What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 130 Stack Exchange communities.

While at least some answers from every question seem to have been up-voted as a "helpful" answer:

  • about 49% of questions have score of 0 or 1.
  • another 21% of questions have a score of 2.

Most questions seem to have an answer or two with score +5. Is it that we have a lot of fluff questions and not useful ones? Or perhaps it just seems more natural to vote for answers than questions? If we want the most useful info on Stack Overflow to filter to the top, how can this disparity be fixed?

I think this may cause real problems for people getting the badges which require +25 or +100 votes on questions.


Edit: I like a lot of the feedback I'm seeing. Two problems are highlighted:

  1. The work flow doesn't promote voting for questions the way it promotes voting for answers. One solution: when a user answers a question, the system should invite them to vote for the question. If someone cares enough to answer a question, then they care enough to vote for it as well.
  2. Users with less than 15 reputation cannot vote for questions or answers. Is this warranted? Maybe these users could provisionally vote for questions and answers and those ratings would only take effect when the user reaches 15 reputation?

Edit: I had no idea this post would be this popular -- or that this would still be a problem by now.

All I ask is this. As you browse Stack Overflow, when you see a good question: vote for it.

If everyone does, then this problem will disappear.

share|improve this question

migrated from stackoverflow.com Jul 25 '09 at 15:19

This question came from our site for professional and enthusiast programmers.

2  
Just commenting that as of 9/9/08 the stats have shifted slightly and more questions are being ranked. Now its ~46% ranked 0 or 1, and ~17% ranked 2. –  Justin Standard Sep 9 '08 at 18:52
32  
I think the real issue is that questions and answers are treated alike in the voting system. As a result, questions get compared to answered in terms of vote and most of the time, answers look more helpful and well crafted than questions. I believe the system should differentiate votes on questions and answers (I think Joel mentioned this idea in the latest podcast(#62)) –  LeakyCode Jul 25 '09 at 16:10
3  
Maybe the people answering the questions are ranking them - by not voting them up? –  Jonathan Leffler Oct 2 '09 at 2:44
15  
Perhaps because the overall quality of the questions are low? –  Lawrence Dol Nov 25 '09 at 19:42
12  
There's also strange situations where a question gets more favorites than it does votes (serverfault.com/questions/3780/…). Are we seeing redundancy due to voting up a question and favoriting a question are basically saying the same thing? –  Dynamo Dec 10 '09 at 21:56
16  
I have favorited, but not upvoted, questions that I think are not great but want to revisit later. I have even favorited questions I have downvoted, or voted to close, because I want to see what happens to them. –  Dour High Arch May 7 '11 at 1:00
1  
@arch check these related feature request questions: meta.stackexchange.com/questions/4719/add-answer-later-tab meta.stackexchange.com/questions/69346/… –  Knu May 23 '11 at 21:17
11  
Is it ironic that this question was voted for over 500 times? –  user23948732856 Apr 28 '13 at 18:56
1  
I think questions need to have two types of votes. One for how well the question is written (code examples, clarity etc). This could be moderated. The second rating would be for value and relevance of the question. Now some of the relevance will be related to the number of views (could be related to title keywords) but more importantly that a question has some value to other programmers. –  Jamie Clayton Jan 5 at 0:50

51 Answers 51

I think this is a usability/motivation issue.

You can't vote up questions on the 'index' pages like you can on reddit/Digg/so on, so people aren't going to go 'that looks cool' and vote it up (or vice versa) before reading it.

IMHO this is a good thing, but...

As I see it, the 'workflow' of reading a question/answer goes like this:

  • User opens page

  • User reads question

    • Unless the question is abnormally good or bad, or otherwise provocative, this isn't likely to elicit any emotional response. It's just a question, carry on.
  • User scrolls down and begins reading answers

    • As there are many answers, and good answers are rewarded by being 'accepted' and also with increased reputation, this puts the user in the mindset of 'make the answers better'

    • The emotional response behind having your answer accepted or upvoted is "I know stuff, I'm smart, I feel good." Likewise, conferring that reward on someone else is quite a powerful thing too. This provides a very strong motivation to rank and provide answers.

