I often ask questions that perhaps get one up-vote, then someone makes an answer that excellently adresses exactly my issue. That answer then gets five up-votes.

If everything was about scoring points, then that would perhaps be entirely appropriate; after all, a person who comes up with a good solution to a problem deserves more credit than the one who could not solve it. My concern is rather of the question gets the attention that it actually deserves, if the question will ever be noticed by other people who may also find it relevant; and whether that happens has a lot to do with the score it gets.

My logic is this: Suppose people found the answer relevant, and that the answer does exactly what I asked for and nothing else (and does not provide any additional perspectives, like explaining how the TeX kernel functions). Then it must be because they found the question relevant, at least to some extent. So why not up-vote the question, too?

  • 9
    In mathematics the art of proposing a question must be held of higher value than solving it. Commented Jan 21, 2015 at 14:11
  • 7
    We have a lot of questions that have got too much upvotes in my point of view while some good answers are basically underrated. Of course good questions should be upvoted, but sometimes even the OP 'forgets' to upvote the answer(s) or even to accept (one of them)
    – user31729
    Commented Jan 21, 2015 at 16:31
  • 25
    I always upvote a question for which I provide an answer. My thinking is that if a question is sufficiently interesting to induce me to write up an answer, it deserves an upvote from me. If I don't provide an answer, I will still upvote the question if I think it is clear and well put, if it comes with a usable MWE, and/or if it's original and interesting. Please don't ask me to define "interesting" -- I know it when I see it... I admit I tend to be more forthcoming with upvoting answers, especially if I learn something I didn't know before about TeX, LaTeX, or typography in general.
    – Mico
    Commented Jan 21, 2015 at 20:22
  • 3
    Answers can show up in searches independent of questions, and there are sometimes good answers to poor questions. A good answer can stand on its own, even if the question itself is not a very useful one. The same is, of course, also true of questions: a good question can stand independently, even if the answers to it (if any) are weak. However, good questions tend to get few votes. I don't think the same is generally true of answers, although some excellent answers certainly attract few votes.
    – cfr
    Commented Jan 21, 2015 at 23:11
  • 3
    I don't know if I'm the only one, but I just forget. Absolutely true. I forget to upvote questions, and answers, but since I usually read first the question, and then the answer, I click the upvote in the answer and forget to upvote the question.
    – Manuel
    Commented Jan 23, 2015 at 23:42
  • related: meta.tex.stackexchange.com/questions/2655/…
    – cmhughes
    Commented Jan 24, 2015 at 19:46
  • 1
    Math.SE is full of well-written answers to poorly-written questions. I have answered questions after downvoting them there. Commented Jan 27, 2015 at 13:21
  • 2
    @MatthewLeingang Then you are encouraging people to post poor questions. Instead of down voting and answering I down vote and comment as to how the question may be improved. If they edit their question, I upvote it and then answer. If they don't I ignore it. Commented Jan 28, 2015 at 20:05
  • Related: When should I vote?
    – Werner Mod
    Commented Mar 9, 2015 at 20:02

4 Answers 4


The following statement

...a person who comes up with a good solution to a problem deserves more credit than the one who could not solve it.

is completely subjective and in my opinion wrong. It's very easy to ask a question (good or bad), but it's not always easy to write a good answer. The SE reputation system subjectively supported this approach by awarding only +5 reputation for an upvote to a question, but +10 for an upvote to an answer (this was updated in November 2019 to +10 on any post). Moreover, question can be downvoted without loss of reputation to the voter. Perhaps the motivation here is that the answer(s) have more value than the question. Of course, this is not a universal acceptance, since we have some good questions on this site, but that should be a consideration.

Good questions require a combination of things, including (but not limited to)

  1. a good title;

  2. a show-of-effort that possibly reproduces the problem;

  3. suggested attempts at fixing it (if possible);

  4. images showing the expected output (if possible); and

  5. a consideration for the community and future visitors.

Here are my motivations behind the above:

  1. The title is important, as it draws the first attention from visitors. Don't use something like "ToC problem" or "Help needed please" or anything that doesn't provide a succinct description of the problem.

    If you don't have a good title, it might not show up as expected in search, and users might not even click on the post when they see it in search results. It's like having a house without a proper address associated with; you won't get many visitors.

  2. When answering a question I don't want to recreate the wheel from scratch. I don't want to copy a code-snippet and spend even 5 minutes trying to wrap it into something that is compilable. I really don't and it's not necessary. That is something that the OP should do, because for them it is super-easy.

    Similar to walking into a big room full of clutter with a lonely light in the middle. Invariably, while I'm walking towards the light to understand what's going on, I'm unnecessarily tripping on other things inside the room. Providing the community with a well-lit room makes the problem so much easier to absorb. Copy-and-paste-and-compile and tadaa. Done. To me that shows that the OP took the time to present the problem.

  3. Surely not everyone enjoys fixing things, and would much rather just scream "Fix it for me." That's okay and it can become a nuisance to request this with every post. However, if you show an attempt to fix it, it also highlights what your train-of-thought is, which could benefit from some explanation in an answer.

