What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

One common complaint on MSO about suggested edits is that some users blindly accept all suggested edits, no matter how minor or incorrect they may be. It seems to come up especially frequently in the never-ending "is a minor edit better than no edit at all" debate.

It reminds me of the complaint we had two years ago about how some users would automatically upvote any post with a score of less than zero because they disliked seeing negative numbers or considered downvoting inappropriate. Jeff ended those "pity/sympathy upvote" complaints — more or less — by asking Is there an actual "pity" or "sympathy" upvote problem? and posting some data in an answer.

Can we get similar data on suggested edit approvals? I don't spend that much time on the /review pages, but I rarely see descriptions like

Username approved n edit suggestions, and rejected 0 edit suggestions.

and when I do, it's usually because the reviewer is new (i.e. values of n less than, say, ten).

share|improve this question
15  
Obligatory –  gobernador Jul 16 '12 at 21:51
2  
There is no problem because there is no such thing as a reject button –  random Jul 16 '12 at 21:57
    
Thanks for the assist, @gobernador. –  Pops Jul 16 '12 at 22:00
    
+1 You were paying attention ;) –  Zuul Jul 16 '12 at 22:05
10  
I have been trying around the suggested edit review queue lately and I do find that minor suggested edit constantly get accepted even if I put my reject vote on them. Some user do seems to take advantage of that loophole, since I'm frequently seeing the same user make trivial edit to post and they almost all get accepted. –  HoLyVieR Jul 17 '12 at 0:17
2  
@HoLyVieR if you feel that suggestion is worth some extra-effort to "lock" your rejection, an option is to hit "improve" and uncheck "helpful" box. In that case, one better really makes sure that post has been improved (or left as-is if it looks already good) –  gnat Jul 17 '12 at 7:02
9  
I'm not good at analysis, but this query might help: Suggested edit acceptance percentage, by user. It looks like ~36% of reviewers have >=90% acceptance percentage (they accept ALL / most of the edits). However, the average acceptance percentage among the top 100 reviewers (by quantity) is ~78%. (This is all with the filter set at >=25 suggested edits voted on) –  jadarnel27 Jul 17 '12 at 13:10
16  
We just had a 1-rep spammer insert links for their commercial project as suggested edits to answers, and two of those were approved by high-rep users. The links were just pasted at the bottom of the edit to the short answers, so I don't think the reviewers even looked at the edits. That's anecdotal evidence, but it might a symptom of a larger problem. –  Brad Larson Jul 17 '12 at 15:40
3  
We've had the same problem on SF; though most of ours seem to come from lower rep users approving anything/everything. @BradLarson I've seen the same thing happen; absolutely blatant astroturfing that gets approved. –  Chris S Aug 3 '12 at 16:52
8  
Since the new review interface was introduced, SO's suggested edit queue is no longer quasi-systematically full. Evidently more people are participating, and I am grateful for that. However, I have the informal impression that the number of incredibly bad edits that were approved by the time I clicked to reject them also increased sharply. I would love to see some hard data about this, but it's difficult: someone has to authoritatively decide which edits are good and which are bad. –  Gilles Aug 3 '12 at 18:22
4  
I just saw an edit approved that used bold instead of code blocks, approved before I could reject. So I went there and edited myself. –  Lee Louviere Aug 3 '12 at 18:43
1  
I've mentioned this in a few other related questions as well, but the number of tag wikis containing plagiarism that have been accepted by 8K, 9K and 10K+ users is appalling. It's not that hard to quickly Google the text. –  LittleBobbyTables Aug 3 '12 at 18:51
    
I've seen a few times that totally invalid edits being approved. Even from 4k+ users. –  Nikola K. Aug 3 '12 at 18:52
2  
Out of curiosity, I just went and had a look at the first page of tags on SO. The #3 overall tag, [php], has a tag wiki with a plagiarized excerpt and first paragraph. (See revision 1 of the wiki and this web page published in 2006.) –  Pops Aug 3 '12 at 19:08
5  
Yet another anecdote: Edits that add nothing but random formatting –  Gilles Aug 7 '12 at 19:51

1 Answer 1

up vote 13 down vote accepted
+500

Since all suggested edit approval/reject are public they are also in the Data.Se in the table SuggestedEditVotes.