  • Because of this motivation, people will put a lot of effort into writing answers (like me with this diatribe) and ranking them.

This works very well for providing and filtering good answers, but there's no such motivation behind voting for questions. For most questions, the strongest response they are likely to elicit is "I have that problem too", which while it's strong, is only going to apply to a small portion of the viewers/answerers.

While I think this is why questions aren't being voted on as much, I don't think you need to go all out to provide more motivation for it, as this would distract from the main goal of writing/ranking the answers. A simple 'nudge' to remind people to vote on questions I think will do the job without any/many adverse effects.

My suggestion for this is simply to make the voting buttons on the question proportionately larger (or make the ones by answers smaller), and possibly change the color or something.

This will draw the reader's attention to them, and send the message 'hey, while you're here, vote on the question before carrying on reading the answers'

share|improve this answer
3  
Any idea why my post is community owned? I don't mind, but I thought that it had to be edited more than 4 times (it's only been edited twice, and by me)? –  Orion Edwards Sep 23 '08 at 21:34
5  
I read the FAQ and realised the question tripped the 'more than 30 answers' thing - didn't know about that. It's a pity, It took me a long time to write that answer. –  Orion Edwards Sep 25 '08 at 2:32
1  
Did you contribute less to the discussion because it's CW? –  Gnome Jan 3 '10 at 14:20
20  
I'm not sure if it's understood in the answer but unstated, or if I'm reading too much into it, but when you're allowed to vote on answers, you have several "competing" answers presented at once---it's easy to rank them and vote for the best/etc. In contrast, you're only allowed to vote for questions when a single one is in front of you, and different questions aren't competing against others per se, and that's the major dynamic difference from SO in reddit/digg/etc. –  Gnome Jan 3 '10 at 14:25
1  
@Gnome: That's exactly right. It's partially usability, but mostly a difference between what makes a "good" answer and a "good" question. Which, in this context, I suppose is a usability issue, of sorts. –  Chris B. Jul 3 '10 at 15:20
4  
Maybe make the vote button under the question instead of next to it, so when they finished reading it, they come across it before continuing to the answers. –  JD Isaacks Jul 23 '10 at 19:42
2  
I do upvote questions, but where the answer to "what is a good answer" is rather clear-cut (and an answer does show cleverness, and knowledge by the writer if it is even only partially right), for "what is a good question" the answer is much more complex. Comming up with good questions is much harder than good answers. Questions do show OP's misunderstandings and problems with the subject matter (why ask otherwise?), so they are at a very distinct disadvantage here. –  vonbrand Apr 9 '13 at 1:14

After tinkering way too much, I've come to the opinion that there's no great reason for down-voting a question.

Remember the old saying "There are no stupid questions" - well, I think that applies here.

  • If the question is too vague, leave a response to that effect (or vote up a comment that states that already).
  • If the question contains a typographic or syntactic problem, edit the question for clarity.
  • If the question is outright spam, or abusive, flag it as such.

But down-voting a question? What's the point? If the question is naive you can simply answer it.

Down-voting lacks any good use cases (as described above) but worse than that, it is open to abuse:

  • If you don't like the person who asked it, you might down-vote it.
  • If you don't like the topic, you might down-vote it.
  • If you don't like the religious or political beliefs of the question, you might down-vote it.

