What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

Early on, there was a question about whether Stack Overflow should add badges for asking lots of questions. At the time, one answer was:

Getting answers is reward enough for asking questions.

Even then, Stack Overflow was getting more than enough questions, so it didn't make much sense to incentivize more questions.

This accepted Meta answer lists the badges that can be earned by asking quality questions. (There are at least 8). But looking at the list of recent Great Questions, it looks more like a list of old and popular questions. These badges are difficult to come by in the first place, and they're much harder to earn on sites like History even though there are some high-quality questions on that site.

It turns out that there's a higher barrier to ask a question on sites other than Stack Overflow. We optimize for pearls, but we still need plenty of oysters. The Student bronze badge is awarded for the first upvoted1 answer. There are no equivalent gold or silver badges. Therefore, I'd like to propose the following:

  • Student—Asked first question with score of 1 or more
  • Sophomore—Asked 10 questions with score of 1 or more
  • Socrates—Asked 50 questions with score of 1 or more

Do you have any suggestions of better names/levels?


To give you an idea of how many of these badges would be awarded, I wrote a query. The results2 for the original trilogy:

users   gold silver bronze none    Site Name 
------- ---- ------ ------ ------- --------- 
3000344 7127 67713  805778 2194566 Stack Overflow
235053  115  1698   53073  181980  Super User
156773  53   1191   40063  116710  Server Fault

The gold version on SO would be slightly more rare than Unsung Heroes and more common than Great Questions. Obviously, there would be fewer badges for smaller sites:

5545    1    18     917    4628    Philosophy
4032    4    41     802    3230    History 
5663    3    28     917    4746    Parenting 

Several older sites would get none at all:

5778    null 29     1063   4715    Quantitative Finance
7884    null 20     927    6957    Project Management
4333    null 17     1630   2703    Motor Vehicle Maintenance and Repair

These smaller sites get at most 5 questions a day. I suspect that a gold or silver badge would encourage site regulars to increase that number without sacrificing quality. Certainly the requirement for those questions to stay in positive territory would prevent a person from flooding the site3 with terrible questions in order to get the badge.


1. The original spec didn't consider voting at all for Student and Teacher.

2. Your results will vary from mine since I'm looking at slightly newer data. The bronze badge count also differs from the number of Student badges because the badge isn't revoked when a user's question is later downvoted or deleted.

3. There is a potential problem with cookie-cutter or variation-on-a-theme questions. However, I think the typical tools for closing and deleting questions will prevent people from gaming the system for these badges.

share|improve this question
9  
I get the feeling that this runs the risk of encouraging badge hunters to spam questions. Which is fine as long as the questions are of high quality. I suggest that we raise the threshold of 1 vote to something higher. And to penalize for negatively voted or closed questions. (i.e. negative voted should deduct from the total) –  Mysticial Apr 9 at 0:01
1  
@Mysticial: Raising the threshold rather defeats the purpose: encourage more questions on small (even tiny) sites. But I could see an argument for a penalty for negatively-scored questions. (It's easy to ask a good question that gets closed for reasons entirely apart from quality. Duplicates, for instance.) –  Jon Ericson Apr 9 at 0:06
2  
Ah. I was only considering SO when I mentioned raising the threshold. I was thinking, "high enough to get past the robo-reviewers and passerbys that don't read the question". A variable threshold isn't going to work since sites can grow making the badge a moving target. –  Mysticial Apr 9 at 0:08
1  
Am I a bad person for looking at this and having as my first thought "Huh, from those numbers, SO must have just recently crossed over the 3,000,000 users mark. Neat!"? –  Billy Mailman Apr 9 at 1:05
1  
@Billy Mailman: Not a bad person; I'm glad to know people read that far down the post before starting to type their comments. ;-) –  Jon Ericson Apr 9 at 1:26
1  
Do we get similar badges for answers (10+, 50+ answers)? Right now the badge balance is already shifted towards questions. –  Toon Krijthe Apr 9 at 9:09
    
@Toon Krijthe: Perhaps. I've written the query which you are free to play with. Note that at those levels, gold answering badges would be substantially more common than asking badges would be. –  Jon Ericson Apr 9 at 16:33
    
closed and deleted questions would better block user from getting silver / gold badges, in order to prevent abuse (at Programmers, there are already complaints about not enough close votes, no need to make things even harder than that) –  gnat Apr 10 at 13:09
1  
Go deeper. If it is truly difficult to ask questions on a site, to the point that you need to artificially encourage the creation of new questions, it's likely that site should not exist in the first place -- at least not on the SE engine. –  Jeff Atwood Apr 15 at 8:42

7 Answers 7

up vote 18 down vote accepted

No.