I figured that if there's a problem with rubber stamping edits it would show when people are accepting/rejecting edits faster that the allowable time to upvote two comments (5 seconds).

So I wrote a query that compares each vote with the one prior. Then I counted the number of times per user they voted twice within 5 seconds.

I've apparently done it 5 times and you've done it 29 times.

Here are the top 32.

User                        UserId  # Of times 2 votes were separated by 5 seconds
Book Of Zeus                905093  2253 
marc_s                      13302   1535 
Oded                        1583    1490 
Haim Evgi                   74314   1253 
Zsolt Botykai               11621   1183 
JB King                     8745    1000 
warren                      4418    995  
Lucifer                     996493  943  
abatishchev                 41956   925  
Bill the Lizard             1288    843  
Uwe Keim                    107625  831  
DVK                         119280  781  
Cody Gray                   366904  725  
pratap k                    899271  670  
Gilles                      387076  636  
Nick Craver                 13249   613  
mKorbel                     714968  530  
Matt                        444991  513  
animuson                    246246  506  
Sean Owen                   64174   471  
balexandre                  28004   469  
Davide Piras                559144  467  
Abdel Olakara               185655  458  
yojimbo87                   217288  447  
???                         596219  402  
bpeterson76                 364708  395  
Alex K.                     246342  379  
OMG Ponies                  135152  366  
Srikar                      147019  356  
0A0D                        195488  346  
Stecya                      606586  339  
Jeff Atwood                 1       328  

Yikes that's a lot of familiar names and a lot of very fast approval/rejections.

Using a later revision of this query I was able to use look for examples where a user did a large # of these quick votes in a single day.

Using that as input for this query through sheer trial an error able to find these Edit votes.

41 Accepts between 2012-03-01 00:29:13 and 2012-03-01 00:30:36. Or 1 accept for every 1.8 seconds.

   userid VoteTime            suggestededitid Vote
-- ------ ------------------- --------------- ----- 
1  905093 2012-03-01 00:29:13 211813          UpMod 
2  905093 2012-03-01 00:29:14 211812          UpMod 
3  905093 2012-03-01 00:29:16 211809          UpMod 
4  905093 2012-03-01 00:29:16 211808          UpMod 
5  905093 2012-03-01 00:29:17 211807          UpMod 
6  905093 2012-03-01 00:29:19 211796          UpMod 
7  905093 2012-03-01 00:29:25 211855          UpMod 
8  905093 2012-03-01 00:29:26 211854          UpMod 
9  905093 2012-03-01 00:29:28 211850          UpMod 
10 905093 2012-03-01 00:29:31 211846          UpMod 
11 905093 2012-03-01 00:29:36 211844          UpMod 
12 905093 2012-03-01 00:29:37 211843          UpMod 
13 905093 2012-03-01 00:29:39 211841          UpMod 
14 905093 2012-03-01 00:29:41 211840          UpMod 
15 905093 2012-03-01 00:29:44 211836          UpMod 
16 905093 2012-03-01 00:29:46 211839          UpMod 
17 905093 2012-03-01 00:29:46 211835          UpMod 
18 905093 2012-03-01 00:29:49 211834          UpMod 
19 905093 2012-03-01 00:29:50 211833          UpMod 
20 905093 2012-03-01 00:29:52 211832          UpMod 
21 905093 2012-03-01 00:29:55 211828          UpMod 
22 905093 2012-03-01 00:29:57 211826          UpMod 
23 905093 2012-03-01 00:29:59 211788          UpMod 
24 905093 2012-03-01 00:30:00 211792          UpMod 
25 905093 2012-03-01 00:30:01 211793          UpMod 
26 905093 2012-03-01 00:30:03 211785          UpMod 
27 905093 2012-03-01 00:30:06 211777          UpMod 
28 905093 2012-03-01 00:30:09 211687          UpMod 
29 905093 2012-03-01 00:30:23 211665          UpMod 
30 905093 2012-03-01 00:30:24 211663          UpMod 
31 905093 2012-03-01 00:30:25 211662          UpMod 
32 905093 2012-03-01 00:30:27 211766          UpMod 
33 905093 2012-03-01 00:30:28 211746          UpMod 
34 905093 2012-03-01 00:30:29 211623          UpMod 
35 905093 2012-03-01 00:30:30 211688          UpMod 
36 905093 2012-03-01 00:30:31 211667          UpMod 
37 905093 2012-03-01 00:30:31 211790          UpMod 
38 905093 2012-03-01 00:30:33 211616          UpMod 
39 905093 2012-03-01 00:30:34 211615          UpMod 
40 905093 2012-03-01 00:30:35 211872          UpMod 
41 905093 2012-03-01 00:30:36 211871          UpMod 

