17

I see quite often, especially when the problems with the review queue are discussed, that there is a proposed problem with people rep farming with edits. The idea is that these people are making poor edits just for the sake of the points you get when those edits are accepted.

If rep farming is a problem though, why offer the reputation in the first place? Two points of reputation isn't really anything, and in my opinion, the people who care about two points of reputation aren't the type of people who you would want to be making edits in the first place.

Wouldn't it be better to offer no reputation for making edits? The people who care to make the edits probably aren't doing it for the reputation anyways, and I'm fairly certain you don't get anything after you gain the privilege to make instant edits.

Even if the number of edits go down a bit as a result, if the number of rejected edits go down as well, I would think that would be a benefit.

My other thought is maybe to only encourage large edits. That way, people who take the time to fix the bad posts that require say, at least 5% of the post to be changed, would get rewarded. But the people who only change a couple letters would only be doing it because they actually care about fixing the couple typos.

Is rep farming actually a concern, and if it is, wouldn't it make more sense not to offer reputation for making edits?

10
  • 4
    I'm not sure the problem is in the 'rep farming', the problem is with the reviewers that blindly approve suggested edits that shouldn't be approved.
    – JonK
    Sep 9, 2014 at 15:15
  • 6
    @JonK Neither would be a problem without the other. The fact that both exist is what causes problems. A lot of effort has been spent trying to deal with the latter, with only limited success. This is approaching the problem from another angle.
    – Servy
    Sep 9, 2014 at 15:33
  • @Servy I suppose the thinking was that if you make it much less likely to get a bad edit approved people would stop making them to begin with.
    – JonK
    Sep 9, 2014 at 16:08
  • 1
    @JonK Sure, and if you come up with a successful way of doing that, then great. So far there has been major problems with bad reviewers for years, and not just problems here or there, but a very significant portion of the reviewer base doesn't review correctly, resulting in a large percentage of improper review actions. Clearly the attempts to solve the problems by fixing the reviewers hasn't been successful. Would it be better to fix the reviewers, yes, but if we can't, then exploring other options is absolutely worth doing.
    – Servy
    Sep 9, 2014 at 16:13
  • 1
    This is only a problem when people approve edits without reading them first. That means it is going to be the most common in big sites anyway, and on those sites, it takes a lot more rep to get the privileges. If a user does use edits to gain rep, and begins abusing gained privileges, he can be suspended or removed. Peoples rep is a very rough way of viewing their helpfulness on a site. One may have thousands of low voted posts, and have more rep than someone else, who has only a few, well received posts. Rep farming by editing won't necessarily produce worse users than mass answerers.
    – Frank
    Sep 9, 2014 at 20:01
  • Both will have access to privileges they may not be ready for, and both will have a lot of rep, but low usefulness and poor contributions to the site.
    – Frank
    Sep 9, 2014 at 20:03
  • 2
    @J.Musser So the fact that there is some other way of earning rep without being helpful means that it's okay to give people rep for these actions, even when they aren't being helpful?
    – Servy
    Sep 9, 2014 at 21:05
  • @Servy No, it's not okay to 'give' people rep for these actions. I only approve edits that I actually think will improve the site, and I only upvote Q/A's that I think improve the site. The problem isn't the rep farming editors, the problem is the reviewers who are in it for the badges. I think the +2 is a good incentive, but everything has consequences. This network is incentive based, so you should expect a good bit of the problems to be incentive based.
    – Frank
    Sep 9, 2014 at 23:34
  • You could have a site with no intensives for your effort, but those would have their own problems (just look at any forum out there). Get used to it; it isn't a perfect world. There will always be problems. We simply need to do our best, and leave it at that.
    – Frank
    Sep 10, 2014 at 0:04
  • @J.Musser If we could actually fix the fact that the review system completely sucks I expect we'd have done something about it by now. The fact that it's been such a major problem for so long indicates that SE doesn't seem willing/capable of solving the problem, after all, it is a very hard problem. Yes, I'd love it if they worked to deal with the droves of crappy reviewers, but until that happens, this is something we can do to help.
    – Servy
    Sep 10, 2014 at 14:05

4 Answers 4

14

Reputation is the carrot to get users to become more active in the community. There are many ways to do this. Asking good questions and writing good answers are two of the most pronounced ways that this happens.

