[Edit: looks like the question is now undeleted; the principle of what should be okay remains]

Those of you without the privilege to see deleted questions won't get anything out of the link, but here is a question that was posted 11 days ago.

Since it had sat there unanswered for a week and a half without any response, I gave a fairly lengthy partial answer yesterday-ish (depending on your time zone it may be 2 days ago), with some explanations of where further clarification would be required for a more complete answer (which explanations were much too long for a comment -- or even two comments)

The OP's response was to delete the question, which of course wasted the effort I had put into responding to the question.

I find something like this happens ... well not especially often, but it's not all that rare either, and it seems to be happening more frequently.

At least partly as a result of events like this, I find I'm less likely to actually post an answer a question from a low-reputation user than I would have been say a year ago.

My answering rate is down to below half what it was say last January -- and of course there are a number of reasons why that would happen -- but this perception of the increasing risk of wasting my time is certainly a contributing factor.

I have on one previous occasion undeleted a post where the OP deleted pretty much as soon as I answered, but that (undeleting it myself) is not something I'd seek to do as a matter of course in relation to posts I'd put an answer on (I'd do it with someone else's answer though, as long as there was some value in the answer, since there's no danger of it being self-serving to do so then).

It looks like OPs can only delete a question with an answer if there's only one answer and the answer isn't upvoted ... but that simply seems to be encouraging OPs to delete as quickly as possible if they decide their question is no longer worth investing time to fix (rather than actually put in a little effort to fix it with an edit).

What's a reasonable response to this sort of thing?

Should I just accept that I should exercise more caution and only answer questions I don't think the OP is likely to delete? If I'm not the only one that feels that way, our answer rate will presumably drop even more quickly than it has been.

Or should I post more low-effort answers, only investing more effort where the OP doesn't see fit to abandon ship at the first sign of difficulty?

Or should I seek some assurance the OP won't delete as soon as I post a substantial answer?

I really don't know what to do, but I'm getting quite cheesed off at the cavalier way answers can be just trashed because the OP feels like it.

[Edit 23Jan: I realize in retrospect the correct response to the question by the StackExchange model would have been simply to comment pointing out the biggest issues with the question and simply put the question on hold until it was improved, or eventually, to delete it if it was not e.g. see the discussion here as it relates to borderline questions. To quote from the link "rather a few borderline posts be deleted than the site be overwhelmed with garbage". So I probably did the wrong thing in trying to partially answer it at all.]

  • 4
    $\begingroup$ If the only reason for not undeleting the question is because having answered it you can no longer be disinterested, why not ask another moderator to take a look at it? (Though I wonder if the consequences of a lack of impartiality in these situations could be harmful enough to be worth bothering about.) $\endgroup$ – Scortchi - Reinstate Monica Jan 15 '16 at 11:55
  • 2
    $\begingroup$ I don't disagree with other answers, but it deserves brief emphasis that the ability of OPs to delete their own questions is more commonly a feature. They may realise that the question is off-topic, already answered or unclear, or that we can't play the invited game of telling them what's the correct analysis or the right way of approaching their project (or even of setting it up). That's consistent with the point of view that regardless of the merits of the question, a good answer based on a lot of effort deserves a prompt upvote, not just on its merits but to protect it against deletion. $\endgroup$ – Nick Cox Jan 17 '16 at 14:21
  • 1
    $\begingroup$ I naturally can't comment on precisely why this thread was deleted and in any case this Meta thread addresses the general principle. Sometimes it appears to be lack of awareness of the goals of the forum, namely that a good thread will be of use or interest to others in the future. Sometimes it's an attempt to remove a footprint from the internet: in some fields, people are paranoid that web searches by employers will reveal a history of asking dumb questions, even if identifiers don't reveal exact names. $\endgroup$ – Nick Cox Jan 17 '16 at 14:28
  • 1
    $\begingroup$ The first "field" that came to my mind when starting to read the post are students - if they ask self-study questions and copy & paste the answers they obtain here, plagiarism detection software (or an inquisitive teacher) may find that the answer is not actually the student's with a simple web search. $\endgroup$ – Christoph Hanck Jan 21 '16 at 13:47
  • 1
    $\begingroup$ Sometimes it's best, I think, to address a couple of issues in comments & only bother with more if the OP responds. On the other hand it's easy that way to end up partially answering the question in an inordinately long comment thread (e.g. stats.stackexchange.com/q/189385/17230 (& I still owe an answer!)). $\endgroup$ – Scortchi - Reinstate Monica Jan 26 '16 at 17:24
  • $\begingroup$ (+1) On stats.stackexchange.com/questions/189183/…, both question and @Glen_b's answer. Also, put in my own short answer. The question is 1) common 2) usually unexplained to newbies and 3) worthwhile keeping. Here is one example of a closed question that I was downvoted on that I understood the OP question and was accepted by the OP with others nipping us for no good reason. stats.stackexchange.com/questions/338802/… $\endgroup$ – Carl Apr 23 '18 at 19:10