Down-voting now attracts a penalty to the down-voter's rep, that's an interesting development, but I think the simpler and more correct variation would be stop the ability to down-vote a question.

share|improve this answer
22  
I think that people lay too much into down-votes. They are not a personal attack or insult. As the hover text says, a down-vote merely indicates that you do not find the question/answer helpful. –  Morten Christiansen Oct 31 '08 at 16:43
36  
There are tons of questions on SO that simply shouldn't be here e.g. "What song were you listening to when you first figured out how to use pointers?" Most users can't close an inane question, but they can at least down-vote it. –  Earwicker Dec 14 '08 at 23:05
10  
Agreed. The other incredibly annoying and common occurrence - which goes against the spirit of stackoverflow is - an answer that is an indignant "why are you even doing it this way?". This is just another way of saying "don't ask questions noob". Every programming related question is valid - it doesn't matter why unless 'why' can help answer the question. –  Justicle Jul 6 '09 at 2:49
7  
@Morten I don't find 80% of the questions on SO helpful because they don't apply to things that I ever work with (does anyone work with a large % of the stuff discussed on SO?). Of the ones that DO apply to stuff I work with, some of it I know already, so it's not helpful either. Does that mean I should be voting them down? No. –  TM. Aug 20 '09 at 18:23
3  
@Justicle +1 I'm glad someone else has noticed that. I tend to lean toward asking really obscure questions and I find that people who don'e have the slightest clue usually give that type of answer. –  Evan Plaice Jun 21 '10 at 20:35
6  
I do agree that questions should be downvoted. If somebody doesn't ask a good question, downvote it to encourage them to drop it and replace it with a better question. If they stand by it and the community doesn't agree then they'll have to be willing to take a rep hit. Personally, when a question I've asked gets marked as bad-example I usually delete it and replace it with a better one. This feature complements the '5-questions-per-proposal' and cuts down on the amount of bloat included in a proposal. –  Evan Plaice Jun 21 '10 at 20:41
3  
I say, give some motivation to vote for questions (+1 rep). There's already a cap on how many questions you can vote for on any given proposal so why not? It's not like somebody can vote 1000 arbitrary questions all at once for easy rep. I agree completely that there is a lack of a motivating factor to vote on questions. –  Evan Plaice Jun 21 '10 at 20:49

I actually think this is not a problem, it is simply nature showing us how the system should be design.

In my opinion, there is no reason to rate questions, and we should not try to "fix" it but understand why it is happening. and a possible outcome (the correct one in my opinion) is to do without the voting on questions.

Voting on answers is great, but I see no reason for the existence of question voting.

Even worse, while the voting is useless (yet harmless), the absolute "evil" baked into the system is the fact that you get reputation when your question gets voted! That is an "evil" incentive to post questions for other reasons beside the ONLY valid reason: you want to get an answer to your question.

Giving people incentive to post questions just so they can get reputation is a just as lame as Microsoft giving you "points" for searching using their search.. Motivation for searching should be getting good results. Motivation for posting Stack Overflow question is to get an answer, period.

share|improve this answer
11  
IMO a vote for an unanswered question should mean either "I would also like to ask this question" or "I would have asked this question but it was already answered". The "favorite question" feature overlaps this purpose. –  joeforker Feb 25 '09 at 15:28
2  
+1 I agree with your entire post –  staticx Nov 25 '09 at 20:33
8  
I think voting on questions is okay, but it is certainly crazy that an upvote for a question gives as much rep as upvoting an answer. The reason I don't upvote questions more is that this person is getting there question answered, its own reward. Answerers get nothing but rep, so they should get upvotes and rep for their hard work. I would upvote more questions is the rep for questions upvotes was less, like 3 or 5 rep points. –  Patrick Karcher Jan 6 '10 at 22:25
1  
I don't agree; I see good reason to allow and encourage question voting. Without it, what incentive does someone have to make sure their question is good quality? With no voting, I only need to provide the minimum information needed to solve my problem, just enough for someone to attempt to answer it. With voting, I am more likely to try to ask the perfect question which will not only help me but will also be more likely to help future visitors. –  Jesse Webb Aug 1 '13 at 17:07

Remember the old saying "There are no stupid questions" -- well I think that applies here.

