19
$\begingroup$

I came across this question and answer today: Recall and AUC of a binary classifier

And I think the accepted answer is wrong. So I wrote what I believe is a better answer, and downvoted the accepted answer explaining why I think it is wrong.

side note: if my reasoning went astray please correct me

side note 2: I actually think in terms of content it will be easy to fix the accepted answer and I also left a comment that I'll gladly upvote a revised version. I don't think it will be a matter where the author of the accepted answer and I cannot reach convergence - I think that he just didn't think of a particular situation which makes the "can never happen" answer incorrect.
It is not like trench wars about the correct prior or appropriateness of Bayesian vs. frequentist approaches

The concept of stackoverflow relies on the fact that other users can up- or downvote answers. Ultimately this should lead to good (including correct) answers floating on top. This concept also assumes that in case some issue discovered, all sides can react: the author of the answer can edit and the OP can even change the accepted answer. This relies on people staying active on CV/SX, but what if they aren't active any more?

Question and accepted answer are 1 1/2 years old. The OP's profile says that the OP never came back to the site after the day that question was asked and "answered".

(The author of the accepted answer was online here more recently (2 weeks ago) so he/she may edit the answer in reaction to my comment.)

I'll flag the question for moderator attention, but I think it would be good to write down here how to proceed

  • in case the OP doesn't show up again and react on the new situation.
  • in future if similar issues arise (the OP could have deleted his/her account meanwhile)

To be even more clear: I think it is a perfectly valid policy to say that we don't do anyting in these situations.

I just think that it would be good to have the policy discussed and spelled out here (whatever it is).

| |
$\endgroup$
  • 12
    $\begingroup$ Please do not flag answers you think are incorrect: moderators are not arbiters of correctness or accuracy and cannot respond to such flags. Some related threads worth consulting include meta.stats.stackexchange.com/questions/647, meta.stats.stackexchange.com/questions/67, and meta.stats.stackexchange.com/questions/1039. $\endgroup$ – whuber Feb 17 '14 at 15:41
  • 6
    $\begingroup$ This subject also has been extensively debated in several comment threads on our main site at stats.stackexchange.com/questions/52319. I do not want to re-open that particular debate, but feel obliged to point it out because it shows that arguments over accuracy can become acrimonious and therefore need gentle treatment. In particular, we should always be aware that an answer may look incorrect because it implicitly makes different assumptions or adopts a different interpretation of the question than we expected. Consider probing that possibility with polite, directed comments. $\endgroup$ – whuber Feb 17 '14 at 16:17
  • $\begingroup$ @whuber: I edited the question: I'm not primarily asking about what to do with a "normal" wrong answer (possibly recent, but importantly: both OP and answer author are active so they react - and if it is by downvoting my answer and insisting that the accepted answer is correct). I'm asking here because I don't know what should be done, but I think it would be good to have a policy for such situations. $\endgroup$ – cbeleites unhappy with SX Feb 17 '14 at 20:02
  • $\begingroup$ I promise I won't flag again in such a situation. I flagged because I (wrongly) thought that moderators may have some whatever tool for such a situation (there are these banners saying that the question is closed or whatever - there might have been a banner saying that there are concerns/controversy but unfortunately the concerned people are not active here any more). I don't think that deleting or closing would be appropriate for the question. And anyways not immediately. $\endgroup$ – cbeleites unhappy with SX Feb 17 '14 at 20:04
  • 9
    $\begingroup$ There is no special policy either possible or needed here. Every action is fallible, but within limits correctable! Original posters may accept poor or even incorrect answers (and then fade away). People may post poor or even incorrect answers (ditto). But people who are competent and confident can and should post good answers, explain firmly but politely the limitations or errors of other answers, and vote up (and down) accordingly. So, you did what you could, and that's fine. $\endgroup$ – Nick Cox Feb 18 '14 at 0:06
  • 1
    $\begingroup$ @NickCox: may I ask you to convert your comment into an answer? $\endgroup$ – cbeleites unhappy with SX Feb 18 '14 at 13:01
  • $\begingroup$ I certainly will if you wish, but I regard it only as a statement of the well known. $\endgroup$ – Nick Cox Feb 18 '14 at 13:04
  • 2
    $\begingroup$ @Glen_b: may I also ask you to make his ans answer? As to (iii) editing: I'm extremely recluctant to edit an answer because I think it wrong (after all I may be wrong myself). Personally, I make only edits where I'm certain that I know what was meant. Which is fundamentally not the case if I think the answer wrong, not even if I have a guess what happened. $\endgroup$ – cbeleites unhappy with SX Feb 18 '14 at 13:04
  • $\begingroup$ @NickCox: I ask about writing down what how we think we should act in such situations. Such a "policy" should ideally align well with what is already well known - otherwise it is impractical. So yes, I think it answers the question. $\endgroup$ – cbeleites unhappy with SX Feb 18 '14 at 13:07
25
$\begingroup$

There is no special policy either possible or needed here. Every action is fallible, but within limits correctable!

  • Original posters may accept poor or even incorrect answers (and then fade away). It is best not to read too much into question acceptance; always remember that this is just one person's judgment. Sometimes an OP chooses what they wished to hear, or what they understood, or what best fitted their specific needs. Often the pattern of voting indicates other answers that are deeper, fuller, or more interesting to many readers.

