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.

Most users seem to vote mostly on answers. How can users be encouraged to vote on the questions as well? (This was discussed in podcast 62.)

share|improve this question
I agree. The problem is that people tend to look at good answers and vote to tell others that this was a good solution, and ignore the fact that the question is good. –  awe Oct 21 '09 at 13:11
To conclude: new gold badge, "Electorate", added today: "Voted on 600 questions and has better than 3:1 answer/question vote ratio.", meta.stackoverflow.com/badges/54/electorate –  Peter Mortensen Jan 2 '10 at 13:34

9 Answers 9

I think we might need a badge specifically for voting on questions. That might be a good start.

I was thinking a badge for 600+ question (not answer) votes -- make it a gold badge?

some quick metrics:

# of answer votes: 1,990,598
# of question votes: 516,783

# of users with > 600 question votes: 34
# of users with > 800 question votes: 9
# of users with > 1000 question votes: 6

# of users with > 600 answer votes: 504
# of users with > 800 answer votes: 294
# of users with > 1000 answer votes: 175

But I'd like to hear other suggestions as well..

marking as complete with new gold badge now implemented:

Voted on 600 questions and more than 25% of votes are for questions

share|improve this answer
Silver for 500, Gold for 1000? –  Damien Jul 23 '09 at 10:15
Jeff, how do the numbers break down for 100, 200, 400 question votes? –  John Saunders Jul 23 '09 at 10:16
Bronze after 100 –  Damien Jul 23 '09 at 10:18
It seems to me that 2M votes on answers and 500K votes on questions does not support the argument of less upvotes on questions. There have to be 4 times as many answers then questions easily. –  malach Jul 23 '09 at 10:31
Since we're already talking about the distribution, I suggest 100, 400, 1500. Gold should not be too easy~ –  mafutrct Jul 23 '09 at 10:34
malach, see updated stats. there are 32x more users with 800 answer upvotes than 800 question upvotes –  Jeff Atwood Jul 23 '09 at 10:59
Jeff, I'd still like to see the breakdown of the lower question vote totals. I'd be interested to see where the inflection points are. –  John Saunders Jul 23 '09 at 11:04
What is the \ratio of questions to answers anyway? –  RCIX Jul 23 '09 at 13:11
hey... I got it :) –  Stefano Borini Jan 2 '10 at 9:08
I think this may promote people voting up questions, even questions that ought to be closed. People aren't as likely to vote down questions because voting a question down incurs a penalty. So you'll see questions that have upvotes that should otherwise have been downvoted and closed. –  George Stocker Jan 2 '10 at 19:47
I also disagree with this. Besides what George Stocker said, the ratio requirement actually discourages voting on answers. –  mmyers Jan 3 '10 at 4:46
The answer part comes quite naturally. That's like saying we're going to discourage men from looking at women. Oh noes! :) –  Jeff Atwood Jan 3 '10 at 7:27
Jeff, a couple comments have asked about the ratios of questions to answers, but you haven't addressed it here. It's pretty important. If there were 4 times as many answers as questions, then the 4 times greater number of votes would imply that people were voting for questions just as often as they were for answers, which would mean that encouraging question voting in the form of a badge isn't really needed. –  gnostradamus Jan 6 '10 at 16:37
...Actually, I just looked at some of the stats Greg Hewgill has mined from the data dump (hewgill.com/~greg/stackoverflow/stats.html). The "Votes per question by week" and "Votes per answer by week" graphs show that questions get on average 1.25 votes and answers get on average 2 votes. I'm not sure if this is for all/non-CW posts or up/down/all votes, but it does suggest people vote slightly more for answers than questions. –  gnostradamus Jan 6 '10 at 17:08

In addition, our user pages only show total up/down votes. Splitting those into questions and answers may also help.

share|improve this answer
Quite right! How are we supposed to know what we've done? –  SamB Apr 15 '10 at 20:10

Could the metrics be a little misleading?

By definition only one question per question can be upvoted.

