Recently, I've seen a few examples of users asking questions and answering their own questions as a way of adding to the question pool here on MSE.

The most recent example I've seen of this is the following question, although it certainly is not the first.

In the past, I voted to close such questions. For example, user @LeAhnDung recently had posted a "question" on ordinal arithmetic, in the hopes of providing a general answer to all questions relating to the subject. While @AsafKaragila agreed with my evaluation that the question didn't fit MSE guidelines, it contributed to the confusion I hold today.

What is the appropriate way of handling these sorts of questions? Do we ignore them? Do we encourage them? Do we close them? Do they count as a form of reputation hunting? (In my experience, I haven't found the users who exhibit this behavior to be doing so for reputation, but it's plausible)

  • 23
  • 5
    @MarkMcClure I'd argue that this is somewhat different than answering your own question. This sort of question is designed to be self-answered from the onset. Moreover, I'm a bit concerned that this could be a slippery slope in the types of questions that get asked. – Rushabh Mehta Nov 5 at 21:01
  • 26
    Hmm... From the help document: "If you have a question that you already know the answer to, and you would like to document that knowledge in public so that others (including yourself) can find it later, it's perfectly okay to ask and answer your own question on a Stack Exchange site." Seems like that's exactly adding "to the general knowledge pool". – Mark McClure Nov 5 at 21:04
  • 2
    Mark McClure: we need, then, to be clear as to what "counts" as "adding to the general knowledge pool." Not every question, self-answer accomplishes that, you much admit. Furthermore, if the "addition" is a "duplicate" or another question/answers post, it merely duplicates. So it cannot be said that all questions+self-answers are explicitly encouraged in the help center. The issue here is to determine when it might be appropriate, and why and when it isn't always appropriate. – amWhy Nov 5 at 21:12
  • 2
    So, @RushabhMehta is there some way that the types of questions that you refer to differ from the types of questions as described in the help center? Or is this just a CRUDE quality thing? – Mark McClure Nov 5 at 21:16
  • @MarkMcClure You can address my comment, as well. Don't take the "Help Center" as authorizing every conceivable configuration of a question with self-answer. Case in point: user asks: "is 2 + 2 = 4?" with a self-answer: "yes, 2+2 = 4, in standard arithmetic on the set of all integers. But watch out!, $2+2 \equiv 1 \pmod 3$, so it isn't always the case that $2+2 =4$!" – amWhy Nov 5 at 21:23
  • 6
    @amWhy Yes, your comment seems quite fine - thank you. I guess the point is, if one wants to close what is (in one's opinion) a poor question, then do so. Whether or not the question was asked in order to provide a self-answer seems irrelevant. I specifically pinged Rushabh because I'd like clarification on whether there some specific distinction between the types of questions he's asking about vs those described in the help center. – Mark McClure Nov 5 at 21:30
  • 2
  • 21
    Just stop making judgement based on who is posting a question and who is answering it. The value of a question/answer pair doesn't really depend on who is the poster. – achille hui Nov 5 at 22:02
  • @MarkMcClure Honestly, I was just trying to start discussion. I didn't have such specifics in mind. – Rushabh Mehta Nov 5 at 22:16
  • 5
    @Mark McClure: the content of the help center that you referred to was not decided by this site, it was added unilaterally without asking us. The actual views of users on this site is more nuanced. – Carl Mummert Nov 6 at 1:20
  • 5
    Personally, I think there is a difference between having a genuine question - something you don't know but wish you did - compared to "playing Jeopardy" by phrasing something one already knows in the form of a question. In my opinion the best use of this site is for genuine questions. Users who just want to post their thoughts would usually be better served by using a blog. So we can expect these questions-in-form-only to be of the highest possible quality, at least, to warrant posting here. – Carl Mummert Nov 6 at 1:23
  • 1
    @CarlMummert second that, and wish it was posted as a solution! – rschwieb Nov 6 at 3:38
  • 1
    Let me also point out that the specific user you mention is posting proof verification questions, and I think that was his attempt to break out of the usual proof verification methods. The problem with proof verification was recently brought up by me on math.meta.stackexchange.com/questions/29119/…. – Asaf Karagila Nov 6 at 12:21
  • 3
    I have so far added a few questions with the purpose of adding to knowledge base. In each case I was prompted by some user via a comment for some explanation. I think an occasional behavior of this kind is fine and in any case if someone is worried about rep hunting, such posts can be made community wiki. – Paramanand Singh Nov 8 at 12:43
up vote 11 down vote accepted

