What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 130 Stack Exchange communities.

I find a lot of suggested edits pretty poor. For me, it feels like that more than a half of the suggested edits doesn't properly address all issues in the post. The trigger of today was this particular example. It does not remove the PLEASE HELP ME screaming of the OP and even more, it introduces 2 bad edits:

  1. using code tags to highlight framework/library name and
  2. using <br> tag to introduce a paragraph. There are many, many things which can be done better. It has been properly rejected as "too minor" once, but it is after all approved twice.

I would like to see kind of a mechanism to "downvote" an approved suggested edit. This should affect both the person who suggested the edit and the persons who approved it anyway. In general, I find that both the edit suggestors and suggested edit reviewers are doing a poor job in such case. Suggested edits shouldn't yield reps too easily and the reviewers should review them more carefully.

share|improve this question
16  
That is a p*sspoor suggested edit indeed. I almost completely stopped reviewing them because edits like this got accepted no matter what I did. That being said, I believe there was a similar proposal a while ago. Perhaps in a somewhat different format. Let me search. –  Bart Sep 24 '12 at 16:50
    
This is the feature-request I was referring to. Somewhat different, but with a similar basis. –  Bart Sep 24 '12 at 16:51
11  
The problem really isn't bad edits, it's the bad reviewers. If the bad edits were actually rejected, instead of accepted, then the person suggesting edits would be told that they make a bad edit (and why) and will hopefully stop suggesting those types of edits. If they continue submitting bad edits anyway then they would be blocked once enough are rejected. Of course that entire system breaks down when the reviewers just accept everything without reading it so that their review stats go up. –  Servy Sep 24 '12 at 17:34
12  
5  
I once had the pleasure of using my mod powers to reject a suggested edit with exactly one vote. Guess what? My satisfaction was short-lived; the user submitted the same suggested edit later, and it was approved by two other users. –  BoltClock's a Unicorn Sep 24 '12 at 18:00
    
@Bolt: that's a huge demotivator :( –  BalusC Sep 24 '12 at 18:06
    
Hm, I would have sworn this was a dupe, but I can't find it. And BoltClock found the good related links while I was searching. The silver lining is that suggested edits rarely make the site significantly worse. Bad suggestions to mediocre posts don't create atrocious posts; they result in posts that are still mediocre, just in a different way. Bad suggestions to great posts don't seem to happen all that much, and the ones that get approved tend to get fixed fairly quickly because great questions attract good users who know when to use "rollback." (That said, I'm all for fixing the system.) –  Pops Sep 24 '12 at 18:16
1  
@Popular: true, we can fix or rollback a bad edit, but the edit suggestor still keeps the undeserved +2 rep and the reviewers still have no idea they were doing their job completely wrong. I can't imagine that the system is designed to work like that. –  BalusC Sep 24 '12 at 18:19
    
I'm with you. I was pointing out a silver lining only. I'd love to see this get fixed somehow. –  Pops Sep 24 '12 at 18:20
    
Yeah me too...especially considering I was one of the approvers :$ I cringed right after I hit approve because I missed the cry for attention at the end of the original post. –  j08691 Sep 24 '12 at 19:54
    
@j08691: Thank you for your honesty. I'm wondering, how are you reviewing the suggested edit queue? Is it in all hassle, because it's "too easy" or because you're under kind of a pressure? –  BalusC Sep 24 '12 at 19:59
    
@Chichiray - Are you asking how I review the queue or why I review it? –  j08691 Sep 24 '12 at 20:05
    
@j08691: The "how". I review them very carefully and only when I am boring and have really a plenty of time left. Personally, if I'm short in time or are in a hurry, then I wouldn't bother to enter the /review at all. I'd imagine that when someone still does that, the chance in making review mistakes is much bigger. –  BalusC Sep 24 '12 at 20:09
1  
@Chichiray - Ah, no I never hit the queue if I'm in a rush. In fact, like you I usually visit it only when I have nothing else to do. And I don't do it for points either. I think that ship sailed a while ago. We all make mistakes -- thanks for highlighting mine today though :P –  j08691 Sep 24 '12 at 20:11
    
I would totally love that feature for this one: stackoverflow.com/review/suggested-edits/1050706. Yes, I could have improved it, but the number of corrections required while typing on my phone would have taken all evening. –  LittleBobbyTables Nov 21 '12 at 3:59

8 Answers 8

I think this is one of the more serious problems with the site right now. There are so many bad edits in the system, but what's worse is that nobody is rejecting them. There are many more bad reviewers than good reviewers. People are able to ride this system to 2k rep by spamming reviews and then they graduate from bad editor to bad reviewer, and the circle repeats.