  • People may post poor or even incorrect answers (and then fade away).

But people who are competent and confident can and should

  • Post good answers. If you have a better answer, post it!

  • Explain firmly but politely the limitations or errors of other answers.

  • Vote up (and down) accordingly.

In this specific case, you did what you could, and that's fine.

Note on editing: The tacit policy seems to be that people with sufficient reputation and a reasonably good eye for language, mathematical and statistical formatting should edit mainly to improve presentation. Supposed technical errors other than trivial typos should be brought to the attention of posters, with the aim that ideally they make corrections themselves. Such diffidence is partly for posters' own education (assuming they accept corrections) and partly because editors may be those who are confused or incorrect.

| |
$\endgroup$
  • 5
    $\begingroup$ +1 for both the first half (on what to do), & the second half (on editing policy). $\endgroup$ – gung - Reinstate Monica Feb 18 '14 at 14:12
  • 4
    $\begingroup$ I would add one more thing: comment on the accepted answer, saying why it is wrong. First, it gives an opportunity for the author to clarify, fix or delete the answer. Second, it gives a warning note for others (and your comment will be upvoted or not - so it does not require special actions). Quite a few times, on StackOverflow, I saw accepted answer with a heavily upvoted comment, that it does not work (or should not be used). Then I naturally head to the next answers. $\endgroup$ – Piotr Migdal Mar 1 '14 at 14:57
  • $\begingroup$ that's a very helpful post. Re the note on editing, I've noticed that this varies across sites. For example on mathematics, edits to make corrections seems to be more likely to be accepted than they are here or on stackoverflow. So it's worth being aware of the site policy if you use more than one site. $\endgroup$ – TooTone Mar 2 '14 at 12:55
  • $\begingroup$ oh, and is it worth adding your note on "mainly to improve presentation" that editing tags is included in that? (I was going to edit that into your post but thought better of it:) $\endgroup$ – TooTone Mar 2 '14 at 12:56
  • $\begingroup$ @TooTone Thanks for your comments on editing. The need for formal "acceptance" depends on reputation, but I've noticed few objections from OPs to edits that are intended as improvements in presentation. Insertion or deletion of tags is clearly often appropriate. I had not intended this answer to be a complete guide to how and what to edit. $\endgroup$ – Nick Cox Mar 2 '14 at 13:09
6
$\begingroup$

(I just realized I forgot to make my comment an answer, as requested)

There are a number of possible actions, including
(i) commenting on the answer (to point out where it went wrong) and
(ii) posting a better one
(and - though it's seldom used* - (iii) to correct easily fixed errors via edit)

* and that's mostly a good thing

If you post a good answer, there's some chance the people who know the area you're posting about will recognize the better answer and upvote your answer.

The correct thing for the moderators to do is to respond to any of those actions you can take as any other reasonably high-reputation user on the site might do in response - in their capacity as users - which is to comment, upvote, revert a mistaken edit, etc as appropriate.

As long as you do generally check things you're not completely certain about, you should not fear to post or comment. Yes, you may be wrong - I don't think anyone here is immune from that** - but if you're right most of the time, and you are reasonably diligent, everything will work much as it should.

If you're going to edit other people's answers to correct errors, you need to be much more careful - but even then there's little to fear, because your edits will be reviewed (and which edits might be rejected if they're too extensive, even if they're right). Even if edits go through, they can be reverted.

** (certainly not me, though I could point to a couple of users who are almost never wrong)

| |
$\endgroup$
4
$\begingroup$

Ask what to do on Meta, linking to the question under discussion; it will attract attention & more up-votes for your correct answer, making it clear to future readers that the accepted answer is flawed.

[Sorry for being flippant. But considering the case where an incorrect answer has been accepted with an up-vote or two, & a correct answer has zero votes or a down-vote, would it be an appropriate use of Meta to ask for more eyes on the question?—after of course following @Nick's advice.]

| |
$\endgroup$
  • 4
    $\begingroup$ not sure to agree here. I think this question is ok, as it seeks for guidelines about what we should do. Once we create the guidelines, wouldn't be the other questions duplicates of this one? $\endgroup$ – Andre Silva Feb 18 '14 at 13:11
  • $\begingroup$ @Andre: Not sure I agree, but I remember seeing similar questions before, with the same consequences, & a desirable outcome seems to have been achieved without recourse to moderatorial action or policy changes. $\endgroup$ – Scortchi - Reinstate Monica Feb 18 '14 at 14:21
  • 5
    $\begingroup$ I think this does touch on a key point, how to draw others' attention to a mess. Substantial change to a post will bring it to the top of the active list; some people will notice comments in "Chat"; otherwise raising it in Meta, as suggested here, is a possibility. Note that moderators rightly emphasize that it is beyond their remit to decide on correctness of answers. At the same time, if every contested question were echoed by discussion in Meta, that wouldn't work well, but as we are a long way from that, I am not so worried. $\endgroup$ – Nick Cox Feb 18 '14 at 16:34
  • $\begingroup$ @Scortchi: by the way I was not aiming at changing any policy but rather on having it written down so a search on meta yields the (already implicitly) existing policy. Kind of: make the policy searchable / findable. $\endgroup$ – cbeleites unhappy with SX Feb 22 '14 at 12:56

You must log in to answer this question.

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