29
$\begingroup$

My question is about the effects of the reputation system, and one's reputation being displayed prominently on his screen and all posts. I understand that reputation is a way to keep the website clean, preserve the intended purpose of the website, but it also seems to create a sub-game where the goal is to maximize one's reputation. Why isn't there a maximum reputation, and why is reputation prominently displayed? I am new here so maybe I am missing something, but when you talk to someone off the internet you don't say "Hello, I am __ , with n peer-reviewed papers." or something like that. One of my concrete concerns is that the reputation system might incentivize answering easy or soft questions to get a quick reputation "high", and that it might also promote some group think. If anyone has a behavioral background, maybe they could comment on what the reputation system does, and if it interferes with the math.

$\endgroup$
  • 26
    $\begingroup$ I agree that some people fish for reputation, but that doesn't mean the answers are bad, which I think is what's important. $\endgroup$ – Git Gud Jan 9 '14 at 23:27
  • 12
    $\begingroup$ If you are worried about other peoples behavior, then the one thing one can say is that these sites work very well compared to similar other sites without reputation and voting. I agree that most advantages come from sorting the answers and from the small reputation changes that prohibit voting and commenting by spam-bots and trolls. However, if you are worried about the effect on your own behavior or well-being then you can install a greasemonkey-script that kills all gamification signs, see meta.stackexchange.com/questions/40777/… $\endgroup$ – Phira Jan 9 '14 at 23:55
  • 5
    $\begingroup$ @Phira The usenet newsgroup sci.math had no rep/gaming and it was actually better than MSE in its heydey. $\endgroup$ – Bill Dubuque Jan 10 '14 at 0:25
  • 10
    $\begingroup$ @Bill: I wasn't there, but if I know my history right this "heydey" was before the Eternal September. $\endgroup$ – Asaf Karagila Jan 10 '14 at 6:49
  • $\begingroup$ Answers to this older question seem somewhat related: meta.math.stackexchange.com/questions/7116/… $\endgroup$ – Martin Sleziak Jan 10 '14 at 7:35
  • 8
    $\begingroup$ @BillDubuque A quick glance at sci.math today shows that it is of extremely low quality, with digressions into theology, holocaust denial and flame-baiting. As Arturo said two years ago: "The signal to noise ratio is increadibly higher [at MSE] than in sci.math, where it has been approaching 0 asymptotically for a while." $\endgroup$ – Chris Taylor Jan 10 '14 at 9:10
  • 6
    $\begingroup$ @Chris What remains today of usenet is just a decaying skeleton - as you noticed. You had to be there in its glory days to really appreciate it. In any case, my point was that usenet newsgroups worked just as well as SE (if not better) and without any of the many disadvantages caused by SE's gamification/rep model. $\endgroup$ – Bill Dubuque Jan 10 '14 at 14:03
  • 18
    $\begingroup$ @BillDubuque I wonder if that was simply due to the smaller user base and higher cost of discovery. The only people who would have read sci.math in the "good old days" would have been computer literate, smarter than average and quite likely employed or studying at a university or other research institute. Such a group of people might be capable of generating good discussion without the need for feedback mechanisms. The same system, exposed to the userbase of the modern internet, is just as likely to generate chaos (which is what seems to be happening at the moment). $\endgroup$ – Chris Taylor Jan 10 '14 at 14:06
  • 9
    $\begingroup$ @Bill Dubuque I first entered the internet one year after "the september that never ended". I liked sci.math and I miss the interactive communication. I dislike the impersonal aspect of the stackexchange model. However, I do not think that it is an accident that this site has become successful and sci.math declined and I am quite frustrated with the quality of the answers on artofproblemsolving.com which has no gaming, no rep, and its presentation is more or less usenet with latex enhancement, so I think that M.SE is the best model available on the web for now. $\endgroup$ – Phira Jan 10 '14 at 17:24
  • 1
    $\begingroup$ As Charlie Munger wrote: Never, ever, think about something else when you should be thinking about the power of incentives. $\endgroup$ – copper.hat Jan 11 '14 at 7:09
  • 4
    $\begingroup$ Specifically for mathematics, the presence of mathjax is, for me, a huge deal compared with text in newsgroups. $\endgroup$ – copper.hat Jan 11 '14 at 7:13
  • 6
    $\begingroup$ See also here $\endgroup$ – Hagen von Eitzen Jan 11 '14 at 15:07
  • 1
    $\begingroup$ @copper.hat Mathjax can be used in any forum with web browser access (including usenet newsgroups like sci.math). If MathJax appeared during the glory days of sci.math then probably it would have spread like wildfire. Instead, MJ did not appear till much later, when s.m was a decaying skeleton. So nowadays there is little motivation to get MJ working nicely on s.m. Further, many of those who remain on s.m have explicitly stated that they prefer their ancient newsgroup browser over modern web browser interfaces. So they wouldn't have access to MJ anyhow. Others resist change or are tech averse. $\endgroup$ – Bill Dubuque Jan 11 '14 at 15:29
  • 1
    $\begingroup$ Exactly this, and in addition I somehow got concerned if a reputation based privilege system is the right way to establish community moderation of high-quality academic online communities after all, on Physics it went wrong :-/ whereas on MathOverflow it happily works so far ... $\endgroup$ – Dilaton Jan 11 '14 at 17:48
  • 4
    $\begingroup$ @Dilaton MathOverflow owes a large part of its success to the actual real-world reputation of its users. $\endgroup$ – Phira Jan 23 '14 at 23:23
12
$\begingroup$

