What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 128 Stack Exchange communities.

Are you influenced by existing votes? That is, if a question or answer has a large number of up votes, are you more likely to vote it up (and similar for down)?

The reason I'm putting this question is this: on these sites we see how other people vote before we do. This almost certainly influences us (see resources, below) - just as we can be influenced by our choice of restaurant (people often naturally pick the busy one).

This trend may also be further influenced by rep - do we give more weight to the answer of someone with 10K rep than someone with 10 - even if they both have e.g. zero votes?

And if this does happen, is it even a problem?

I'm not looking here for solutions - right now I'm just curious as to whether this is something you recognise.

Clarification edit: I'm referring to voting here. Example: if you see a question or answer with 20 plus votes, are you more likely to vote it up than if it has no votes, or a negative number vote?

Bounty: I've been so impressed at the quality of answers and comments here that I'm offering a small bounty to encourage more great responses and discussion.

Wrapped up: Bounty awarded to the answer that received the most votes. In light of the question, it only seemed appropriate...

Resources:

share|improve this question
16  
+1 just to see if it gets the ball rolling :) –  user27414 Sep 3 '10 at 18:52
1  
@Jon B: Well, it worked with your answer, but then you've got 16K rep... =:-) –  Wikis Sep 3 '10 at 18:55
4  
I really love the clonk of my bell when I move across the meadow. –  Pëkka Sep 3 '10 at 21:12
6  
Actually, the issue is probably even more complex because of badges. Quite often people seem to slam on the brakes when an answer/question is at +9, hesitating to hand out the 'Nice answer'/'Nice question' badge and waiting for someone else to do that. Once someone does cast that one crucial vote, the voting quickly regains momentum, only to lose it again at +24 ("Should I really give him that silver badge? I dunno, let someone else figure it out.") –  ЯegDwight Sep 3 '10 at 22:29
    
@Reg - interesting - why would people hesitate to hand out a badge? Currently this question has 8 votes, so that might happen here as well! –  Wikis Sep 3 '10 at 22:54
    
Related: Should good answers always be upvoted? –  ire_and_curses Sep 4 '10 at 5:42
1  
There is an attempt to start a series of 'collaborative' data analysis project on stats.stackexchange.com. The first project kickstarted with this question: What should be our first Polystats project?. You may want to consider adding the above issue as a project using SE data as an answer to that question. –  user149013 Sep 12 '10 at 13:15
    
@Skrikant - thanks, done! stats.stackexchange.com/questions/2512/… I'd be interested if this works because it would provide an objective answer to this question. –  Wikis Sep 12 '10 at 15:11
1  
@Pekka: Definitely needs more cowbell! –  Piskvor Sep 13 '10 at 8:32
    
@Piskvor ahahahaha! I mean, moooooooo –  Pëkka Sep 13 '10 at 8:52
    
@RegDwight - that was an interesting observation (about badges). I didn't see it operate here, however. –  Wikis Sep 13 '10 at 16:15
1  
Herds illustrated: dilbert.com/fast/2010-09-23 –  Wikis Sep 23 '10 at 7:41
    
Yes votes and the outcome of discussions here are strongly determined by what the most influential people out there do and say: I even saw it happen that an inoccent question asking for help and opinion about an issue, asked by the OP in good faith and intention to help, was voted to a net score of above +10 and then this net score fell like a stone to currently about -40, as soon as a well known and by the community highly respected user had posted his answer wherein he strongly (and IMHO unjustified) questioned the good intentions of the OP. Such herd behavior is IMHO a big problem. –  Dilaton Nov 11 '13 at 15:40
    
I always find myself influenced in the Review Queue... I can see that 3 people have voted to close for a particular reason, I probably will select the same reason –  Dan Hanly Nov 11 '13 at 16:08
2  
@Dilaton I didn't say I wanted to do it! It's something that I've noticed that I do whilst I review. In fact, I'd go so far as to suggest that I think the votes of others should be removed from the review process. I don't like the fact that I'm influenced, but I still find myself influenced. –  Dan Hanly Nov 11 '13 at 16:24
show 1 more comment

7 Answers 7

up vote 33 down vote accepted
+50

Yes, I am influenced by existing vote counts and the user's rep.