I disagree. There clearly can be, and have been, questions that are pointless or off-topic. If Stack Overflow is to stay useful as an information source and not turn into just another Digg or redit programming section, then there is a need for community policing. I think the system right now is pretty decent. By docking a small amount of reputation for every negative vote cast, it encourages people to be more careful with downmods.

I believe that votes to the answers should contribute to the questions rank as outlined here:

http://stackoverflow.uservoice.com/pages/general/suggestions/15865

I agree that answer votes should count towards question score somehow. In that thread, I advocated automatically upmodding a question when you upmod an answer.

I believe that this would have the benefits that come from giving a bonus to the question based on the answer score while still giving freedom to vote the question down if necessary.

Already, I've run into situations where I have downvoted questions I felt were bad even though I upvoted the answers. I believe that any system where the answers contribute to vote score must take this sort of situation into account, and I feel that the method I proposed would be the simplest way to do it both for the users and for the site developers.

share|improve this answer

I think the "Votes" filter tab on the main page should take into account the up mods for answers in calculating a question's total score.

I would like to be voting more questions up, but I still don't have 50 reputation.

share|improve this answer

Also, when you start using the system, you can't vote at all. So a lot of questions and answers might be lacking in votes.

It says I must have 50 reputation in order to begin voting. I'd like to be voting up questions and answers now.

share|improve this answer
3  
It annoyed me at first: but 50 rep isn't that hard, and I think it improves the quality of content in the community when you have to actually get involved to do most things :) –  singpolyma Dec 17 '08 at 15:34
3  
I also found this (50) somewhat daunting at first but it seems really easy to amass enough points. I started earning badges and points right after my first question, which fired me up to earn more. –  Chris Duncombe Rae Dec 27 '08 at 21:47

I think it's more natural to vote for answers rather than questions. To me, the default reason to vote for a question would be to encourage others to answer it. If it's already answered, so the thinking goes, why not just vote up the answer rather than the question?

I don't think that line of thought is the best thing for the site in the long run, but it may be a behavioral issue right now, especially as young as the site is.

share|improve this answer

What about automating it (I'm a programmer ;))?
If you answer it, the question is voted up automatically. If you don't think it's worth it, you can vote it down afterwards.
Will increase reputation points inflation, but there is always a drawback.

share|improve this answer

I think the reason why there is no immediate ability to upvote is to prevent people making lots of sockpuppets and upvoting themselves - enough of that and they would be able to start editing pages and pose a potential problem if there isn't a way of stopping this sort of gaming of the system.

The current situation may not be the best solution, but I do think it is better than leaving the site open to that sort of abuse.

share|improve this answer

I agree that one big reason is the +25 reputation restriction to upvote.
I understand that since this is a beta, people can come up with a handful of new questions and earn those points quickly, but when SO gets more filled with questions, it will become more common that new people come here, search and find the answer they are looking for. And I think finding what you are looking for is totally worth the upvote. Besides, allowing it early would encourage the good practice of searching instead of creating repeated (or re-worded) questions. I've seen may good questions so far that I would like to upvote but I don't want to just throw a random question just to earn the reputation needed to do that.

share|improve this answer
2  
The +25 rep restriction (+15 by now) cannot be an explanation for why people vote more on answers than on questions, because it applies to both, questions and answers. –  Alexander Tobias Heinrich Apr 8 at 14:56

There is a difference between questions and answers (that sentence alone should warrant a "No s**t, Sherlock!" Badge...).

Often, I see questions that are of no personal interest (and therefore not useful) to me, so I see no reason to vote it up. But the answer to those questions maybe useful for me, by giving some additional information that I can use. Or I just think "Whoa, that is some quality content for the site".

On the other hand, when the question is useful to me (because I asked myself the same thing) or if I believe that a question is good and very useful, I +1 it, which does happen a lot more seldom than upping answers.