I would love to see either some form of downvoting or flagging of bad edits and (though I can't imagine how) bad reviews. The problem is that, when I see somebody doing an awful job of reviewing edits, I have no recourse. I've gone so far as to find people's Twitter handle in their profile and ping them on Twitter to ask them to please stop allowing awful edits through.

I also find myself racing to "improve" an edit, just so that I can uncheck the "This edit was helpful" checkbox and deprive the editor of the small amount of rep they're after, before the "I ACCEPT EVERY EDIT" train comes through and auto-approves everything.

share|improve this answer
    
What minimum rep do you think a good reviewer should have? 10K? Further, you can use @nickname comment reply in the question as well to notify an editor. –  BalusC Sep 24 '12 at 17:53
11  
I'm worried about the larger issue of who should be able to review edits, and being able to contact reviewers, not editors. There is nowhere where you can @name a reviewer of a suggested edit, except for finding a recent question/answer and doing it there, which is an incredibly ugly solution. I would love some way of flagging a review as wrong/bad, and maybe accruing enough (moderator-supported) flags would make somebody unable to review further edits. –  meagar Sep 24 '12 at 17:54
    
You're completely right. –  BalusC Sep 24 '12 at 17:55
2  
Sometimes when I get annoyed, I chide the reviewers in my edit summary when I roll back or edit over an approved suggested edit. That doesn't notify them either of course. Just... ugh. –  BoltClock's a Unicorn Sep 24 '12 at 17:59
1  
Minor nitpick: you can earn a max of 1k rep from edit suggestions, not 2k. You're probably thinking of how you can't earn rep at all from suggesting edits to regular posts after reaching 2k because your edits are automatically accepted at that point. (You can still suggest edits to tag wikis and their excerpts until 20k.) –  Pops Sep 24 '12 at 18:04
1  
Also, you sound like me at the end, with the racing-to-improve bit. –  Pops Sep 24 '12 at 18:19
1  
"I ACCEPT EVERY EDIT" train comes through and auto-approves everything so true: The robo-approvers are killing my will to review edits "...people keep approving crap while I'm looking at it... a lot of people are stupidly approving and actually driving me away from the queue." –  gnat Nov 27 '12 at 9:23
    
Farming reputation on suggested edit? Why? They just look for active question, fork one of the answers and hope someone won't note a time difference and upvote it too. –  РСТȢѸФХѾЦЧШЩЪЫЬѢѤЮѦѪѨѬѠѺѮѰѲѴ Jul 4 '13 at 12:06

Don't get me wrong, I like the review system. But I think it can be even better.

The problem with bad reviewers is that there is no automatic way to stop a person from reviewing. The only tool a bad reviewer has to improve himself, is looking at the results of other reviewers. And of course common sense.

In my opinion the system is a bit in favor of accepting:

  1. Accepting is easy, it only takes a single click, no thinking required. And twenty days of fifty clicks gives you a gold badge.
  2. Reject votes are not visible unless you click on "reject". So there is no visible sign that an edit is suspect.
  3. You get badges and your name is shown in a top 20 list if you have enough clicks that day.

Some improvements could be:

  1. In order to be accepted, each reject vote should be countered by two approve votes. So if normally three votes are required to accept an edit, it needs five if it has a single reject vote.
  2. There should be a visible clue if a reject vote has been cast.

Ideally there should be a mechanism to stop bad reviewers. But the problem is that this requires a review of the reviewers. Maybe a way to flag accepted edits that are wrong, and a certain amount of flags stops the person from reviewing for a time. But that mechanism is vulnerable for misuse.

Another problem is minor edits. Often an edit is good, but it is too minor (one typo, one tag). So maybe we need an accept as minor. Which preserves the edit, but the suggester gets no reward other than the eternal fame.

share|improve this answer
3  
8  
I'd like the 'accept as minor' option because I often notice "lose" vs "loose" errors and the like. For things like that, I don't care about the rep, I care about not having to read the incorrect English -- I'm finding that the continual exposure to the error is affecting my writing. I want a "minor edit in self defence"! –  retracile Sep 24 '12 at 21:43
1  
Yes, I am pretty particular about my own English and it is kind of painful to see one or two characters off that do not satisfy the minimum edit criteria. I have usually looked for - and found - something else to pass the 6-character limit though... that said, I would personally not mind not getting any rep for edits under, say, 20 characters or something. Up to the reviewers. –  icedwater Jul 6 '13 at 5:42

I was (maybe still am) someone who was not the best editor. I would see a mistake, go in and correct it, but I wouldn't be as thorough in reviewing the rest of the post as I should have been, so other mistakes would slip by. I've been reading MSO a lot more recently and am starting to learn the error of my ways (and have tried to make better edits in the recent past, although I probably still have a ways to go).