If I come across an old post, I usually read the top few answers. I don't bother reading the ones at the end, since everyone else is tipping me off that they're not as good. This is undoubtedly flawed.

I also pay attention to rep, as well as my own personal experience with a poster. If I come across a [c#] question where Jon Skeet has answered it, I know he's almost certainly right. Again, this is flawed.

Is this a problem?

Short answer: yes. Long answer: no.

If we were to show all answers anonymously until you voted, we would solve the problems of my lazy behavior described above. However, that would make the system less usable in general, and I would probably not bother reading/voting at all.

The root cause of the problem is laziness and human nature. Anything else we might try will also be tainted by those issues.

So are we doomed?

Contrary to these tendencies, there are plenty of examples of posts where the opposite has occurred. Sometimes high rep users are downvoted while newbies (with better answers) are upvoted. Sometimes a new answer will appear when there is already a clear leader, and then overtake that leader.

In conclusion, it ain't perfect, but it's the best we got.

Stack Overflow will continue to refine its model to minimize these problems. One day there will be a paradigm shift and we'll all be doing something completely different instead (notice that we're not discussing this on phpBB right now).

share|improve this answer
24  
I was thinking of upvoting this, but no one's upvoted it yet so I'm wary of doing so. –  James McNellis Sep 3 '10 at 18:40
    
@Jon - thanks, that's a very detailed and well thought out answer. And honest. One nuance to my question (perhaps it was not clear?): I'm especially thinking here of question / answer votes to go quickly in one direction. E.g. if a question has 20 up votes, are you more likely to vote it up than if it is on zero - or has a negative figure? –  Wikis Sep 3 '10 at 18:40
    
@James - nice =:-) and now you can! –  Wikis Sep 3 '10 at 18:41
2  
@Mark One is not enough. It could be a stray. –  Pëkka Sep 3 '10 at 18:41
    
Need to watch the vote count to see whether somebody else does something about this. –  Pëkka Sep 3 '10 at 18:42
    
@Pekka- just wait a few minutes - once three or four people vote the other votes will pour in... =:-) –  Wikis Sep 3 '10 at 18:44
9  
Upvoted because Jon has 16.5k Rep. –  gnostradamus Sep 3 '10 at 18:46
1  
@gnostradamus good point. Upvoting too. –  Pëkka Sep 3 '10 at 18:49
1  
@Mark - personally, I am more likely to upvote only to a point. I would probably +1 a question with a score of 5-10 more likely than I would if it was a 0 or negative. However, when I come across one of those epic +100's questions, I usually don't bother. –  user27414 Sep 3 '10 at 18:50
11  
I didn't even read this, but I upvoted because everyone else is doing it. –  mmyers Sep 3 '10 at 18:51
4  
Achievement unlocked: [rep-whore] –  user27414 Sep 3 '10 at 18:53
2  
@Jon B: I read the comments, of course. That's how I know what the cool kids are doing. –  mmyers Sep 3 '10 at 18:57
1  
Some of these comments are very funny. It was worth posting this question just to see the community cracking jokes... –  Wikis Sep 3 '10 at 18:57
3  
I did read this, and I liked it, and I wanted to upvote it. But then, right after reading the last sentence, I noticed Jon's rep, so I concluded I must be biased and abstained from voting. Take that, Jon, I won't be fooled by the herd! –  ЯegDwight Sep 3 '10 at 22:07
5  
@Reg - you're a non-conformist, just like all the other non-conformists. –  user27414 Sep 3 '10 at 23:41
show 3 more comments

Yes, I would admit casting/not casting a vote based on the current activity.

The "Everything's in Order Here" Syndrome

If I come in after the voting activity has already settled in, and the voting levels seem to reflect the best posts, I often feel like "everything's in order here... move on."

The Exceptional Standout

If any particular post is truly exceptional, I might want to throw my vote into the ring, despite the already-high vote score.

The "Unsung Hero"

But, if activity is low and I see something interesting, I will more likely up-vote it (to celebrate the "unsung hero", so to speak).