However, there is more to being part of the community than just the questions and answers. Care taking of the site is key in maintaining the quality that we wish to present to the world. The quality of questions and answers is what sets apart Stack Exchange from Yahoo Answers. That quality has two parts to it - there is the actual material there, but there is also the presentation of that material.

When someone writes a post that is in all lower case, unformatted, full of SMS speak, and mistaking which "its" or "your" to use, that shows a poor image of Stack Exchange to the world. There is far too much of this for Peter and Marc to do by themselves (yes, there are many more... just picking some of those that I often see). Thus, we need to incentivize some people to help out for awhile (and maybe get the "I'll fix things for free" after they get 2k rep or do 1000 edits).

There is no problem with farming reputation from edits if they are good edits. It is improving the quality of the site and making it a better place.

Yes, there are people who go through and try to fix old posts that need more serious moderation to handle it. This isn't a problem with rep-farming edits - it is that more moderation is needed. Some old crap question gets bumped from an edit? Close it and delete it (and that +2 rep disappears). Do this (moderating and cleaning up) enough and people will realize that sometimes those ugly things in the back should get thrown out rather than polished.

People making unnecessary edits - like changing the html from <font color="grey"> to <font color="gray"> because of some mistaken belief that one spelling is correct and the other isn't - those need to be rejected explaining that it doesn't help.

But, if it is an improvement, it is an improvement. It might also help the people asking questions and dumping in unformatted things to see how things should be formatted and what options are available.

Additionally, it is hoped that those people who are helping caretake the site with little edits will go on to be people who help close and delete crap as they gain in reputation and privileges. It is not always the case, but you need to dangle the carrot for them early on to have hope that they'll be the ones cleaning without reward later.

4
  • Second to last paragraph, you make it's vs its mistakes...
    – bjb568
    Sep 9, 2014 at 23:57
  • 7
    @bjb568 I'll give you two rep if you edit it correctly.
    – user213963
    Sep 10, 2014 at 0:00
  • Can't find 6 chars of edit tho.
    – bjb568
    Sep 10, 2014 at 0:04
  • 1
    @bjb568 well, thats annoying. (Turned out I had a few of them when I looked through it)
    – user213963
    Sep 10, 2014 at 0:26
6

We started offering reputation to people who suggest good edits for a single, obvious reason that everyone is in agreement with: to incentivize people to start making good edits and continue contributing positively to the site.

So to me, the only real question is how eliminating the reputation boost would affect users' willingness to continue making/proposing edits. Is the warm fuzzy feeling you get deep inside from knowing that you've helped contribute to making the site a better place enough to encourage people to edit? Or do we need to incentivize it with magical unicorn points?

At first blush, it seems like we don't have any hard data on this. But perhaps we do, from a slightly different angle. Once users have earned 1k reputation from suggested edits, they don't get any more reputation from their edits being approved, and once users have full editing privileges, they don't get diddly-squat. I haven't run the numbers, but I can say qualitatively with a fair bit of confidence that users don't stop editing once they've maxed out on reputation or earned full editing privileges. As such, it seems to me that we do not need to incentivize editing any further by awarding reputation. People will suggest edits anyway.

The only possible counter-point that I can envision is that suggested edits work like training wheels for editors, and giving people this reputation incentive during their early days of editing is critical in cementing their long-term participation. I'm skeptical of that, but I'm not sure how we can prove definitively one way or another. Maybe we could go back and look at the data for editing before suggested edits were introduced and compare that to the levels of participation we have now—do we have more people editing (percentage-wise, of course) now that we've catalyzed their early participation with reputation? Or maybe we could run an experiment, simply removing the reputation incentive and watching to see how much participation is affected. There will probably be a fair amount of whining on Meta, but we're all used to that. The operative question is whether behavior actually changes with regard to making edits that substantially improve the site.