As I do not have 2K rep yet, I can't speak for how to fix the reviewers, but I think part of the problem with the editors is the fact that there is no real notification of a rejected edit. It just disappears unless you start looking through your edit history to see what happened. I have had 5 rejected edits and I don't recall getting any kind of notification that the edit was rejected or feedback beyond "too minor". I think most editors have the best intentions, but do not know they should be doing more. What is needed is some way to provide (anonymous) constructive criticism as to why an edit was rejected (or even if it was accepted by was borderline). A rejected edit should appear in the inbox and maybe offer more than a one-liner as feedback. Also maybe to help editors get off on the right foot is to require mod approval of edits from users with less than 10 approved edits? At least this way, the user should get more constructive feedback initially and try to steer them in the correct direction.

Another problem I've found in the process is somewhat of a confusion conflict of "rules". One one hand, edits are supposed to "substantially improve the post" but on the other hand, we have MSO discussions like this one that say titles are not supposed to include tag data, or this one that says posts are not supposed to include "thanks" or "please help me" or other extraneous remarks. But if that is the only thing that is wrong with the post, the edit might be rejected as too minor. So what is the line between cleaning up a question and too minor of an edit to be approved?

share|improve this answer
1  
The main issue when it comes to "too minor" is, are you cleaning up all, or almost all, of the problems in the post, or are you just cleaning up 1-2 and leaving lots more. If you are fixing most of the problems then it really shouldn't be rejected as "too minor". –  Servy Sep 24 '12 at 19:04
    
@Servy I agree.... but some reviewers only see very minor changes without looking at the entire post, so they get rejected. –  psubsee2003 Sep 24 '12 at 19:25
1  
Well, I see way too many minor posts being accepted, then minor posts that fix all of the issues being rejected. –  Servy Sep 24 '12 at 19:34
4  
@Servy: To be honest, I never cared much for that particular rule, the "you didn't fix everything, so I won't let you fix anything." Since I have free editing privs, I generally fix what I feel like fixing with my edits. I don't feel that the standards should apply so differently for people who don't have free editing privs. If someone sees a typo and edits it to fix it, why should that be rejected because they didn't scrutinize the post to fix everything that <insert reveiwer here> thought was wrong with it? That means waiting longer for someone to come along and fix the typo. –  Nicol Bolas Sep 24 '12 at 20:14
1  
@NicolBolas Well, it's been discussed in a few million meta questions, so I won't go into detail. In short, it's mostly because the post needs to be reviewed, and it wastes valuable community resources to spend lots of time reviewing suggestions for a single typo. The community as a whole benefits more if you spend your time making more substantive edits. Additionally you really shouldn't be making lots of minor edits even with 2k rep. You are still encouraged by the FAQ to make substantive edits. –  Servy Sep 24 '12 at 20:25
1  
@Nicol - There must be some limits, otherwise we end up like this stackoverflow.com/review-beta/suggested-edits/687660#./… where we get one misspelled word fixed, and two capitalized Android's, but everything else is just terrible. And this was of course approved. :-) –  Bo Persson Sep 25 '12 at 7:57
    
@BoPersson that's why I like retracile's comment to one of the other answers. If there was a way to approve a minor edit that would approve the edit but would not award rep, it would help. This doesn't apply to your example but I've seen plenty were the question is so awful, you can't figure out what the OP was trying to say. So you fix what you can and hope someone else can figure out the OP's question. –  psubsee2003 Sep 25 '12 at 8:03

I support this because it's a way to get the attention of edit suggestors and suggestion reviewers. However, that alone is not enough. Once we have their attention, we need to direct them to the guidelines for editing and reviewing. That way, they learn how to do things correctly, and just as importantly, the rationale behind correct behavior.

If we don't take the opportunity to teach, then we'll just be treating symptoms, not causes. People might understand that they're doing something wrong, but they won't understand what/how, so they won't improve, and we'll be stuck in a never-ending cycle of meta-moderation that takes up time and effort but doesn't solve the problem.

Also, there's one potential significant flaw with using downvotes here: users can't fix mistakes in suggested edits. Regular downvotes are attached to posts, and if the posts get improved through editing, the downvotes can be removed. There's no way to change a suggested edit after it's been judged. At most, I would support removing the +2 rep reward for having a suggestion approved if a "downvote" is cast. Tacking on a permanent/irreversible -2 seems too harsh.

(In fairness to the OP, it's possible that this is what he meant all along. The definition of "'downvote'" in the question was left ambiguous.)

share|improve this answer
2  
Removing the +2 is perfectly fine. The problem is not the suggestor, but the approvers. I could live with for example a -1 penalty for both, or at least some kind of mechanism for them being blocked from reviewing suggested edits for a while. –  BalusC Sep 24 '12 at 20:30

One suggestion would be to increase the reputation needed to review edits. But that would burden the already-heavily-"privileged" users with even more work.