  4. An image shows me a 1,000 words without typing it. Even if you create something in MS Word/LibreOffice (because of ease-of-use, say), then it's better than 5 paragraphs of textual positioning and fine-tuning. It makes the consumption of information just so much more fluent.

  5. Even though you may have a problem that needs fixing right this very minute, your post forms part of a community of knowledge. So, consider that when you ask a question, and incorporate all of the above suggestions.

    Previously questions that were so specific and/or didn't show much community traction in terms of usefulness, may have been closed as "Too localized".

Voting is very subjective and should therefore not be considered too personally. As a consequence, people's feeling towards reputation also vary considerably. Some don't care for it, while others obsess over it. For some, it may be

...[all] about the points...

but for others it's not.

Voters could be more prone to voting for certain people rather than the type of post. Sometimes you want to vote without reading the post (question or answer), but because of a previous record from that specific user.

I occasionally use a very literal interpretation of the voting buttons:

  • Upvote: This [post] shows research effort; is useful and clear

  • Downvote: This [post] does not show any research effort; it is unclear or not useful

If any of these motivations are not met adequately, then I might abstain from voting and consider my contribution neutral (even if I answered). Perhaps I don't understand the field being investigated, so it doesn't benefit from either up- or downvoting.

  • 2
    These are just my thoughts on the subject...
    – Werner Mod
    Commented Jan 21, 2015 at 20:34
  • I was reminded of your second thought, because it's the same for me at How to add backref citation page-number in the bibliography? natbib package and abbrvnat style :-)
    – Johannes_B
    Commented Jan 22, 2015 at 21:52
  • 1
    @Werner: Well written!
    – user31729
    Commented Jan 23, 2015 at 8:46
  • 1
    Your thoughts ignore what a "question" means. On many SE sites a question is often a sentence or two, and may represent as little as a minute's worth of thinking and typing. On research sites like MathOverflow and CSTheory.SE (and perhaps some other SE sites), a question may require hours of intense thinking about some issue, formulating multiple versions of the question until there is one that makes sense, finding and checking relevant literature, and then writing it all up -- a highly nontrivial exercise, even if the question ends up meaningless or uninteresting due to some missed insight. Commented Jan 23, 2015 at 11:16
  • I believe asking a good question is usually more difficult than providing a good answer. If the opposite were true, we wouldn't have so many poor questions. Remember you have to assess the difficulty from the person writings perspective. It is extremely difficult to ask a good question when you are starting out in a subject (many of our users). Commented Jan 28, 2015 at 20:00
  • At least the +5/+10 distinction is gone. I'm sure some of you benefited as well. I also read that initially there was no such distinction but it was later introduced. Commented Apr 8, 2020 at 22:05

My logic is this: Suppose people found the answer relevant, and that the answer does exactly what I asked for and nothing else (and does not provide any additional perspectives, like explaining how the TeX kernel functions). Then it must be because they found the question relevant, at least to some extent. So why not up-vote the question, too?

I can tell you from experience that this is not always true. Sometimes I run across an answer which is very comprehensive and well written, even though I don't actually care about the problem it was written to solve. I often upvote those answers, but that doesn't make the question relevant to me.

Besides, I don't upvote questions just for being relevant. A question worthy of an upvote, as far as I'm concerned, generally has to be reasonably well written and demonstrate prior effort on the asker's part to solve the problem. Granted, most of the time when a question actually helps me solve a problem I will upvote it, but if the question is particularly terrible I'll just abstain because it seems to be doing the site a disservice to upvote such questions.


Note that there is a difference in the texts on the up/down vote buttons for questions and for answers.

For answers it is simply "This answer is useful"/"This answer is not useful". For questions however, the text is "This question shows research effort; it is useful and clear"/"This question shows no research effort; it is unclear or not useful".

There is thus a higher barrier to a question getting an upvote than an answer; it has to satisfy more criteria.

  • 1
    Yet with a higher barrier to entry, questions still are "rewarded" less than answers. Either SE needs to stop pretending that questions are important and are the seeds from which great collections of answers grow, or the system needs to stop subtly hinting that good questions are not as important as good answers. That attitude makes sense for SO, not so much for several other SE sites. Commented Jan 23, 2015 at 11:22

There is no relation whatsoever between the voting pattern and the usefulness of the question/answer. Everyone has a different model in mind. I use the upvote button for Mark as read. Others vote for appreciation of technical whizz. Some vote for stuff they like. Some use it for evaluation. We have the most upvoting squad of the network. But our question quality is probably very similar to other networks. This implies that our behavior is different than others.

You cannot impose meaning on arrows as Stack network hopes to implement. There are more exceptions than the rule.

You might say this is a vote-for-familiar-user but not-for-strangers OR belittling-the-noobz OR I-don't-care-about-the-OP-give-me-problems-to-procrastinate or something else.

All or none of them might be true. Though I feel we are becoming more and more the first item. Stick around a few more months and the numbers would balance up. Plus we have more obey-SO-rules police officers than ever. Plus our network has changed quite significantly since last year so it is even harder to judge.

  • 7
    one addition to your "something else": i-don't-know-anything-about-this-so-i-can't-make-an-intelligent-decision. (doesn't matter whether it's a newbie or an old hand.) Commented Jan 23, 2015 at 13:38

You must log in to answer this question.

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