What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 122 Stack Exchange communities.

I'm writing a paper comparing the mechanics of Stack Overflow to Reddit, and one thing I can't find a solid answer to is why Stack Overflow makes a user's reputation so visible.

A user's karma on Reddit is essentially invisible -- unless you're looking at a moderator user list, or have clicked through to a user's profile, you really don't see their karma scores. I'm thinking this is to let content to stand on its own without having a particular user's involvement/influence (or lack thereof) overly influence the response to a submission.

Comparatively, Stack Overflow lists a user's reputation on every question and answer.

I have a few thoughts why this might be the case. For instance, if dealing with sensitive data and needing an authoritative answer, it might be more preferable to take the answer of a user with more rep than an unregistered user's response if neither answer has been upvoted. But that's a fairly big assumption to make, and a fairly rare scenario.

Am I missing something obvious?


share|improve this question
add comment

4 Answers

up vote 5 down vote accepted

On stack overflow an answer either fixes the problem, or it doesn't. The solution is not subjective - you type it into the editor, compile it, run it, and it either works or it doesn't. There are subjective aspects to the posts (such as explanations, how well researched/referenced the post is, etc) but if an answer won't work, it gets fewer upvotes (generally) than an answer that does work.

Most other websites that implement similar-looking systems are fundamentally different at this level. User vote on what is interesting to them, and those votes give a user a social currency boost - but ultimately from subjective material.

Here the reputation is partly social and subjective, but largely when I vote something up it's due to its correctness and according to my set of best practices.

On meta you'll find the voting wildly different because it is primarily subjective.

But on Stack Overflow the reputation isn't just a number, and isn't just a game (although it's both) it's a number that can reasonably be used to measure one's expertise and participation level on Stack Overflow. It's not just how many people liked their posts due to interestingness, but how many other programmers believed that their answers are not only correct, but a good approach to the problems.

So you can't compare reputation with karma and other social and sharing websites without taking into account the nature of the posts and voting here.

The reason it's so blatantly exposed is firstly to encourage participation and secondly so people can compare answer providers beyond the initial does-it-work test.

If you have two people giving you two solutions to your problem (there is more than one way to do something), and one is an expert with a decade of experience, and another is relatively new to the field, you'll be more likely to choose the path of the expert. Usually that's visible in the way the answer is written, the reasoning they give you for the solution they suggest, and so forth, but that's not necessarily easy for a new programming to measure, so reputation gives them an instant, "The community has this much confidence in their past answers" measurement which will guide them in their choice.

But the primary reason, again, is to encourage highly competitive programmers to climb the ladder. There are two systems - reputation, and badges - and they are both always shown when the user's avatar is shown. The two systems appeal to different aspects of competitive behavior.

There is a very informative post by Rands called Gaming the System which goes over in some detail about how to set up a game around business objectives that will draw geeks in. After explaining how and why it works, he reiterates the key points:

The Rules of the Game

Now that we understand how games float the geek boat, we can tease out rules you can use to build your own business-centric games. This is will take a creative leap on your part because I don’t know how your particular situation is grim. Perhaps your bug count is crap like mine? Maybe you can’t hire fast enough? Maybe you can’t measure how screwed you are? I don’t know what game you need, but I know you need to follow the universal rules of games:

The rules need to be clear. Whatever game you design must stand up to scrutiny. Test the rules with selected geeks before you roll it out. Find the holes in your game before you’re standing in front of the team describing a game that makes no sense. Ambiguity, contradiction and omission are the death of any good game.

The rules must be inviolable. Enforce rules with an iron fist. A rule not followed is twice as bad as a poorly defined one. A violation of the rules is an affront to a geek. They react violently to violations of the rules because it’s an indication that the system is not working. Rules make a game fair, and when they stop being followed, the geeks stop playing.

The playing of the game must be inclusive, visible, and broadcasted. Include everyone on the team. Those not on the team should be aware of the progress and implications.

Only use money as a reward as a last resort. It’s a knee-jerk management move to use money as an incentive. Problem is, money creates drama. Money makes everyone serious, and while you may be in dire straits as you design your game, you don’t want the team stressing about who is getting paid; you want them to stress about the work.

This is not to say that rewards in a motivational game are verboten, but step away from the money and think about achievements. One of the best trophies I’ve awarded was a horrifically ugly ceramic blue rhino the size of a pit bull. The winner proudly displayed the rhino achievement in his office for years.

It’s not a game. Just because I’m using the word game all over this article doesn’t mean it’s trivial, simple, or something not to be taken seriously. Your geeks will treat the game as a motivational tool as seriously as you choose to treat it in building and rolling it out — because they want to win.

The playing of the game must be inclusive, visible, and broadcasted.