A useful principle governing these situations is

Respect what the community contributes.

This clearly implies that whenever someone has made a substantial reply to a question, the thread should remain visible to all. If the question or any answer has an upvote, preserving visibility becomes even more imperative.

As a moderator, whenever I have encountered a unilateral deletion of such a thread I have reversed that deletion and left a comment to the O.P. explaining why. I can't remember any situation where there was significant objection.

This principle applies, albeit more weakly, to threads in which there appear to be important contributions in the comments.

Other courses of action are possible. For instance, one could unilaterally repost the deleted question and self-answer it. Although I can't immediately come up with a strong objection to this maneuver, I don't like it for its artificiality.

As with any important principle, there will be notable exceptions. One that arises about once a year occurs when the O.P. discovers they have posted sensitive material. In those cases we delete the question and also contact the SE team to remove the edit history.


It seems to me the overarching raison d'etre for CV is to create a body of high quality information about statistics, machine learning, etc., and only secondarily to provide short term assistance to individuals (although that is clearly worthwhile as well). Working from that principle, I see several conclusions:

  1. If there is valuable information in the answer (even though the OP will most likely never provide the requisite additional information), then the thread should be undeleted.
  2. If allowing such threads to remain deleted will discourage people from providing high-quality content in the future, the thread should be undeleted.
  3. If neither condition is met, the thread can remain deleted.
  4. If the conditions are met, but the OP objects, I believe moderators developers have the ability to remove the connection between a post and its original author so that the post is subsequently shown as being owned / authored by the community user. If that can be arranged, it seems like a viable compromise.

Another useful principle is

Respect what the original poster asked

SE's own 2015 statistics highlight the surprising fact that all questions posted on this (entire) domain are down voted once. Given that not all SE forums are equally harsh, some must be worse than others to maintain the average.

Some of these down voted questions are asked by non-stupid people who have taken the time to register and ask what they feel are legitimate questions in areas they are not expert in.

Others ask simplistic questions (relative to expert standards) that then receive post doctoral level answers that are of absolutely no use to the originator. If a child asks why is snow white, don't hand over a print out from a photo reflectivity analysis and a slideshow of electron energy jumps.

If the true purpose of SE is not to provide assistance to individuals, it might be worth while changing the access policy to not allow all and sundry to post questions.

I can well understand posters deleting questions that receive useless answers. I have done the same, and have to convince myself to ask further ones. I challenge you (as statisticians) to determine how many posters only post one or two questions, are patronised, and are never seen again. Your example at the top is a case in point.

I finish with something I'd like all to dwell on: "Acceptable to whom?"