Badges are for demonstrating participation in the community. That means answering questions, commenting, voting, flagging. Asking is part of that, yes, but...

Finding that you have a problem that needs solving is not a choice you make. You should not be rewarded for it inherently. Sure, having the courage and humility to ask a decent question is the sign of a good "scientist", but...

In principle, I agree with what you're trying to get at. But, frankly, upvotes have been shown to be a poor indicator of what constitutes a "good" question: more often, they are simply a sign of popularity through simplicity or Skeet tweets… or, worse, a tag that traditionally attracts the LCDs.

In practice, all this would do is encourage people to ask yet more stupid, worthless questions and we have quite enough of that as it is.

Me, a cynic? Never… ;)

Update: I would probably support this proposal if the barrier to entry were, say, a score of 3. Not 1 though.

share|improve this answer
1  
Which is exactly what I'm afraid of. But I'm not against the idea if there is a heavy penalty for negatively voted questions. –  Mysticial Apr 9 at 0:09
    
@Mysticial: Crap questions often end up with +1 (as opposed to +20), not -15 (as opposed to +2). Such is SO in the 2010s. –  Lightness Races in Orbit Apr 9 at 0:10
    
I'm thinking something like: For the gold badge, you need: 1) 50 questions with >= N score. 2) > 95% of all questions (including deleted ones) are >= 0. –  Mysticial Apr 9 at 0:13
    
@Mysticial: Rewards are great but you have to ask why those rewards are there and what behaviour they encourage. Thing is, everybody thinks their own question is good, even though relatively few actually are. So the promise of a badge isn't going to magically increase question quality, and if it's code quantity you're after increasing then we really don't need that. –  Lightness Races in Orbit Apr 9 at 0:15
    
(FYI, my answer and comments pertain to SO only.) –  Lightness Races in Orbit Apr 9 at 0:16
1  
Just to be clear, I'm in favor of people asking good questions. They're so rare that we can use more of them. True, everybody thinks their question is good, but the ones who are wrong will quickly find out when they get downvoted. The negative reinforcement of being pushed away from the badge will probably "correct" the badge-hunters just like the review audits for the robo-reviewers. –  Mysticial Apr 9 at 0:23
8  
If upvotes don't indicate quality in questions, we should fix that. –  Jon Ericson Apr 9 at 1:28
3  
Agreed, @Jon Ericson, but it's an [almost?] intractable problem. How could we possibly ensure that enough users vote on a merit basis? Also, the sentiment on meta is largely "we like voting on these metrics, but really use your votes as you see fit", which is rather contrary to that goal. Once there a given community is sufficiently large (and that bar isn't very high), I think we must accept that votes will always track popularity rather than quality. –  Esoteric Screen Name Apr 9 at 1:54
1  
Yep, sums up my opinion pretty well. 2 things to add: think of all of those Regex questions that get quick upvotes. And also, do we really want to reward the people who ask questions all of the time and almost never answer? –  hichris123 Apr 9 at 2:14
    
@JonEricson: You'd have to fix people. I've danced around this issue a few times before, but given up now. –  Lightness Races in Orbit Apr 9 at 9:05
    
(and I think the other comments above are very observant) –  Lightness Races in Orbit Apr 9 at 9:11
1  
@hichris123: I took a look and while 56% of non-negative regex questions have non-zero score, that's less than .net, c, c++, python, and c#. I'd say it still takes some skill to ask a good regex question. –  Jon Ericson Apr 9 at 21:10
    
@JonEricson To look at only the percentage of upvoted questions is to assume that the question quality in those tags is comparable. I think there is a general feeling that certain tags, such as regex, tend to have a lower average bar, so while the percentage of upvoted questions is lower than these other tags, if the actual question quality is much lower, then that upvoted percentage could still be a lot higher than it should be. –  Servy Apr 11 at 13:41
1  
I think what @jon is implicitly saying here is exactly this: "Finding that you have a problem that needs solving is not a choice you make" these sites aren't actually solving problems for anyone, are they? And that's kind of the point of the SE engine. So the question is, does the topic belong on this engine? –  Jeff Atwood Apr 15 at 8:43
2  
Finding you have a problem isn't a choice. Taking the time to search the site properly for an existing answer before posting, however, is. As is presenting the question in a clear, logical, and correct fashion, with all the necessary supporting code/etc, and having done some actual debugging yourself first. –  DeadMG Apr 15 at 9:53

Maybe, perhaps with the addition of negative badges, eg:

'Deja-vu': Duplicate Questions

'Nobody cares': i++ i ++i

'Clueless': Floating-point compares don't work.