'One of my concrete concerns is that the reputation system might incentivize answering easy or soft questions to get a quick reputation "high", and that it might also promote some group think'

I think your concerns are valid and the high-end users probably feel strongly about this.

I don't know how this can be addressed on a grand scale, but I know getting rid of rep system wouldn't be good. Reputation is not only a number for showing but also it identifies users who made significant contributions to the website and determines who should be allowed moderation privileges. The users keep the site running.

Having said that, as individuals we can

  • upvote more difficult questions and interesting problems

  • be less tempted to vote for questions/answers which are soft or its answer could have been found by googling. (i sometimes even downvote these questions, especially the ones which shows no research efforts)

  • flag the easy questions for closure for 'off topic due to lack of context'

I think the main problem is that we are not doing the last thing often enough and answering these questions in a way encourages more people to do the same. I think there are too many people who uses this site to do their homework without having a go themselves.

I personally have rule of only answering easy questions if they are short (less than 5 mins to type the answer), unanswered and motivated. I think this is fine. In a sense, diversity in level is a nature of this site.

I notice the OP has a relatively low rep right now, I actually encourage you to reputation hunt until you get the basic priveleges if you get certain moderation rights, if you actually care about the site and wants to make a contribution to it in long run - in my opinion, at least in this circumstance, the end justifies the means.