  • $\begingroup$ Hi Paul, thanks for your comments. However, it's not quite clear to me how your advice would work in the context of the question we're discussing. Could contribute an answer there that follows the principles you'd like to have seen followed, by way of example? There's little doubt I have something to learn, but I don't see exactly how to approach it in this instance. I'm sure such an example could be helpful to a number of us (as well as to the asker). $\endgroup$ – Glen_b Jan 21 '16 at 23:55
  • 1
    $\begingroup$ "all questions posted on this (entire) domain are down voted once": that is not true of CV. Are you alluding to an average of some kind? Can you show or cite results from a website or reproducible analysis we can run? As for "useless answers", "patronised", "harsh": we should discuss such criticisms if true, but these are strong claims. What fraction of posts do you think fall under one or more of these headings? If you see this specifically, do post better comments or more helpful answers. The implication that this is general needs more substantiation. $\endgroup$ – Nick Cox Jan 22 '16 at 1:58
  • $\begingroup$ My comment doesn't relate to the issue of deleted questions, but I just wanted to say that I agree with Paul's comment about some answers being too long and detailed. I sometimes see answers (often with the highest number of votes) that I think would be more helpful if they were a quarter of the length that they are. I know a lot of people will disagree with me, but I value conciseness. For that reason, I really dislike the message "We're looking for long answers...", as though an answer has to be long to be good. $\endgroup$ – mark999 Jan 22 '16 at 2:07
  • $\begingroup$ I should also mention that I'm not suggesting that long answers are necessarily bad. I've seen long answers here that I thought were excellent. $\endgroup$ – mark999 Jan 22 '16 at 2:08
  • 2
    $\begingroup$ @mark999 In such cases, why not post a shorter answer yourself? or a summary comment that you think contains the nub of the matter? or a comment that said "Helpful, but please add a short summary"? I know that if I posted an answer that was longer than my average and someone chipped in with a pithier summary, I would want to regard that as constructive. $\endgroup$ – Nick Cox Jan 22 '16 at 2:31
  • $\begingroup$ Thanks for your comment @NickCox. Posting a shorter answer might be a reasonable thing to do, but it would be essentially posting a subset of an existing answer as a new answer, which I wouldn't want to do. Also, for the majority of questions here, I don't feel like I know enough to post a good answer. $\endgroup$ – mark999 Jan 22 '16 at 2:52
  • 3
    $\begingroup$ Fair points both. However, but what then should be done if you are right? A long answer can always be skimmed and skipped. An answer that is short can be more useless and I see that as a more common problem. I always try to be mindful of the other people who might be interested by an answer. I am not happy at the thought of people leaving the site after a brief visit who are frustrated or disappointed at the answers (or lack of them) that they got, but in many such cases the questions just were not clear or the posters stopped trying. That's frustrating and disappointing for us too. $\endgroup$ – Nick Cox Jan 22 '16 at 3:11
  • $\begingroup$ @NickCox I agree with everything you wrote. I don't think it's a matter of anyone being "right"; I have my preference and other people have different preferences. It's true that long answers can be skimmed and skipped, but I've seen answers where the preamble and extra details have made it relatively hard to find the actual answer to what was asked. Since writing my intial comment, I saw this: meta.stackexchange.com/questions/268421/… I agree with Tomas but clearly many people don't. $\endgroup$ – mark999 Jan 22 '16 at 4:25
  • 1
    $\begingroup$ @mark999 I take the point. In practice, that message is sent here quite frequently but only if the answer is undeveloped and incomplete. It's not the only principle there is. I go along with those who think that the spirit is right, but the initial wording could be improved: "long" could be just "high quality", followed by hints on what that means. Moderators and high-reputation users are often positive about high-quality answers that are concise and complete, say one paragraph long, from anyone. $\endgroup$ – Nick Cox Jan 22 '16 at 9:53
  • $\begingroup$ Thanks for your reply @NickCox $\endgroup$ – mark999 Jan 22 '16 at 10:15
  • 2
    $\begingroup$ I'm sorry you've been frustrated here, Paul. It might help if you open a new meta.CV thread to discuss the problems you mention, & especially if you could give some concrete examples. For instance, you claim that all questions on CV have been downvoted. My impression is that only a small percentage are; in particular, I notice that none of your 5 questions have been downvoted. I could go on, but my point is that your perception of the site is quite foreign to me, but even if only partially true, it is something that we need to address. $\endgroup$ – gung - Reinstate Monica Jan 22 '16 at 17:05
  • $\begingroup$ I think Paul's comment on downvotes per question was based on figures that are network-wide, not CV only -- likely based on the numbers for stackoverflow (2.5 M questions, 3.2 M answers, 2.3 M downvotes) and for the rest of stackexchange other than stackoverflow (almost 1M Q's, 1.3M A's, 1M downvotes) mentioned here. There's perhaps a mistake in thinking that questions that are taking all or nearly all of those downvotes, but network wide there's still a high rate of downvotes per post. $\endgroup$ – Glen_b Jan 23 '16 at 2:35
  • 2
    $\begingroup$ However, from looking at data.stackexchange.com/stats , here on CV last year we averaged about 2200 questions per month (we have a similar rate of answers but I haven't looked those up) and 70 downvotes per month (across questions and answers). (We upvote about 30 times as often as we downvote.) ... a lot of those upvotes come from familiar names (If we we having this conversation late last month I could point out the top voters for 2015; I know who they were but it would be nice to be able to point straight to the figures) $\endgroup$ – Glen_b Jan 23 '16 at 2:41
  • $\begingroup$ For where I got those numbers form, see votes per month and questions per month. Note that the dates on the questions per month are not "start of the month in question" $\endgroup$ – Glen_b Jan 23 '16 at 2:42
  • 3
    $\begingroup$ Paul, as my figures above clearly show, on Cross Validated our downvote rates are very low -- averaged across both questions and answers less than 2% are downvoted. Even if you assumed every downvote was applied to a question it would only be about 3%. Instead of lots of downvoting and closing, we lean towards requests for clarification and edits (and a large fraction of those requests for clarification and those helpful edits fall on very few shoulders). $\endgroup$ – Glen_b Jan 24 '16 at 23:20