Stack Overflow is not much more than a massively multiplayer problem solving game.

And there are cookies.

share|improve this answer
"On stack overflow an answer either fixes the problem, or it doesn't. End of story. It's not subjective - you type it into the editor, compile it, run it, and it either works or it doesn't." That's simply not true for many questions. An answer which explains why things work the way they do is more valuable than one which doesn't - and the quality of explanations can certainly be subjective. –  Jon Skeet Jan 20 '12 at 16:53
@JonSkeet Answers either solve the problem or they don't - while explanations are good, solutions are king. There is a lot of content which is judged subjectively, but stack overflow discourages questions which are not problems to be solved, and answers which do not solve the problem. The solution - whether an explanation is included or not - either works or it doesn't. Yes, there is a lot of content beyond the solution which is judged subjectively, but if the answer is wrong, it doesn't matter. So I still stand by the idea that whether the solution works or not is largely objective. –  Adam Davis Jan 20 '12 at 17:39
When the question is "Why isn't this working?" the answer is an explanation. Your second sentence is "end of story" which seems in direct contradiction to "there is a lot of content beyond the solution which is judged subjectively" - clearly there's a lot more "story" than your first paragraph would suggest. –  Jon Skeet Jan 20 '12 at 17:42
@JonSkeet I believe I understand your point, and have modified my post. –  Adam Davis Jan 20 '12 at 17:50
I found this answer to be the most helpful and complete. Marking as accepted. –  aendrew Jan 22 '12 at 13:34
add comment

Reputation is meant to be part of the incentive to contribute well - the more visible it is, the more incentive it provides. There's a reason it's called "reputation" instead of "karma" after all.

Part of the "game" of Stack Overflow is to improve your reputation - hence the visibility. At least, that's my take on it.

share|improve this answer
Thanks for the response! Would you say reputation serves little purpose in helping users decide which answer to select? Would you ever be more predisposed towards a particular answer if its author has a high reputation? –  aendrew Jan 20 '12 at 15:46
@aendrew: I wouldn't personally, but then I'm probably not an entirely typical user. I'm sure it makes some difference to some people, but hopefully not too much. –  Jon Skeet Jan 20 '12 at 15:48
Coming from the man who makes reputation moot for the rest of us. =) –  casperOne Jan 20 '12 at 15:49
@aendrew For reference, Jon has the most reputation of any user on any Stack Exchange site. He also has almost 50% more reputation on Stack Overflow than the second highest user by rep on Stack Overflow. Additionally, his reputation is exponentially higher than most people in the top 10% of Stack Overflow. Oh, he's also Chuck Norris. –  casperOne Jan 20 '12 at 15:54
@aendrew: There is a perception that the more rep you have the more you get 'rep goes to rep'. –  Iain Jan 20 '12 at 15:54
Dammit, I just clicked the upvote button.... –  casperOne Jan 20 '12 at 15:57
@aendrew Relevant to your interests regarding voting: blog.stackoverflow.com/2009/04/… and meta.stackoverflow.com/q/115368 (particularly this post in one of the answers: stats.blogoverflow.com/2011/08/…). The conclusion is that users get high rep because their posts are good, as opposed to their posts being viewed as good because they have high rep. I would say users are more inclined to trust Jon Skeet et al. if they are clueless :P –  Matthew Read Jan 20 '12 at 16:13
@aendrew To answer your question about reputation helping users decide which answer to select, I think it's safe to say that the answer is yes, and you can look at Jon's reputation/upvote patterns (btw, the data explorer is great for this) to see the effect that reputation has. At this point, he'd probably hit the rep cap every day even if he wasn't so prolific because in this environment, because of his rep (which is a reflection of the consistency, quality, and persistence of his work) people make the (often correct) assumption that his answer is right. –  casperOne Jan 20 '12 at 17:07
@casperOne: I suspect I'd also hit the rep cap because at this point I've got over 18,500 answers. If 0.11% of my old answers get a single upvote each, I hit the rep cap. That's just a matter of having a very long tail - not a bias over answer selection. –  Jon Skeet Jan 20 '12 at 17:10
@JonSkeet I think it's a matter of both, but I can't say for sure, I don't know what's in people's mind when they vote (you make a very good point about the long tail though). I think that it's fair to say that for new answers, there is some sort of effect because of your reputation. It's also the same effect that say, Eric Lippert has when he posts. –  casperOne Jan 20 '12 at 17:17
@JonSkeet BTW, stop interactively debugging with the users in the comments! =) Go earn more rep or something =P –  casperOne Jan 20 '12 at 17:17
Ohh, actually, here's a challenge for you. Try to hit the rep cap by asking questions (even if you bust, try to get 40+ upvotes on questions per day to synthesize hitting the rep cap with questions). =) –  casperOne Jan 20 '12 at 17:21
@casperOne "At this point, he'd probably hit the rep cap every day even if he wasn't so prolific" The last time I calculated this I came to a number ranging between 400k and 650k rep would suggest one has enough posts that residual voting would hit the rep cap every day - including weekends. I have noticed that the long tail isn't as predictable now as I figured when I did the original analysis - on the other hand I was using my own data, and Jon's posts are more mainstream, so his tail is not only long, but particularly popular. So that's 73k rep per year if he stops posting. –  Adam Davis Jan 20 '12 at 17:47
Can we just declare you the winner? –  LarsTech Jan 20 '12 at 17:54
The challenging thing is, as the person selecting an answer (from what was never really intended to be too subjective of a question in the first place), I now have to consider whether the fact this response currently has over 3 times as many upvotes as its closest rival is a result of that "rep begetting more rep" tendency mentioned earlier. Hrmm...! –  aendrew Jan 20 '12 at 20:38
show 2 more comments