I'm sure you could find other examples or examples where for a given time period users rubber stamped all available suggested edits (they're in the table too).

share|improve this answer
1  
Interesting indeed! Although for that first table, I wonder how many were times where someone checked out an entire page of 30 suggestions -- or whatever the number used to be -- and didn't click the judgment buttons until after reading them all (or at least a screenful). –  Pops Aug 3 '12 at 21:27
    
Certainly a possibility, especially if you've got 3 27" inch monitors –  Some Helpful Commenter Aug 3 '12 at 21:31
    
I don't suppose there's a way to split out acceptance votes from rejection ones, is there? I know that I've rapid-fired rejections when I saw a stream of spam or vandalism from anonymous or 1-rep users. We don't really have a problem with excessive rejection, it's the approvals that are the issue. –  Brad Larson Aug 7 '12 at 19:29
    
@BradLarson Here you go. If you were thinking of a different look let me know –  Some Helpful Commenter Aug 7 '12 at 19:36
    
@SomeHelpfulCommenter - Weird. Unless I'm reading that wrong, there are a lot fewer rejections than I'd expect on that list. I'm pretty sure people on that list like Bill aren't just approving things willy-nilly. –  Brad Larson Aug 7 '12 at 19:42
1  
@BradLarson well here's the day where Bill made the most judgments. A lot of them were related to the great tag clean up question | 10K only link | I suspect many of the people on this list may be here because of mass edit activites –  Some Helpful Commenter Aug 7 '12 at 19:59
    
@BradLarson I totally messed up the link to "here's the day" here it is corrected –  Some Helpful Commenter Aug 7 '12 at 20:57
1  
Wow. Surprised to see my name so high up there. Though I did use to go on edit approval/rejection blitzes :) –  Oded Aug 9 '12 at 20:41
    
The first table is extremely misleading; it needs to also have the total number of votes in another column. –  Matt Fenwick Dec 17 '12 at 16:32
    
@Matt I don't understand why that matters, or why that would help to answer the question. However you should feel free to modify the query and add your own answer. –  Some Helpful Commenter Dec 17 '12 at 17:13
    
@Matt here's the query with the total, but I still fail to see how that matters. People aren't robots even a snack can change behavior –  Some Helpful Commenter Dec 18 '12 at 16:35
    
Your original first query answered "how many times did a user speed-vote?" This number is misleading without also answering "what percentage of the time did a user speed-vote?". The answer to Q1 would say that Bill, abatishchev, and Gilles look similar, when the answer to Q2 shows that, in fact, their speed-voting behavior is very different. –  Matt Fenwick Dec 18 '12 at 18:47
    
I scan and read very fast so of course my approvals are going to be fast. It's not indicative of my effectiveness, I rarely fail audits. –  staticx Mar 27 '13 at 13:04
    
@0A0D Take a look at this Looking at the time stamps alone it really looks like you were trying to get your vote in before the edit was approved. This doesn't mean you always do it. This just happens to be your fastest day –  Some Helpful Commenter Mar 28 '13 at 15:01
    
@SomeHelpfulCommenter: I am sorry, but I read what you just wrote several times and I am not following you. Please clarify. –  staticx Mar 28 '13 at 15:21

You must log in to answer this question.

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