$\endgroup$
  • 2
    $\begingroup$ I'm relatively new to the site, and have enjoyed "reputation hunting." It's given me practice at framing answers that are useful and usually relatively brief, and given me the chance to exercise my skill at guessing what a badly-written question means. At the same time, i've got a couple of older questions marked, and work on them from time to time, in hopes of resurrecting them if I ever get anywhere. This seems to be like a good mix, i.e., reputation is driving my activities in just the right directions. (A couple of downvotes have also taught me not to be too speedy in answering!) $\endgroup$ – John Hughes Jan 10 '14 at 17:34
  • 1
    $\begingroup$ @John, Yeah, reputation does point you somewhat in the right direction, but so does citation count in research. @ Lost 1, my contributions will go down as I go back to class, but they might be improved. I act out of self-interest mostly, and do analysis like this partially because I'm too tired to self-study independence proofs. $\endgroup$ – Jacob Wakem Jan 10 '14 at 18:39
  • 2
    $\begingroup$ Define soft. What you think is soft, a person with little ability at maths but an awareness of its importance who wants to pursue it further may not find it soft. As a university tutor I've had one or two students who clearly spend a lot of time and effort trying to figure things out and mostly continue to struggle no matter what they do. Generally, students in this category also find it hard to articulate just what their problem is. Maths.SE is interactive, Google is not. I'm happy to give such students as much help as they need. $\endgroup$ – Geoff Pointer Jan 11 '14 at 0:55
  • 3
    $\begingroup$ @GeoffPointer i define questions such as 'why is maths useful in the real world?' or 'how many digits of pi we currently know' as soft. I agree with your last point but only to the extent that i think they should have a go at themselves or where they are stuck on. So many people just post what smells like hw, followed by 'thanks in advance' or 'i have no idea what to do'. That is not on. $\endgroup$ – Lost1 Jan 11 '14 at 1:01
  • 2
    $\begingroup$ There was a "debate" between several of the more experienced users and moderators here recently about the difference between Maths.SE and other Whatever.SEs. They were disagreeing about a variety of things but one I was surprised wasn't mentioned was the homework thing. For example, TeX.SE doesn't have to worry too much about this because TeX/LaTeX is a tool not a subject per se, but Maths has a very obvious homework problem. I've been trying my best to distinguish the homework sidesteppers lately. I don't think Maths.SE should just be a site for hints though. It's complicated, for sure. $\endgroup$ – Geoff Pointer Jan 11 '14 at 1:10
  • $\begingroup$ @GeoffPointer i hate people asking and answering this sort of thing. math.stackexchange.com/questions/634274/… i actually think op learnt nothing from thus $\endgroup$ – Lost1 Jan 11 '14 at 1:18
  • $\begingroup$ @GeoffPointer from very limit experience, at least maths at this level is not learnt by seeing a solution. This is basic applying method if you know it, impossible if you don't. At least the op, imo, should state any approach he or she knows and how he or she is strugglinh. $\endgroup$ – Lost1 Jan 11 '14 at 1:18
  • $\begingroup$ @Lost1 This is none the less how you feel about that material. As I said, the students I was referring to had trouble articulating their difficulties and I was only able to help them during a one on one session in a maths consulting centre. $\endgroup$ – Geoff Pointer Jan 11 '14 at 21:37
  • 2
    $\begingroup$ Maybe we should award more reputation for answering the question by user who has a lot reputation. This follow the assumption that users with lot of reputation does not ask easy question which may be googled in 5 minutes. $\endgroup$ – Trismegistos Jan 16 '14 at 10:52
  • $\begingroup$ @Trismegistos i wish, we can just generally award more rep for answering difficult questions, but this is very opinion based and hard to distinguish. $\endgroup$ – Lost1 Jan 16 '14 at 11:03
10
$\begingroup$

Do you feel this policy will encourage you to be a reputation hunter? If you believe that you're above this behaviour - and I certainly believe that I am above it - then maybe giving other people the benefit of the doubt in general is the most consistent way to proceed.

In my short three months here I have found that the reputation system works very well compared with some other places I've been and I haven't noticed any significant negative consequences and it seems pedagogically sound. I particularly like that it's a whole community driven thing and isn't just handed out by moderators.

If everyone that uses this site participates in voting honestly, one way or the other, for what does or does not benefit themselves, or others, in their study of mathematics I'm sure the reputation system will continue to be a largely positive thing.

Addendum 1: I've read the OP again. Bad answers get very quickly voted down. Reputation hunters can make good teachers. You say that a reputation boost makes you feel good. A reputation negative can make you feel bad in equal measure. As a new user here, I enjoyed receiving up votes as confirmation that I was actually helping. Not to steal Git Gud's thunder, who has commented as such already, but bad answers won't earn reputation.

Addendum 2 (16th August 2019): This question has stayed on my mind. Has it been conclusively answered? Who wouldn't want to have a good reputation whatever they do? My original comment that I'm above reputation hunting meant that I believe I'm above seeking a reputation just for the sake of it. Education for me is a passion and I'm a social animal. I desire to make a positive contribution and a positive reputation is feedback that I'm on the right track. A quick look at my answers here will show that I answer plenty of questions that aren't easy.

The point here is to distinguish between desired behaviour and undesired behaviour. The OP was concerned about encouraging people to only answer "easy" questions, but someone has to answer them and they'll only get upvotes if they're useful to someone.

A related problem is the use of down votes. I've just noticed that my answer has attracted a down vote, but what can I learn from that if I don't know why? What concerns me about down votes is the ability to give them anonymously without explanation. Every time I down vote a question, I write a comment explaining why, but there is no obligation here to do so. Surely, if we're a genuinely educational community then we can be open about our criticism?