Am I missing something obvious?

Hell yes! The bigger the number by your your avatar is the bigger your...uuhhh...well, you get the idea.

On a more serious note, it is a simple reward-system. You do something good, you get reputation, you do something bad, you lose reputation. While reputation is something virtual which is of no real value, it shows how much people have contributed and that those contributions were good. You can easily compare yourself to other peoples, which is a pretty good stimulus to get better and earn even more reputation. And that results in more good contributions to the community.

It is also used to give only good contributors and community members certain privileges.

share|improve this answer
Thanks for the response; same question as with Jon -- would you be more predisposed to choosing an answer by an author with a higher reputation? –  aendrew Jan 20 '12 at 15:49
@aendrew: I personally try not to be predisposed as such, but I'd be lying if I said reputation didn't have an effect. –  David Jan 20 '12 at 15:59
@aendrew: That's totally personal preference. High reps can be as wrong as a 1-rep user. Personally I accept the answer which is*the answer* and is better. But you'd have to ask every single user on SE about that to get an definitive answer. –  Time Traveling Bobby Jan 21 '12 at 18:17
add comment

I have a few thoughts why this might be the case. For instance, if dealing with sensitive data and needing an authoritative answer, it might be more preferable to take the answer of a user with more rep than an unregistered user's response if neither answer has been upvoted. But that's a fairly big assumption to make, and a fairly rare scenario.

Sensitive data should never be published on any of the stack exchange sites. If it's sensitive,then by definition it shouldn't be published. That said, the scenario you outline either doesn't exist or is extremely temporal as most editors will happily black out the truly sensitive things (usernames/passwords) fairly quickly.

You asked a question as to why SO displays rep so prominently but your assumption is more geared towards the impact of publishing rep. These are sometimes very different things.

There are very few people who can authoritatively answer the question as to why SO displays rep scores the way they do. Even Jon Skeet can only take a guess, even an educated one as it is. So, instead, you should posit that to team@stackoverflow.com.

Moving on to the next logical question: What impact does it have?

This varies by user and there are four situations:

For some the rep score is meaningless in evaluating whether a response is "true" or not. Unfortunately, they are in the minority. These users will tend to upvote all of the correct answers they see on a single question.

A second group uses rep score as a secondary motivator. For instance when there are two very similar answers the one with the higher rep wins.

A third group is similar to the second but with a counter approach: they vote the lower one up instead. Kind of a counter-culture type of thing, but again, they are a minority. I've found myself more in this group lately than anywhere else; as it helps encourage a broader community.

A fourth group uses rep score almost exclusively without regard for the actual answer provided. I believe they are still somewhat of a minority, and I like to call them lemmings. ;)

Bear in mind that there are other contributing factors. For example, those people with high rep scores tend to provide accurate answers faster than others. They also tend to take the time to answer the more difficult questions.

Due to how SO functions, the first one with an answer has an advantage in gathering rep over late comers. After all, if I click on a question and the top answer seems to provide the solution then I am much more likely to upvote that answer instead of providing a competing but similar answer.

Also, those who provide answers to difficult questions have a large advantage over those seeking to only answer the "low hanging fruit" type of questions. The competition is orders of magnitude smaller and, therefore, it is much more likely to get both upvote AND accepted answer points.

share|improve this answer
Thanks for response. I probably should have been clearer about the sensitive data thing -- I'm thinking more in terms of using code that has the potential to unintentionally create a security hole of some sort, and less about posting any sensitive information online. It probably is a bit of a red herring, though -- I've noticed users on SO very quickly jump on potential security vulnerabilities in answer code and give ways to prevent those problems. Good idea about contacting the team directly, though! –  aendrew Jan 20 '12 at 20:32
add comment

You must log in to answer this question.

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