The Intervention (aka, you haven't hit bottom, yet)

And, yes, I don't tend to down-vote an already beaten-down post... unless I think it is really so bad that the barrage of down-votes has not yet truly reflected its deserved bottom.

share|improve this answer
3  
+1 Very well put. These tend to be my "patterns" too. –  Pëkka Sep 3 '10 at 20:20
5  
Then there's the "great answer, but I answered too" syndrome, where you don't +1 a good answer because it's competing with your own. –  user27414 Sep 3 '10 at 20:28
    
@Jon oh yeah, I know that too. But as long as the answers both go into the same direction, and are roughly of the same quality, I find that okay. ("Quality" is of course tougher to classify on Meta than on SO proper.) –  Pëkka Sep 3 '10 at 20:34
    
@All - another great answer (Robert) and great comments. This is turning out to my more enlightening than I imagined. I'm really enjoying reading the thought provoking posts and the humourous one. Thanks! –  Wikis Sep 3 '10 at 20:44
2  
@Robert Cartaino: Out of interest, why did you choose to copy this answer verbatim from meta.stackexchange.com/questions/15706/…, rather than just referring to it? –  ire_and_curses Sep 4 '10 at 5:45
1  
If anything, this is the opposite of herd behaviour: you're mainly voting in disagreement with the consensus, rather than agreeing with it. –  Andrew Grimm Sep 4 '10 at 11:03
2  
@ire_and_curses: Because it felt really lazy to drop a link to go read another answer which only coincidentally happened to answer this question. –  Robert Cartaino Sep 5 '10 at 15:38
add comment

And if this does happen, is it even a problem?

Problem?? Herding behavior is the very foundation of social media and the social web!

A while back I wrote about intermittent (random) reinforcement. But if voting appeared to be completely random, people would just post junk all the time. Community members still need to be told when and why they were given reinforcement, even if it's only given some of the time.

Herding behavior, information cascade, and all of the related phenomena are the counterbalance to truly random reinforcement. Advanced technical questions ask voters to make difficult decisions and those voters don't have the time or inclination to test and thoroughly evaluate everything. Most will simply eyeball it and maybe vote based on their immediate reaction; if other people have voted similarly, that makes them more likely to act on that instinct.

And because of that, there emerges a quasi-realistic picture of which questions and answers are actually "good", and what you need to change (or continue doing) to pick up more votes.

Without this, at the speed most questions fly by, votes on your answers would all be little onesy-twoseys. You'd get the same kinds of feedback on 1-line half-assed answers that you would on picture-perfect reference code.

Of course, this leads to several undesirable side-effects, most notably the bikeshed effect (soft/easy questions/answers getting the most votes). But on the whole, it is a good thing. It is evolutionary. It is how people learn to function within a group. During the early stages, people aren't ready to think and act independently; the only thing keeping them from fleeing is the knowledge or perception that if they do what everybody else does, they'll be safe from harm or ridicule.

Why do you think internet memes are so popular? It's the safest form of humour. It's guaranteed not to offend the majority of the community (although it will obviously irritate some). It's a way to feel accepted without actually being noticed.

Eventually most people grow out of this stage and do learn to think and act independently, but for many, that takes a long time. In the meantime, they do what they're biologically programmed to do and herd. (Some people never leave this stage - obviously, then it becomes a problem for them personally.)

The "wisdom of crowds" is far from perfect, but absent some sort of automated testing facility, it's about the best we presently can do, and that's largely thanks to herding. Herding doesn't just influence these sites; it defines them, and almost every other form of social media.

And for the most part - we shouldn't worry about it. It's part of how stable societies are formed.

share|improve this answer
2  
@Aarobot - reading your answer just confirms my comment above. This is amazing stuff, really eye-opening to me. It's very interesting how you've taken my question, "is this a problem?" and looked at it from a completely different angle - as a blessing! But I have to add: it's not very herd like of you... =:-) –  Wikis Sep 3 '10 at 21:09
5  
This answer (being the last one added) is currently lowest in terms of votes, shows a frightening amount of independent thought, and when I scroll to the bottom, I can't see the other answers any more. Exactly the kind of stray answer they will come for! (Looking around nervously) Sorry Aarobot, but I shouldn't be seen in your company. Was nice knowing you –  Pëkka Sep 3 '10 at 21:40
    
@Pekka - fortunately, nobody except us will get far enough to read any of these comments... =:-) –  Wikis Sep 3 '10 at 21:42
    