At the end of the day, the site is about personal benefit: Whenever an article helps me to gain something that helps me in my work, it gets +1.

share|improve this answer
2  
+1 Questions have to compete with thousands of other questions within the problem domain of the site while answers only have to compete in the much smaller domain of a question. An answer may be more useful than the others, but that doesn't necessarily mean the question is more useful than others. –  Leigh Riffel Mar 2 '10 at 19:49

We are only seeing the positive votes, if we could see both +/- votes it might make more sense.

share|improve this answer
1  
you can, once you've gathered 1k rep: stackoverflow.com/help/privileges/established-user –  Alexander Tobias Heinrich Apr 8 at 14:59

I don't understand the motivation behind down-voting a question. Down-voting an answer I get, but unless the question is offensive or spam or something (and there are different ways to deal with those), I'm not sure why you'd need to down-vote a question.

I also don't quite get why you need a +15 reputation to up-vote anything.

share|improve this answer

Well, I know why I haven't voted on any questions. It's because I can't - apparently you need a reputation of 15 just to vote something else up?

While I can understand restricting the ability to vote down, a restriction on the ability to vote up seems a little to much - but this is my first day, so maybe I'm just not used to the new system yet.

share|improve this answer
3  
It really doesn't take long to gain 15 or even 50 reputation, so I don't see it as a problem. I guess it weeds out people trying to play the system by signing up for multiple accounts and upvoting one question that they created themselves. –  DisgruntledGoat Jun 23 '09 at 21:02

Why aren’t people rating questions? You need 15 reputation to vote. (Hopefully this brings me a little closer to being able to do so)

share|improve this answer

Food for thought: While the number of votes for questions is lower, the ration of up/down is relatively close since the end of the Happy, Happy! Joy, Joy! times of Beta.

Average Score of Questions vs. Answers by Month (from Aug 2009 dump)

Stack Overflow: Average Score of Questions vs. Answers by Month

share|improve this answer
3  
Thanks for the graph, but considering that an answer with no votes really drags down the total I'm not sure that chart conveys what some might think, that there some balance between the votes answers gets vs questions asked). Do you have numbers for average and median votes per question vs answer? Or votes cast for the question vs total votes for all answers of that question? Upvoting your post. –  user136460 Oct 1 '09 at 22:52

I upmod things that are interesting to me, that I feel I might need in a future project, that I think would be a valuable FAQ or that I think need a knowledgeable answer to a well asked question.

I can't comment on others motivation.

share|improve this answer

Why people don't vote on questions? Because most questions don't trigger the "oh, that's helpful" response that comes with a good answer. Mostly I guess because questions themselves don't contain much information.

I tend to up-vote only questions that are * well written or * where someone has taken the care to aggregate answers into the question or * which I would have had to type up, if they weren't already there.

share|improve this answer

I'm going to put a vote in agreement with everyone else in that the reputation required to up-vote a question might be hurting the system a bit. Personally, I tend to also up-vote questions that I think are interesting or that I would like to know the answer to as well, but that is generally when there aren't many answers to the question yet. Once there are more answers to the question I tend to up-vote the answers instead.

One thing that I think might be useful in causing more questions to be up-voted is what others have suggested in lowering the bar for when you can start up-voting questions; however, I would go so far as to say as soon as you have 10 points (i.e. one good question or answer worth of up-votes) you should be able to up-vote questions. Then the bar for up-voting the answers could be moved up a bit to say 100 points or so.

However, one thing that might be skewing things a bit is the member base. I'm not sure of the exact numbers of users; however, I have started to notice some common names in answers and it might be that the lack of a broad user base is hurting this part of the beta. I know that I personally tend to ignore questions that I know there is no way of me knowing the answer to so I wouldn't be surprised if other users might be doing the same thing. If you look at the questions with the most up-votes, they tend to either be related to the site itself (i.e. tagged with stackoverflow) or tend to be broad base fundamental topics that everyone would likely be familiar with (i.e. tagged language-agnostic).