2
  • 2
    "suggested edits work like training wheels for editors" -- this was most definitely the case for me. Including even edit-ban I've got once upon a time, about 3 years ago. As of now I've got full edit privileges and almost full pack of edit badges at all 4-5 sites where I am interested in editing but I am still grateful for lessons learned back then, when my suggestions were reviewed
    – gnat
    Sep 10, 2014 at 9:15
  • As to your last paragraph, a large number of the people that do a lot of editing with edit privileges didn't do much editing at all before they had full editing privileges. They learned how to edit after getting to 2k rep.
    – Servy
    Sep 10, 2014 at 14:09
5

Well, since you asked… I very much DO agree with your premise.

I always thought the beauty of the reputation model was that reputation could only be GIVEN to you by someone else. You cannot "take" reputation through an action of your own, as it is always awarded by the community. The intrinsic value of reputation is preserved when you realize that it can only be earned when others find value in something you have done.

But somewhere along the way we started adding these little mechanical exceptions that say, "if you push button {X}, we'll give you a little reward" — So they push button {X}… a lot. It's a production-line motif that detracts from the intrinsic value of reputation, and I am not a big fan of it.

I don't want to overstate the extent of this problem — I really don't think folks will reach the top ranks of a site simply by approving edits — but the motivation driving these behaviors is that these micro payments do add up.

I dislike any feature that "pays" reputation for an action that isn't explicitly awarded by a community member for a job well done. That's not how the reputation system is supposed to work. Reputation is not a payment for work recieved… and I wish we would stop offering them.

3
  • 1
    I really don't think folks will reach the top ranks of a site simply by approving edits It's worth noting there is a 1k limit of reputation through edits, but 500 bad edits is still a pretty big negative impact on the site. People can and have submitted that many edits, with most being largely harmful or at least not helpful.
    – Servy
    Sep 9, 2014 at 20:58
  • 11
    Actually, I think approval of the edit is the other community members voting to give the 2 rep. Sep 9, 2014 at 21:08
  • 1
    Do you think +2 for accepting an answer is bad?
    – bjb568
    Sep 9, 2014 at 23:50
0

It's all about incentives. Stack Exchange thrives because it has a lot of community members willing to do lots of work. Technically, for free, but there is an incentive through badges and rep to help motivate them. Not all need that level of motivation, but it certainly works as far as getting lots of work done. All processes have problems, and there may be a way to adjust it, but for now remember that the community members approving the edit are voting to give 2 rep to that editor.

8
  • The review queue isn't the concern to me here. I'm proposing that it's the incentive of getting those points that's encouraging people to make edits for the wrong reasons, or simply attracting the wrong people altogether.
    – Thor
    Sep 9, 2014 at 21:23
  • 3
    Well downvotes on questions drive questioners to make their questions better, so negative rep if they fail an edit review might be effective. Sep 9, 2014 at 21:24
  • 1
    I could get behind that idea. Though to be fair, and consistent with the rest of the reputation system, it should probably be something like ony -1 for rejected edits.
    – Thor
    Sep 9, 2014 at 21:30
  • @LanceRoberts The whole premise here is that the bad edits actually get accepted (due to the droves of bad reviewers). If all/most of the bad edits actually got rejected then we wouldn't have a problem as bad editors would just get edit banned, so clearly adding more significant consequences for a rejected edit won't solve the problem.
    – Servy
    Sep 10, 2014 at 14:07
  • @Servy, I think the acceptance/rejection process works just fine to achieve concensus, and those bad edits that make it are outliers. Sep 10, 2014 at 14:08
  • @LanceRoberts While I strongly disagree, if you actually believe that everything is working why propose a change?
    – Servy
    Sep 10, 2014 at 14:22
  • @Servy, to disincentivize bad edits. While they are mostly handled, they take time and energy away that could be used on more productive things. We should incentivize things that make the site more efficient, and disincentivize those things that don't. Sep 10, 2014 at 14:29
  • 2
    @LanceRoberts If bad edits are outliers that almost never happen, then clearly it's already not a problem. If you're worried about having reviewer's time wasted then perhaps you should consider the effects of actively encouraging people to make pointless edits by removing the "too minor" rejection reason that is undoubtably going to result in tons of wasted reviewer time reviewing edits that are adding no value.
    – Servy
    Sep 10, 2014 at 14:32

You must log in to answer this question.

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