In this context I want to bring up two possible uses of self-answered questions. My goal in posting this as an answer is to poll our user base, and learn whether people find this palatable.

I am aware that theses cases form a tiny fraction of the possibilities related to self-answering. For the purposes of this discussion the point is that may be we can try and use these to get an idea of the scope of uses, and also get a feeling of how well they fit the SE format (neither is clear to me).

  • Q&A pairs prepared for referrals. An example of such due to yours truly. My plan was to use that as an extension of the relevant tag wiki, and then refer to it in a tag wiki as well as whenever it comes up. Opportunities for this do occur, but may be not frequently enough to make it necessary to have a policy.
  • Joriki suggested that self-answering could also be used for modularizing answers. That old meta thread never got enough attentin (IMHO). May be the modularization idea could also be used as an extension of abstract duplicates.
  • 5
    I am always a bit wary of opening up too much for self-answered questions, since my impression is that those most likely to post such are not the ones with anything of substance to add (we already have plenty of random complicated integrals posted as challenges). – Tobias Kildetoft Nov 6 at 12:08
  • 3
    There is also proof verification style questions, where you ask a question, and post an answer as a proof. Whether or not it is a legitimate way to ask people for verifying your proof is debatable. – Asaf Karagila Nov 6 at 12:22
  • 3
    @Tobias Kildetoft I suppose I don't see the harm of having posts on MathSE that lack substance. Or any harm in asking and answering questions that are basic or uninteresting. Ideally, such questions will receive few upvotes, and sink to the bottom of the MathSE pile of questions, only findable by the OP, who evidently had a need for such a post, and to anyone who searches for exactly the right thing. I mean, it's not like we're running out of room in the question pile. :) – Mike Pierce Nov 6 at 17:18
  • 3
    @Mike: The question is what do you think this site should be. If you think this site should be a repository of stuff, then yes. The pile is ever larger, and everything is fine. If you want it to be a library of knowledge, then some modicum of moderation as to what enters or at least what should enter is probably wise. To wit, not everyone can post on arXiv, even if it's a free website, and if you want to overcome this, you need someone to look at your paper and say "this is fine". – Asaf Karagila Nov 6 at 20:40
  • 3
    @AsafKaragila why not both? Knowledge is knowledge, interesting or not, and can be thrown in the pile (just gotta manually moderate the stuff that's wrong or incomprehensible). Then, ideally, the votes separate the interesting from the uninteresting, and design the facade of our library – Mike Pierce Nov 6 at 21:03

Since my self-answered question is the most recent example that OP has seen of this, and since I've posted such self-answered questions quite a bit (see here and here and here and here), I figured I should post an answer to at least make my own motivation clear. :)

I'm currently a teaching assistant for a precalculus class that I've been a teaching assistant for a few terms now. Each term, the students have nearly the same troubles with the same exercises, and I end up typing up the same solutions to email to them, or having the same conversation in office hours each term. This seems rather silly. So this term I've been quite a bit more active in using MathSE as a tool to post Q&As that I know my students would benefit from seeing. I realize that most of my more recent posts do not contain interesting questions. They're just exercises, and usually pretty basic ones at that. But they aren't bad questions either. They are questions that people legitimately have, in particular my students, so they certainly are adding the pool of useful content on this site, and certainly aren't doing any harm by existing. So to actually address your question of how to handle such posts: these self-answered questions should be judged by exactly the standard as any other post. If the questions is useful/clear/interesting then upvote it, and if the question is unclear/useless/disrespectful then downvote it. If a answer is good and addresses the question well then upvote it, and if the answer is wrong or doesn't address the question then downvote it.