To the OP, I looked at your long answer and some of it sounds rude. So it might have been painful to the questioner and maybe that's why he deleted his question and the thread. Here's a particular bit I thought was rude:

 Now this here is a technical measurement. (you quoted the questioner's
 irrelevant colloquialism)

Your terminology is unfamiliar. Can you either avoid the jargon or explain its meaning? (you made fun of the guy essentially)

I'm sure you didn't intend to make fun, but I second a word used by Paul Uszak, "patronizing."

  • 5
    $\begingroup$ I would definitely encourage readers to go to the thread in question and read my original (absent the interpolated comments above) and draw their own conclusions. I'd also encourage anyone who wishes to contribute a better answer in that thread to actually do so. I'd be keen to upvote (and learn from) any answer better than my own. The OP at the original question appeared to me be using the term "technical measurement" to mean a very specific thing (i.e. to be using it as a term of jargon with a particular meaning) but I had no idea what that might mean in the context he was using it $\endgroup$ – Glen_b Jan 23 '16 at 2:48
  • 2
    $\begingroup$ Specifically it seemed to me that understanding the distinction he was drawing between "technical measurement" and "natural quantities" seemed directly relevant to answering the question properly (hence why I asked). $\endgroup$ – Glen_b Jan 23 '16 at 2:55
  • 2
    $\begingroup$ Sorry, iko, but I disagree. If you call this discussion rude, then most of the answers on this forum also qualify. Someone is confused, unclear, wrong, not even knowing what language to use to discuss their situation: these are common difficulties on this forum and in the thread being discussed. Some times we can translate easily to statistical terms; often we can't. I see @Glen_b's style of answering as civil and intended to be helpful. I have no idea either quite what the original question meant here and if I had answered I would have had to ask similar questions. $\endgroup$ – Nick Cox Jan 24 '16 at 11:09
  • $\begingroup$ When someone is "confused, unclear, wrong, not even knowing what language to use to discuss their situation," I think this person is in an especially vulnerable position. It takes them courage to ask an expert and they're afraid their question makes them look stupid. When I try to help other students, I'm careful not to pick them apart and to sound as neutral as I can saying "X and Y is unclear" or "could you explain what you mean by X and Y for me?" $\endgroup$ – iko Jan 24 '16 at 12:36
  • 2
    $\begingroup$ We agree. But that is exactly what @Glen_b was doing. A lesson that vulnerable questioners need to learn quickly is to try not to be offended by tone. I've been told that British people (I'm one) often sound terse and blunt, whereas we are just trying to be concise and go straight to the key points and believe that waffle wastes time and helps nobody. Glen_b is not British, but he really isn't rude either. More generally, we're all going to draw the line on rudeness in different places. "That's a stupid thing to say" or anything ruder would not be tolerated here if anyone noticed it. $\endgroup$ – Nick Cox Jan 24 '16 at 13:44
  • $\begingroup$ At the general level, we agree. As I said, I don't agree that is exactly what this particular answer was doing. With the best intentions, the OP's answer sounds provoking and patronizing to me, because of how it asks for clarification. A good answer would ask for clarification in a neutral way and be brief and to the point. It's not my place to flag such an answer, but I can understand the questioner deleting his question after this answer. I think un-deleting it does not respect the questioner's privacy (right to withdraw own question from public place). $\endgroup$ – iko Jan 24 '16 at 15:05
  • 1