If users with an "accept-everything" mentality are banned from doing so for a while, then, once again, the people who are doing a good job will get more work to do.

For free. Which can get old pretty quickly.

What about some swag? T-shirts, mugs, anything? What about a point system that has some actual trade-in value for something besides more advertising for a particular question? Like trade-in for swag?

Wikipedia can claim non-profit reasons for thing, but Stack Exchange, Inc., is very, very much for profit. Am I the only one that can see the Invisible Hand here? (Oh, wait ...)

share|improve this answer

I see vote down requires 125 points, and vote up requires 15. Perhaps have a similar rep requirement: 2k to reject, and 4k to accept. That way people get used to actually rejecting edits before they have the power to approve.

Also, remove badges relating to reviewing. I would prefer people who review because they want to, rather than those that review just because they want another badge

share|improve this answer
    
today mindless approvers would just as easily convert into mindless rejectors if reject button would be the only one they've got (as you suggest), wouldn't they? –  gnat Nov 20 '12 at 17:26
    
@gnat, that's why I also suggest removing the badges. If they have nothing to gain, then the mindless would not bother. (at least, that's my take on it - see the edits people do to get the archeology badge) –  Sean Cheshire Nov 20 '12 at 17:29
    
I see. Removing badges is quite popular idea... –  gnat Nov 20 '12 at 17:35

As many have pointed out a bigger issue then bad editors is bad reviewers. A big part of the problem is that reputation by itself doesn't make someone a good reviewer.

As a though what about if in addition to using a minimum amount of rep to enable the review queue, we also required a minimum flag weight. The advantage of using the flag weight is that it is, I think, a better indicator that a user has some positive experience actually doing some reviewing and that they have at least paid some attention to improving the quality of the content on the site.

Perhaps a slightly more complicated version of the above would involve having bad reviews contribute towards your flag weight, with bad reviews being determined by moderators as many have pointed out that you can't rely on what the majority of reviews on a particular review say, since the problem is that the majority is very often made up of poor reviewers.

I don't think this would by any chance eliminate the issue, but it is perhaps a (small?) change that can signification improve the situation.

share|improve this answer

There's this website called newgrounds.com has a system where people can vote on submissions and get points for it.

One thing they do, is they give you additional points if you've voted the same way that others have.

I think that we can take a similar approach to the suggestion reviewing issue.

Only award the point toward the badge, if the reviewer had approved an overall approved post or if the reviewer had rejected an overall rejected edit.

Now this might lead to the obvious problem of approvals being more likely than rejections regardless, but this can be mitigated by splitting the badge into 2 badges. One for voting to approve edits that have been approved, and one for voting to reject edits that end up getting rejected.

With such a system in place, people will have more of an incentive to be correct in their reviewing. rather than just choosing something.

share|improve this answer
2  
Similar concepts have been suggested a number of times, and they all have the same flaw. It's not just one or two percent of reviewers that are bad, a significant percentage of reviewers are bad (i.e. just approve everything, or close to it). This means that many items will be approved because all 2-3 of the people reviewing it just rubber-stamped it, and any good reviewers who reject it would be the ones punished, not the poor reviewers. –  Servy Nov 21 '12 at 16:50
    
@Servy thus, the splitting the badge in 2 –  Sam I am Nov 21 '12 at 16:54
    
The point is that people will still be voting just to get the badge, not by reading the content to see what the proper course of action is. You'll either get people just start rejecting everything, or just rejecting anything without accept votes to try to get that badge, even if the content shouldn't be rejected. At the end of the day it doesn't encourage anyone to actually read the content. –  Servy Nov 21 '12 at 16:56
    
@Servy you can review at max 20 edits per day. There are 2 badges people will be seeking. one for accepting and one for rejecting. If a badge farmer ignores one of them and just auto-accepts or auto-rejects, he will get substantially less than 20 ticks on the day. Splitting the badges is different from just negating it. they'll both exist at the same time. –  Sam I am Nov 21 '12 at 17:01
1  
I suspect most people will accept all until they get the one badge along with the X reviews badges, and then switch to rejecting all to get the other. Or, as I said, they'll just do exactly what every previous voter did rather than actually thinking about what should be done. That's maybe not quite as bad as just approving everything, but it's still undesirable as it's not solving the root problem of reviewers voting without actually reading the content. –  Servy Nov 21 '12 at 17:04
    
@Servy it will take them twice as long (as in twice as many days) to do that, than if they were able to vote right every time. –  Sam I am Nov 21 '12 at 17:05
    
And yet in terms of minutes spent reviewing it will be many, many hours less spent than actually reading the content and doing what's appropriate, so I still think it won't do much to get people to put that kind of time in. Reviewing well is hard, pressing a button without reading is very easy. –  Servy Nov 21 '12 at 17:07

You must log in to answer this question.

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