'Codeless' Absolute bone-idle.

'Linked list': Can't be bothered to do ANY debugging.

'Irritating': URGENT

'Spraycan': Walls of code.

'Deadbeat': Cannot use Google.

'Notwork': Any network code with str*() calls in it.

'Troll': 'Member for today' plus downvotes.

'Chutzpah': Wanting a precis because, if you idiots recycle freely-avilable info into a different form, it might not be caught by my Uni anti-plagiarism scripts.

share|improve this answer
    
+1: go on then why not –  Lightness Races in Orbit Apr 9 at 0:17
    
Joking aside, sometimes I think it's a shame that there aren't actual negative badges until I remembered that they'd be 'abused' by people who want to attract negative attention. –  Rapptz Apr 9 at 0:19
1  
+1 because this is entertaining to think about. :) –  Mysticial Apr 9 at 0:32
    
"'Notwork': Any network code with str*() calls in it." - you could always get a job as copy-writer somewhere :) (Could you clarify "precis", by the way?) –  sehe Apr 9 at 0:35
3  
WHAT ABOUT QUESTIONS IN ALL CAPS? SHOULDN'T THEY GET A BADGE TOO? –  Mysticial Apr 9 at 0:37
7  
[Not sure if serious] –  Robert Harvey Apr 9 at 1:22
    
@Mysticial We used to have an all-caps detector... we should get that running again! :P –  hichris123 Apr 9 at 2:17
    
Disagree. It's just not good. people will run away. –  Isaac Rajaei Apr 9 at 8:29
    
Bad idea, badges should encourage good behaviour. –  Toon Krijthe Apr 9 at 9:09
2  
Of course it's bad idea, but it was a good laugh. I may have been a bit drunk when I wrote it:) –  Martin James Apr 9 at 10:17

I've read through the answers so far and the common thread is that asking a lot of questions isn't desirable behaviour and that Score > 0 is not a sufficient condition to prevent abuse. These are fair criticisms and I honestly never thought about the second. Giving a gold badge for spamming the site with 50 questions that are crap, but manage +1 despite being "meh" is obviously a bad idea.1

It turns out that if we eliminate the score requirement altogether, there would be 18,762 gold badges on SO. Simply by requiring each question to have a positive score, we cut the pool by 38%:

Score threshold  Gold badges
---------------  -----------
none             18,762
1                 7,107
2                 3,010
3                 1,434
4                   784
5                   465
6                   292
7                   204
8                   137
9                   104
10                   67

Notice that increasing the barrier to entry only decreases the number of people who get in by about half.2 I could see arguing that the majority of the people who ask 50 or more questions are help vampires3, but the argument gets strained as you crank up the score requirement. Far and away the most common question score is 0; it's hard enough to get a question upvoted even once.

I believe the number of people asking 50+ questions with a score of X drops off so slowly is because asking well is a skill and question score approximately values that skill. Average question score increases as more questions are asked:

Average question score by number of questions asked.

The data gets increasingly noisy as you extend the graph past 100 asked questions because so few people ask that many. It's worth noting that everyone who would earn one of these gold or silver badges has an average question score > 0 even when including closed and deleted questions.


Quite a few suggestions amount to roadblocks in front of badge hunters. I understand the concern, but that seems the wrong tactic here. If you can't prevent gaming the system, it's better to give rewards as quickly as possible. To me teaching people how to ask positively-scored questions is a win. But if not, the best thing to do is to give them their shiny object so they will turn their attention elsewhere. If it injures our sense of fairness, so be it.

A more intractable issue is whether questions are upvoted because they are "useful and clear" or for some other reason. However, the evidence shows that question downvoting increased consistently since the cost was removed:

Question up/down voting by time

The recent sharp downturn likely is the result of questions destined to be deleted and removed from the dataset in the weeks ahead. I've heard a lot of criticisms of Stack Overflow, but being too easy on questions isn't one of them.