    $\begingroup$ A problem intrinsic to the internet is that all paralanguage (ie, tone of voice, facial expression, etc) is lost. Something said & meant in a respectful way may be interpreted differently. This isn't necessarily the fault of either person. Having observed Glen_b's behavior on this site over several years, I have a good sense of him. Like @NickCox, I did not interpret anything as rude or patronizing. Someone who doesn't know him & is feeling especially vulnerable might read that in to his words even though it wasn't there. $\endgroup$ – gung - Reinstate Monica Jan 24 '16 at 15:25
  • 1
    $\begingroup$ @gung So should the criterion for neutrality be: a person who doesn't know you either way (for good or bad) thinks that your writing sounds reasonably neutral? Or should it be instead, people who know you're a lovely person think that your writing sounds neutral? $\endgroup$ – iko Jan 24 '16 at 15:30
  • 1
    $\begingroup$ Not knowing you, @iko, "people who know you're a lovely person..." comes across to me as having an edge. I generally give people the benefit of the doubt. You make a perfectly reasonable point that some askers may be feeling vulnerable & take a neutral comment as derogatory. I often explicitly try (& I know Glen_b does, too), but I don't doubt I occasionally fail nonetheless. $\endgroup$ – gung - Reinstate Monica Jan 24 '16 at 15:51
  • 1
    $\begingroup$ The right to privacy is important, but a red herring here. People can attain privacy by never posting in the first place or by choosing an anonymous or cryptic identifier. Once you post at all you forfeit a quantum of privacy (meaning, just a little). Let's not forget @Glen_b's notional right not to have his answer deleted too, although clearly forum rules allow deletion of an answer which hasn't been upvoted, which was the original trigger for this thread. $\endgroup$ – Nick Cox Jan 24 '16 at 16:17
  • $\begingroup$ Why a red herring? Yes, some privacy was lost just from asking the question. But does that mean my question stays up in a public place, after I try to take it down? At the outset, the questioner didn't agree not to delete his question. If later he feels strongly about it, why not let him take it down? He's not necessarily being perverse, uncooperative or ungrateful. Maybe he feels ridiculed in a public place. I think it is this that triggers his right to privacy. $\endgroup$ – iko Jan 24 '16 at 16:45
  • 1
    $\begingroup$ Well, by asking a question you're contributing to a communal work; your proprietary rights to the foundation are the more attenuated the more others have built on it. Nevertheless there is a mechanism for you to dissociate yourself from the whole thing should you feel strongly about it, as discussed in @gung's answer. $\endgroup$ – Scortchi - Reinstate Monica Jan 24 '16 at 21:22
  • 1
    $\begingroup$ This makes sense: "your proprietary rights to the foundation are the more attenuated the more others have built on it." Is it sufficiently captured by the rule that upvoted answers are not deletable, since an upvote signals significant progress on the foundation? Conclusion 4 in @gung's answer also sounds fair, but I think the exact words or images in the original question don't need to be used after un-deletion, and can be summarized to protect OP's anonymity and choice to withdraw question. $\endgroup$ – iko Jan 24 '16 at 21:53
  • 1
    $\begingroup$ ... (ctd) the question is the extent to which they should be able to delete another person's content. Presently they can do it, but should they? (Conversely, I can undelete it... but should I?) $\endgroup$ – Glen_b Jan 25 '16 at 0:49
  • 1
    $\begingroup$ I doubt that Glen_b was deliberately being patronising, but I can easily see how his answer could be interpreted that way. In particular, "Well, no, it's not done", especially with the italics, appears to me to be talking down to the question-asker. But even if the question-asker was offended, I don't think that should be a justification for allowing a question to be deleted. My preference would be that a question can only be deleted if there are no answers and no comments. $\endgroup$ – mark999 Jan 25 '16 at 1:08

You must log in to answer this question.

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