To address the rep-farming bit, I didn't post any of my self-answered questions to get reputation; it's really just to have some Q&As posted somewhere convenient for me and my students that others might appreciate too. And honestly I'm rather flabbergasted at the voting patterns that I'm seeing on my self-answered posts. When I post such a question, I expect that the question itself will receive no upvotes (they are usually pretty basic questions afterall), but that my answers would receive a couple upvotes since I am making an effort to be thorough and write good solid answers for my students. But then you get cases like this answer of mine receiving 8 upvotes that, really, isn't worth 8 upvotes (originally there was no MathJax in the title of this post and it might have landed on the HNQ as quite a few basic questions do; none of these basic questions belong on the HNQ). And also this basic question that go two upvotes, whereas my answer got nothing. This illustrates how bizarre voting trends on MathSE can be, but I mean, really, this doesn't actually matter.

  • 2
    Keep up the good work, they really helped me a lot! – Raptor Nov 7 at 4:14
  • 4
    I would tread very carefully here. IMHO the onus is on you to verify that the questions are nowhere near being duplicates. I do belong to the school that thinkg 95% of (pre)calculus questions are duplicates, but... my gut feeling is that I don't like this. – Jyrki Lahtonen Nov 7 at 4:27
  • 5
    @JyrkiLahtonen I'm concerned that any post that I find that is a duplicate of the question I want to type up for my students would not be as helpful for my students to read as my own write-up. Like, if you or I see a duplicate and can say "oh yeah, that's basically the same question," my precalc students may not have the mathematical maturity to see that they're basically the same question. Broadly speaking, I don't think linking to an abstract duplicate question is helpful for precalculus level questions. – Mike Pierce Nov 7 at 5:00
  • @JyrkiLahtonen To talk about something concrete though, suppose I want to type up a Q&A for my students, and find the "canonical" duplicate post for that same question on the site. In terms of site organization, what would you think to be the best course of action? Should I add my own answer to that "canonical" post (if I don't like any of the current answers) and send that link to my students? Should I still type up my own and then mark mine as a duplicate of the canonical one? If I think I can ask and answer better than the canonical one, then should I type up my own ... – Mike Pierce Nov 7 at 5:04
  • ... and mark the canonical one as a duplicate of mine? :P Should I just write mine and then link to all the questions I find that are basically the same? For an actual example post, see this self-answered question of mine and all the versions of basically the same question I linked to in the sidebar, this top linked post being the obvious "canonical" version. How should all these questions be organized? – Mike Pierce Nov 7 at 5:09
  • 4
    Use community wiki for answers, if you truly don't care about gaining rep from both asking, and answering, and it is sincerely a mode you use for pedagogical purposes. – amWhy Nov 7 at 16:18
  • 3
    @amWhy That's not what Community Wiki is for though? Like, the questions I've posted are not general enough be wiki posts for users to expand on: they are specific questions. I'd considered marking my answers CW before, but that seemed inappropriate. They are my answers afterall, and not a general responses to be expanded and refined by the community. Also I think these questions would benefit most from the community adding distinct answers from a different perspective rather than expanding on mine, and marking my answer CW would suggest the latter. – Mike Pierce Nov 7 at 17:04
  • 4
    Community wiki is used for a lot of reasons when answering questions. One of those reasons is to make clear one is seeking only to add to the "general knowledge pool" (period). That's how I've occasionally used it, though I haven't asked any question only to turn around to answer it. I'm sympathetic with Carl Mummert's position on this. I prefer sincere questions, and reading the answers in response to the asker. I'm not so much interested in reading contrived questions where the purpose in asking them cannot be separated from the motivation "to ask in order to answer." – amWhy Nov 7 at 17:15
  • 3
    @amWhy Oh, yeah, I'm more aligned with the whole "StackExchange is a library of Q&As, and a user getting help with a question they earnestly have is just incidental ..." viewpoint that is expressed somewhere in the original general SE documentation or mission statement or wherever. Like, the purpose of every post is just to add to the general knowledge pool. So I think this is where we run into our disagreement over the best-use/purpose/philosophy of MathSE (again :P). – Mike Pierce Nov 7 at 18:35
  • 1
    Mike, we used to have too localized as a close reason. And here I argued that a series of posts designed to help students of a single course at a single university/college surely fits that description. We know longer have that close reason (not sure why, searching the meta would probably help there). Anyway, yours truly (and possibly a few others) don't like this style of posts out of that inertia. But, if we polled opinions of users, I doubt we would get a very clear majority leaning either way :-) – Jyrki Lahtonen Nov 8 at 8:20
  • 3
    Anyway, those of us who think that SE is a Wiki first, including the founder of all of this, may instinctively disprove of using the site that way. – Jyrki Lahtonen Nov 8 at 8:25
  • Related to "judged by the standard of any other post" - part of the usual standard is that the OP does not know the answer to the question, which is why they are posting it. I don't interpret "question" as "ends with a question mark" - I interpret it as "something someone does not know, but wishes they did know". The purpose of this site, IMO, is for people who have questions like that to post them and find answers. If someone makes a post for which they do not have an actual question, they have already distinguished the post from other posts. – Carl Mummert Nov 10 at 13:31
  • The other thing that many of your posts lack is the source and motivation of the problem. Where did the exact problem arise, and why are we interested in that exact problem? As described in How to ask a good question, this is info that some of us would like to see in all "here is a problem that looks like it came from a textbook" posts. Posts such as this one have no motivation for why someone wants to solve the problem, or its source. – Carl Mummert Nov 10 at 13:36
  • 1
    @CarlMummert: "part of the usual standard is that the OP does not know the answer to the question" - I have never heard that standard articulated anywhere before, and it seems like if that were part of the standard, self-answered questions would serve no purpose (but the tour and the user interface both actively encourage people to post them, so clearly this is an MSE-specific thing, not SE-wide). Can you please link to the meta policy where it was decided? – Kevin Nov 10 at 22:00