1. This is true even on sites not named "Stack Overflow" since we'd like to see more sites approach our flagship's popularity. If a badge goes from being a net positive to a net negative as a site grows, we should find ways to fix the problem. (It's worth noting that these badges, like Student, ought to be awarded one time per user. I assume this was factored into the objections, but I might as well make it clear.)

2. Actually, the decrease per level is often far less than halving. I find the table utterly surprising. Look at that last number: there are 67 people who currently have 50 or more questions scores of at least +10. There's no particular incentive for doing this: some people are just good at (and presumably enjoy) asking questions.

3. I suspect that rather few true help vampires are able to get anywhere close to 50 questions, however.

share|improve this answer
    
Have you considered something that rewards based on a running average of questions? Asked more than open N questions with a score > M, which account for at least 75% of your total (including deleted) questions. –  MichaelT Apr 9 at 22:53

The recurring objection to this proposal, and it's a fair one, is that it encourages question spamming. I think this suggestion could be significantly improved by shifting the badge goalposts:
Student - 1 question with score >= 1
Graduate (I think this is a better name than sophomore) - 10 questions with score >= 2
Socrates - 25 questions with score >= 3

Numbers are adjustable, of course. Adding score as a metric ensures that questions will (well, hopefully) be of at least marginal quality, and the goal is to encourage the repeat asking of decent questions. A quick and dirty poking around in SEDE indicates that SO would yield about 5000 gold badges with these adjustments. Sorry, no robust stats, as using it on mobile is a real pain and I'm lazy.

I also think we should add corresponding badges for answers, since we've got the teacher badge already, and should maintain question/answer badge parity.

share|improve this answer

This seems to make sense only for certain sites. Namely, beta sites that are struggling.

But since badges get rolled out network-wide, a badge for flooding, say, Stack Overflow, might add more problems than it solves.

There's an intrinsic value to organic questions being the focus rather than inventing questions for the sake of a badge.

Perhaps to encourage deep thought before posting a question for the sake of a badge, the silver and gold badges would need to require significantly more votes: maybe 10 and 40 votes? Furthermore I think average question score needs a threshold as well.

share|improve this answer

We optimize for pearls, but you still need plenty of oysters.

A quick note about optimizing for pearls...

From: http://en.wikipedia.org/wiki/Pearl

Almost any shelled mollusk can, by natural processes, produce some kind of "pearl" when an irritating microscopic object becomes trapped within the mollusk's mantle folds, but the great majority of these "pearls" are not valued as gemstones. Nacreous pearls, the best-known and most commercially significant pearls, are primarily produced by two groups of molluscan bivalves or clams. A nacreous pearl is made from layers of nacre, by the same living process as is used in the secretion of the mother of pearl which lines the shell.

So... While almost any shelled mollusk can produce some kind of pearl, it is really only certain kinds of shelled mollusks, those lined with mother of pearl, that produce pearls that are actually valuable.

Lets think about ways to cultivate questions lined with the kind of quality that will hopefully generate real valuable answers rather than trying to simply cultivate more questions.

share|improve this answer

Positive badges: Yes.

Negative badges: No. I think there are enough votedown, close question,.. This is not good idea to look at the negative points.

Even though there are many people and developers are afraid of asking a question. You are not doing crime if you don't have knowledge of some thing. There are lots of language barrier as well which community recognized Editing and Editing badges so it will help the community and make useful questions for others.

There are many people that are getting job from their reputation of StackOverFlow. There are many students and fresh developers that might not be as strong as elder developers and this negative points will effect on his/her reputation. It cause people to leave the community. I know already 3 people that left the community. They just look at the answers but they don't never answer any more.

share|improve this answer
1  
Umm, I don't see the request to add negative badges... unless you were referring to one of the answers. I find OP's request has a potential in growing sites (unlike SOFU) –  Andrew T. Apr 9 at 8:45
    
There was not enough place to comment that's why I pointed as answer –  Isaac Rajaei Apr 9 at 8:49
2  
But by posting this, your answer is referring to OP's request, not that answer, and I believe it's just a joke. –  Andrew T. Apr 9 at 8:52

You must log in to answer this question.

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