More than one answer per question can be upvoted. Whilst this applies especially perhaps to the questions that ask for many answers (often Community Wiki, e.g. What are some of your oldest programming books that you still use?) it can apply to many other questions. I have only asked one question on SO but upvoted several of the answers as they were helpful to me. If the upvote criterion was changed to 'single most helpful' or 'best' or 'most correct' answer then the question votes and answer votes would be more comparable.

share|improve this answer
This question is a case in point. I've upvoted the question and two answers. So in this case my voting record is 2:1 (A:Q). While I agree more could be done to encourage voting on questions, I'm not 100% convinced it's a huge problem. –  ChrisF Jul 23 '09 at 11:11

Could one option be some kind of banner after you've answered a question that says (something like) "Have you considered up voting this question?" if you haven't voted at all -- After all, I'm only going to answer well written questions; and it is well written questions that we should be encouraging to be upvoted (rather than "What’s your favorite “programmer” cartoon?" or "Is Mono ready for prime time?")

Questions that need clarifications tend to get comments asking for them, not that there is any way to find out if they are updated in the future, but that's another rant altogether ;)

That's how I tend to behave, anyway...

share|improve this answer
I mean the other way around. Say I ask a question, you come along and write an answer, then you'd get prompted to vote for the question (unless you already have) –  Rowland Shaw Aug 16 '09 at 7:34
seems a bit naggy to me; the answer is an implicit contribution of time. We have (sadly) many users who ask tons of questions but never vote; I suspect there are almost none who answer tons of questions but never vote. Answerers understand the value of votes most of all... –  Jeff Atwood Jan 2 '10 at 8:14
@Jeff: it is easy to forget to vote for questions. Possibly a per-user preference would be useful here? –  SamB Apr 15 '10 at 20:12

How about a badge for voting on, say, 100 questions?

share|improve this answer
Just wanted to note that this can be achieved in 4 days, for better or worse. –  Stu Thompson Jul 23 '09 at 9:52
True. Should obviously be a bronze badge, though. In any case, some of the "beginner" badges can be obtained very quickly. Furthermore, making it easily accessible would train new users in the practice of voting on questions, and not on only answers. –  Ola Eldøy Jul 23 '09 at 10:22

I'm compiling my remarks into an answer. Just to make clear: I am in favor of a separate badge for voting on questions, I just do not see a case for it just looking at the number. As I wrote:

It seems to me that 2 million votes on answers and 500,000 votes on questions do not support the argument of less upvotes on questions. There have to be 4 times as many answers than questions easily.

Jeff Atwood then added some numbers about users extensively voting. My answer: So less than 0.4% of the users vote extensively (>800 answer votes). This is statistically not an interesting group. (I would expect more people to vote on answers extensively as there are more answers in the first place.)

What would be interesting to prove the case is the vote/answer and vote/question ratio. My last answer had the id 1170939. Taking away the questions (shared id), using Jeff Atwood's number of 1990598 votes on answers, that gives 2.12 votes per answer. Stack Overflow shows 232022 questions at the moment, with 516783 votes. That is 2.23 votes per question.

When we look at certain posts about when to downvote, what a downvote is to mean, asking for a certain ratio of upvotes and downvotes: what would the criteria be to up/downvote a question?

  • useful for me +
  • useful for the community, beginners +
  • well put +, poorly formulated -
  • earn me a badge +/-
  • prevent from being closed +
  • signal that this should be closed, and I do not have the reputation points to do it -

Again, I would see some badges (gold, silver, bronce). Are there some significant threshholds, points people tend to camp out? Then maybe these points rounded up some are good guesses for badge achievement.

share|improve this answer

Question votes and Answer votes could have a separate, independent voting limit, so that voting on questions would become essentially "free", answer-wise.

Note that the current voting limit is rather low, and on active days using the site without being able to upvote the obviously better answers is a bit frustrating.

share|improve this answer

You could also add equivalents of the Epic and Legendary badges: "used up all votes on 50/150 days". Now that the last 10 votes are only good for questions, that would encourage more question votes even if it wasn't counting only votes on questions.

share|improve this answer

There is already a badge at 300 votes, the (silver) Civic Duty badge. I personally think people are quite free with the up votes, and too stingy with the down votes.

share|improve this answer
Badge(s) for votes on questions only would make a difference –  Ola Eldøy Jul 23 '09 at 10:17
I am not denying that, I'm just pointing out that there is an mechanism to promote voting. It is not like there is no incentive, which could be inferred by your question. –  Stu Thompson Jul 23 '09 at 10:36
The point with this was to encourage voting on questions. There is not so much problem with voting on answer, as this works quite well already. –  awe Oct 21 '09 at 13:09

You must log in to answer this question.

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