$\endgroup$
  • 3
    $\begingroup$ Sometimes I feel proud if I get up voted.Maybe its for altruistic reasons though, knowing I am helping, or maybe I'm just excited something is happening (between semesters). Overall, I think math.SE is working very well. Seeing as I posted this in the first place, and the community didn't rep-attack me, it couldn't be too bad. $\endgroup$ – Jacob Wakem Jan 10 '14 at 0:33
  • $\begingroup$ But people think they are influenced less than they really are. Wouldn't you want to optimize the system using what knowledge you can of human behavior? $\endgroup$ – Jacob Wakem Jan 10 '14 at 1:03
  • 4
    $\begingroup$ @JacobWakem Yes, but I'm wary of patronising anyone and applying techniques that I wouldn't welcome myself. I have been in situations as a teacher where experience and professionalism do put you in a different position than your students and various aspects of that relationship are necessarily patronising - but not in the negative interpretation of that word. The largest part of my answer expresses how I feel about the community here in particular. If it works, don't fix it. I am open to ideas that will improve Maths.SE if there is research that makes it clear enough. $\endgroup$ – Geoff Pointer Jan 10 '14 at 1:47
  • $\begingroup$ no you are not above it $\endgroup$ – user88576 Jan 13 '14 at 22:55
  • 1
    $\begingroup$ @KarlKäfer Would you care to elaborate? $\endgroup$ – Geoff Pointer Jan 14 '14 at 3:31
3
$\begingroup$

I don't really see why the fact that the system encourages answering easy questions as well as hard is a bad thing. If this encouragment would not exist, you would have 2 negative consequences:

  • Less answers to easy questions. A lot of newcomer users would give up because their simple question was ignored. These "newbies" could one day prove invaluable to the community.

  • Smaller quality of answers to easy questions. In a way, it's just as hard to answer an easy question because you must be sure not to drown the one asking with too much knowledge (you can't just answer "in book A on page x you see that Sterling's formula holds, therefore you..." if someone clearly does not understand more simple things). At the same time, you must keep a balance of what to assume and what not so you don't end up rewriting whole book chapters. Because of this, good seasoned users are needed on easy questions as well as hard.

$\endgroup$
3
$\begingroup$

Reputation systems are problematic and also ubiquitous on online forums.

Reputation serves two purposes:

  1. reduce spam
  2. increase posting by top users with gamification

The downside is false consensus and fallacious authority. Upvotes do not mean "correct" or "good" but rather "applauded within this context". Funny answers, 'politically approved' answers, first answers, and high-quality answers which were made by someone else but copy-pasted by the upvotee all yield more upvotes than their "hard" value.

The first end could be served with a limited system like what you are proposing. I believe Hacker News has come to downplay the visibility of points because of bad effects on the dialogue.

$\endgroup$
-1
$\begingroup$

In addition to the other answers, to address your problem that

"reputation is prominently displayed", it's because it should be. A user with 10k reputation will likely say something which is true and you can trust. One with 100 is not untrustworthy, of course, but less than the previous.

I agree, though, that often incomplete or superficial answer posted by people with very high reputations will get a lot more upvotes than other better answers posted by recent user, but in the end it's not that big a deal.

$\endgroup$
  • 1
    $\begingroup$ I guess I didn't understand this because I thought it is math and the question asker should ultimately see that the answer is logically necessary. But I guess it makes sense if the asker is talking about something far out of their range. $\endgroup$ – Jacob Wakem Jan 11 '14 at 7:19
  • 12
    $\begingroup$ The danger of reputation driven privilege systems, such as applied by SE, is that power hungry rep hunters who are not knowledgeable about advanced topics of the site at a deep level but very interested in SE policing for example, can take over a site by obtaining rep from answering basic/easy/low-level questions such that it is them who determine the scope and rules of the site, what questions are allowed, etc ... They can start to dominate the review queues and vote to close on questions even though they have no (deep) clue about the topic etc ... $\endgroup$ – Dilaton Jan 11 '14 at 18:37
  • 5
    $\begingroup$ @ Dilation, yeah, a high reputation could mean you have no real math life outside math.se, or it could mean you are somewhat involved and make important contributions (are knowledgeable, careful, etc.). $\endgroup$ – Jacob Wakem Jan 11 '14 at 20:09
  • 1
    $\begingroup$ @Dilation Your point is a very good point and I have noticed a little of this happening here, but not often. If you don't want to add your own answer I could add your point as an addendum to increase the scope of my own answer. $\endgroup$ – Geoff Pointer Jan 28 '14 at 22:29

You must log in to answer this question.

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