • 16
    $\begingroup$ I generally agree with this sentiment, but you are missing another important possibility, which is that the author does not understand why their answer is incorrect and sincerely believes it is correct even after their error has been pointed out. $\endgroup$ – Eric Wofsey Nov 26 '17 at 0:29
  • 5
    $\begingroup$ @EricWofsey: That is acceptable in other (subjective) subjects, but not mathematics most of the time, where it is really a matter of manually verifiable fact whether or not something has been rigorously proven. Sure, for difficult or long proofs this is time-consuming, but it cannot be that two sincere mathematicians insist that one another is wrong about whether a proof is correct and refuse to work together to figure out the problem. $\endgroup$ – user21820 Nov 26 '17 at 6:43
  • 1
    $\begingroup$ Couldn't agree more with your remark on sincerety. +1 if one is sincerely trying to post incorrect answers then there is not much hope left. $\endgroup$ – Paramanand Singh Nov 27 '17 at 8:23
  • $\begingroup$ @ParamanandSingh: The irony is the downvoters on my answer here. I cannot fathom which sentence is objectionable to those who treasure mathematics. However, I do know quite a number of Math SE users who have refused to fix errors that are pointed out to them, sadly. $\endgroup$ – user21820 Nov 27 '17 at 8:47
  • 2
    $\begingroup$ Your response to @EricWofsey seems to me to miss the point: while A and B are discussing why they disagree on whether A's answer is correct or not, should A's answer be visible or deleted? $\endgroup$ – Peter Taylor Nov 27 '17 at 8:48
  • 2
    $\begingroup$ @PeterTaylor: I did not miss the point. During an ongoing discussion in the comments, I have no problems with the answer being there. Eric asked a different question about the answerer sincerely believing their answer to be correct. I submit that this should be very rare (if at all) for professional mathematicians. I also submit that people who are not experts in a field may have so-called sincere belief that they are correct exactly like people who have so-called sincere belief that the earth is flat. In that case, experts should vote to delete mathematically wrong content. $\endgroup$ – user21820 Nov 27 '17 at 8:56
  • 4
    $\begingroup$ Just something to vent out... as a fresh user I posted an unuseful answer which I wanted to delete. Found no ‘delete’ button on my Android StackExchange app. Thought it was something to do with my reputation and tried to do other useful stuff hoping that soon I would be allowed to delete that answer. Was finally more than relieved when others deleted it for me. Finally found out that, had I used the browser instead of the app, I could have deleted the answer all along... $\endgroup$ – user491874 Nov 28 '17 at 1:03
  • $\begingroup$ @user8734617: Yes that is an often encountered problem and worse still there is no obvious indication so many users do not know this, including some who think (wrongly) that the poster does not want to delete. $\endgroup$ – user21820 Nov 28 '17 at 7:17
  • $\begingroup$ @user21820 What about things like definitions? $\endgroup$ – PyRulez Jan 18 '18 at 7:06
  • $\begingroup$ @PyRulez: I don't understand your question. What specific scenario do you have in mind? Do you have an example posts that you would like to know my opinion on? $\endgroup$ – user21820 Jan 18 '18 at 7:52
  • $\begingroup$ @user21820 I don't really have an example on hand. I was just noting that any definition is subjective (although usually easy to answer) and questions about definitions are on-topic on this site. So this answer only applies to completely objective answers. (Also note that soft-question is also subjective.) $\endgroup$ – PyRulez Jan 18 '18 at 7:54
  • $\begingroup$ @PyRulez: Yes I have no problem with soft-answers that do not claim any incorrect facts. I also have no issues with opinion-based answers about definitions that don't have clear-cut factual answers. $\endgroup$ – user21820 Jan 18 '18 at 8:05
20
$\begingroup$

Sometimes I have posted a wrong answer. (Hard to believe, I know.) But it seemed to me that the wrong method might be considered by others, so I left the wrong answer, and added a boldface This answer is wrong at the top, and an explanation of why it is wrong.

You may consider doing that when your wrong answer may still be instructive to others. But it could happen---and has happened to me---that (even though marked This answer is wrong) you still get a few more downvotes for it.

$\endgroup$
  • 1
    $\begingroup$ Yes, this is exactly the issue discussed here: Is sometimes keeping wrong answer reasonable?. (And I certainly agree with the approach you suggested.) I will point out this question was intended to be about deletion by users other than the OP. (Since it is not clear from the answer whether it was intended as an answer to this issue.) I apologize if I formulated the question in an unclear way. $\endgroup$ – Martin Sleziak Nov 25 '17 at 15:06
  • 1
    $\begingroup$ I think the important point is the attempt to make explicit an error in one's answer, and an effort to say why it is wrong. I'd never delete such a post (as you describe). Alternatively, (and I would hope to see) in addition, "so the better step here" (the more accurate...etc, ) would be recognizing "blah .... blah" even if merely by identifying another answer (if there is one) exhibiting an approach that avoids the error. $\endgroup$ – Namaste Nov 25 '17 at 20:42
  • 1
    $\begingroup$ But without making exceedingly explicit one's wrong answer, and (through a well labeled header "edit"), adding what the correct answer looks like, then I don't think an uncorrected wrong answer should remain. (I've encountered a handful of users who, even when challenged, even when proved their answer is wrong, stubbornly resist deleting it, or correcting it, or even acknowledging their error. $\endgroup$ – Namaste Nov 25 '17 at 20:59
  • $\begingroup$ And, one final thought, @GEdgar. That approach (consistent with my earlier comments) may be acceptable or possibly helpful in some cases of posting a wrong answer, but I don't think it's a blanket solution in all such cases. Focusing too much on "this is wrong": "blah blah ... blah", actually has the drawback of unwittingly reinforcing the error, since some users recall salient approaches better than they remember which are wrong, and which are right. $\endgroup$ – Namaste Nov 26 '17 at 21:05
  • $\begingroup$ @amWhy I'll add that your last comment is rather close one of the points made in Phira's answer in the previous discussion: Is sometimes keeping wrong answer reasonable? $\endgroup$ – Martin Sleziak Nov 27 '17 at 5:46
  • 3
    $\begingroup$ In this situation, I will always ask myself: If I knew it is wrong before posting the answer, will I still post it? If the answer is no I will delete the answer. $\endgroup$ – user99914 Nov 27 '17 at 18:36
3
$\begingroup$