@Mark Yeah. Except them. Let's get out of here now –  Pëkka Sep 3 '10 at 21:55
1  
@Pekka: There's only 3 answers to the question, and I was an hour later than any others, so that's not shocking. –  Aarobot Sep 3 '10 at 22:17
    
And @Mark: I've been here way too long to still be herding. –  Aarobot Sep 3 '10 at 22:18
1  
@Aarobot I said as much: It's lowest voted because it's youngest, not because it's of bad quality. –  Pëkka Sep 3 '10 at 22:25
1  
@Aarobot: again, great answer. Now - is this a competition to see who is going to stick around the longest to prove we're not herding, or are we now going to venture into the three dimensional world...? =:-) –  Wikis Sep 3 '10 at 22:31
    
@Mark: Certainly not. I doubt very much that anybody who actively participates on MSO is really part of the "herd" (although everybody herds sometimes, it's just a question of when and where). –  Aarobot Sep 3 '10 at 22:32
    
@Mark But if this is an psychological experiment, wouldn't then the fact that we, the participant, know that we're in an experiment affect the outcome of this experiment? ;) –  Yi Jiang Sep 4 '10 at 15:47
    
@Yi Jiang, That too, is part of the experiment. –  Zoredache Sep 7 '10 at 18:30
    
@Yi Jiang / @Zoredache - I must confess I'm a little bit lost, now... –  Wikis Sep 7 '10 at 18:38
    
@Yi Jiang / @Zoredache / @Mark Robinson : Can't you just take that to meta.meta.stackoverflow.com ? –  Piskvor Sep 13 '10 at 8:30
    
@Piskvor - if only it still existed (not). –  Wikis Sep 13 '10 at 16:20
add comment

Somewhat related to what Aarobot said: It (herding) is not generally a problem, because it acts as an amplifier. By taking a dataset with low signal-to-noise ratio and amping it up, it becomes slightly easier to recognize the signal.

share|improve this answer
    
so you're saying that herding acts as a positive filter or catalyst for already good questions / answers? –  Wikis Sep 8 '10 at 6:22
2  
@Mark: No, he's saying that there's a lot of noise in the voting patterns, and herding basically increases the votes without significantly changing the SNR; it's like tuning in to a bad radio station and having to crank up the volume in order to hear the voices over the static. Or, alternatively, like blowing up a tiny image; it's still fuzzy and pixelated but it's still sometimes easier for your brain to make out the details if they're big. Other types of "amplifiers" like Electorate are non-linear; they increase the votes but increase the noise even more. –  Aarobot Sep 13 '10 at 15:52
    
@Aarobot - thanks for the clarification. –  Wikis Sep 13 '10 at 16:21
add comment

On SE I upvote an answer only if it's correct and helpful. Others also found it helpful, so it's upvoted. But not vice versa. I don't cast votes just because it's heavily upvoted or posted by a high rep user.

And if this does happen, is it even a problem?

Not on SE, but on Meta sites upvotes mean agreement, downvotes mean disagreement. This kind of voting isolates the community, it tends to extinguish opinions that's differs from the opinion of the majority. So customs will develop, and participation becomes increasingly difficult.

share|improve this answer
add comment

Yes, herd behavior does play a part in SO.

An iconic example: http://stackoverflow.com/users
Jon Skeet has more nice answer badges (1448) than anybody else has combined badges.

No offense to Jon, but it's unlikely that he's so far ahead of everyone else.

share|improve this answer
    
I'm sure you're right. @JonSkeet has his own tag & meme. There are others who are notable, sometimes for less noble reasons... :) –  Wikis Sep 21 '10 at 7:15
add comment

Are you influenced by existing votes?

No. I believe that factual information outweighs groupthink.

share|improve this answer
    
@geekster - thanks for the reply - so, do you read all answers? –  Wikis Sep 13 '10 at 16:23
    
@geekster - although, at time of writing, you haven't posted any questions or voted...! =:-) –  Wikis Sep 13 '10 at 16:28
1  
Every victim of groupthink believes that factual information outweighs groupthink. That's how the phenomenon works. But social herding and groupthink are not the same thing, so I guess the point is moot anyway. –  Aarobot Sep 13 '10 at 16:54
add comment

You must log in to answer this question.

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