I have self-answered a few questions. At that time, I was reading a textbook thoroughly and I wanted to document answers to some questions that arose in my mind while I was reading the material. I felt that they were reasonable questions in the sense that any other student going through the same material would very likely have the same confusions that I had.

Here are the questions that I asked and answered while reading Hoffman and Kunze's Linear Algebra:

  1. Proof of Theorem 7 (Chapter 5) in Hoffman and Kunze's *Linear Algebra* is unclear
  2. Doubt in section 5.7 of Hoffman and Kunze's *Linear Algebra*

Here are the questions that I asked and answered while reading Patrick Morandi's Field and Galois Theory:

  1. Doubt in proof of Theorem 3.19 from Patrick Morandi's *Field and Galois Theory*
  2. Doubt in proof of Proposition 4.5 from Patrick Morandi's *Field and Galois Theory*

I had also asked this question on meta whether it would be appropriate to self-answer questions for the purpose of documenting clarifications to Morandi's textbook. The consensus was that if the proper context is provided then it should be okay. Indeed, some of my questions are not of the highest quality (clearly the titles are cringe-worthy), but they were not received negatively by the community because, I believe, the necessary context and motivation was provided.

My point is that even though I have self-answered questions before to add to the general knowledge pool, I cannot say with confidence when a self-answered question accomplishes this and when it fails to. However, I can definitely say that this comment was very useful to me in deciding which sort of questions to leave out:

Just please make sure you ask a question in each and every post, and don't turn your ongoing posts into a blog. I.e., don't over do it. But the Question Answer format is a must. – amWhy Jun 28 at 23:29

To me, this "blog test" is the most useful method to decide which questions shouldn't be going on MSE.

Again, I'm not claiming that all of my above questions are a good fit for MSE either, but I know that without applying this discriminatory tool I might have posted many more lower quality questions for the purpose of "adding to the general knowledge pool".

You must log in to answer this question.

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