It is obvious from my reputation (as for the time I'm writing this) that there is still time before I'll be able to vote for deleting, but I still want to say my opinion because I do think it is an important matter.

I think that the function of voting for deletion should be used only on very unique cases:

1- the answer is not related at all(less relevant to this discussion).

2- the answer(wrong or not) is using methods that was pointed out by the OP of the question as something he/she is not looking for. I think that the most common case is when asking a question in limits without L'hospital's​ rule. In my opinion those kind of answer should be deleted because they serve no purpose for the question.

3- the answer is wrong and the mistake doesn't help. For example if someone misunderstood the question he is answering.

Apart from those cases I think that mistakes can be very important and should never be deleted. But I'm not saying they should be leaved.

Each mistake is from one of 4 main types of mistakes:

1- answering using blunt wrong information, for example using a lemma that wasn't proven.

In this case this question should be downvote and commented about the mistake and giving the OP to decide what to do(delete/edit)

2- misusing something, for example using a theorem before all of the conditions to use it are indeed there.

In this case I think that downvotes are unnecessary, you should add a comment pointing it out or edit it to be better. Because even if this, say, theorem can't be used at all in that question pointing out the mistake will make people recognize the mistake and will help them in the future. If such a mistake will be downvoted I think a lot of people that can learn from that will just ignore it.

3- some algebra mistake

This should be just edit but whomever spotting it

4- a "mistake" in how can we use something. For example if I am saying that $\lim\limits_{x\to0}\frac{\sin x}x=1$ because $\sin\theta=\theta$ for small angles.

This case is the most complicated, I think that this case should be judged by each person. If you are think this is a serious mistake downvote and comment it, If you think it is a small detail just comment

I think that none of those mistakes should be deleted because people can learn from those things a lot.

The last thing I want to talk about I'd about answers that one simply don't understand, in those cases one should ask for explanation or just ignore it.

$\endgroup$
  • 9
    $\begingroup$ I will without a doubt vote to delete an answer to $\lim_{x\to 0} \frac{\sin x}{x}=1$ using $\sin x=x$. $\endgroup$ – user99914 Nov 27 '17 at 19:09
  • 1
    $\begingroup$ @JohnMa well, small angles is actually something that I learned in physics and in some area we did use it(again small angles, not all angles). Ways to calculate things related to waves in microscopic size. But the main point stand either way! Not all mistakes are "wrong", some are just not something you accept. $\endgroup$ – Holo Nov 27 '17 at 19:14
  • $\begingroup$ @JohnMa look at this en.m.wikipedia.org/wiki/Small-angle_approximation $\endgroup$ – Holo Nov 27 '17 at 19:20
  • $\begingroup$ Well, it is always $\sin x \sim x$, which I have no objection. $\endgroup$ – user99914 Nov 27 '17 at 19:22
  • $\begingroup$ @JohnMa this is exactly the point, for some people using this to prove that $\frac{\sin x}x\approx 1$ is enough to implies $\frac{\sin x}x=1$(when taking the Lim to 0). You think it is wrong so you can comment about it but I think delete it because of this is something you must not do. Again, my opinion $\endgroup$ – Holo Nov 27 '17 at 19:26
  • 1
    $\begingroup$ So $2x\sim x$ implies $2x=x$? To me the use of $\sin x=x$ falls into the category of "it is wrong and the mistake doesn't help". $\endgroup$ – user99914 Nov 27 '17 at 19:29
  • $\begingroup$ @JohnMa first let me make it clear, in small angles(again, only small angles) we have $\sin\theta\approx\theta$ and not $\sin\theta\sim\theta$, but I won't enter to why now. Secondly, you are right $a\sim b\require{cancel}\cancel{\implies}a=b$ but for $\approx$ we can use $=$ sometimes $\endgroup$ – Holo Nov 27 '17 at 19:54
  • 4
    $\begingroup$ Using $\sin{x} = x$ to prove this limit is circular. I would consider this a major mistake, and would definitely support it being deleted; seeing this kind of thing is misleading. I don't think a comment underneath helps avoid this confusion. $\endgroup$ – Morgan Rodgers Nov 28 '17 at 18:53

You must log in to answer this question.

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