This is definitely something that needs to be monitored, but I am quite curious to see if it starts to resolve itself a bit as the user base increases.

share|improve this answer
  1. I seems quite likely to me that both questions and answers will follow distribution that approximates a power law. I have no justification for this option, but I thought I'd throw it out there.

  2. It's likely that the set of programming questions that are seem as most relevant to everyone is a fairly small subset of all the questions that get asked, unless the community is very homogeneous.

  3. Each question is likely to have an answer that is viewed as being the best - if I view question which already has a good answer, I'm more likely to upmod the answer that add my own answer to the question.

share|improve this answer

Many questions address a very specific problem and the answers are not helpful to the casual reader. While we can still recognize the correct/helpful answers to those questions, we do not judge the question to be 'good' or 'bad'. The question just 'is'.

I'm not sure what would constitue a good question. My answer would be: a question that evokes discussion, because it draws attention to something that software engineers should investigate. However, that is just the kind of question for which this website is not intended.

share|improve this answer

I try my best to vote for anything I think is helpful. I think that seeing both up and downs (as kevin d suggested) is a good idea but I would suggest making it so that when you someone up-votes something you up-voted, you get 1 point.

share|improve this answer

There are no good questions, only good answers.

Maybe the quality of a question comes from the sum how many people answered it and how many people rated the answers?

share|improve this answer

I would only vote for question that makes me think: Gee, I'd like to know that too. However, I yet to run into such question. I guess it's the same for most other people. You usually reply to question others don't know, but you do.

Also, it says 'this was helpful' when you hover over a question, which does not make much sense. Answers can be helpful. Maybe something like 'this is interesting' or 'this is a good question' would be much better.

share|improve this answer

I tend to upvote questions when the answers are generalizable, or the problem is a question whose answer has implications for a lot of other problems. I do this so that when I look at tag, say python, then the questions I see will be ones that are almost guaranteed to have relevance to *me, for many values of me.

What is question upvoting really supposed to be for?

share|improve this answer

This is being overthought.

I call Benford's Law on the voting distribution.

share|improve this answer

I would have up voted several things in the past 20 minutes if there wasn't a silly reputation restriction on up votes. I can understand down vote limits, and an can understand a probation period, but it shouldn't kick in automatically it should be applied when your reputation drops, not when you're first starting off.

Now the initial impression is "Why did I just register? I can't do anything." It will also lead to far more duplicate answers because I can't easily say "I agree with this existing answer." I either have to post my own 90% similar answer, or say "@johndoe I agree" which doesn't elevate the answer in the rankings, requires the reader to parse and scroll back to see what johndoe said, and gives johndoe no benefit from their good answer.

This post is a perfect example of this, lots of other people have said "Vote Up shouldn't be restricted" but I have no way to reinforce their point without you reading all this garbage I just wrote...

share|improve this answer

The answer ratings have a graphical interpretation. The little up-arrow pushes the answer up the score-scale, and thus up the page; and the down arrow pushes it down the score-scale, and thus down the page. That a question has a score of 20 vs 200 vs 2000 doesn't tell me anything, as there is no implied comparison with other questions. I have to think about the meaning of the question voting widget every time I see it, as it does not intuitively mean anything.

I would suggest that the question's rating (for whatever purposes that rating has) be inferred from the answer ratings. Maybe the question score could be the sum total of all answer votes.

Inferring from the number of answers is less attractive, as some answer sequences are actually conversation threads, and do not reflect on the quality of the question.

share|improve this answer

Couldn't it automatically infer the value of a question by other metrics like # of answers and # of up votes on answers, # of views, etc.

Answers that are voted up are usually deep answers, but they typically require a deep question. Don't they? Let's turn to the DB to find out.

share|improve this answer

From a design standpoint I reckon you should have the question rating on the browsing pages. It I don't like a question I'm not going to bother with opening it and then rating it. I'd rather have a "bury/promote" function in the question listings themselves.

share|improve this answer

You must log in to answer this question.

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