Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 158 Stack Exchange communities.

What is meta?
Here's how it works:
  1. Any Stack Exchange user can ask a question
  2. The community provides support, votes on ideas, and reports bugs
  3. Your voice helps shape the way Stack Exchange operates

Rejecting edits from anonymous users adding "plz to help of working things" is getting a bit tiresome. I haven't had to do it a lot, but I don't believe I've ever seen an anonymous edit that was worth anyone's time, which makes me wonder if this feature has any benefit.

What proportion of anonymous edits are accepted? Is this proportion enough to outweigh the crappy suggestions that people have to review?

Edit: It appears that 13/66 have been approved (3 of those improved) on Android, though that includes a few that should not have been approved. Possibly my view is skewed by this and other sites have more success with them, especially given my small sample size. I'd love to be proven wrong here :P. (Another even smaller sample: 3/6 on Music.)

share|improve this question
I'd be shocked if the data show this feature to be of any real value. – Adam Rackis Jan 31 '12 at 17:15
Further, intelligently judging whether this feature is useful will be hard. Even if stats show that, say, 50% of suggested edits are accepted, that still won't show whether the quality of these edits justifies the time wasted in tossing out all the junk. – Adam Rackis Jan 31 '12 at 17:32
Maybe if you helped them more they wouldn't have to edit the question! These poor help vampires need your help! – user7116 Jan 31 '12 at 17:32
@six - funny, but the help vampires are registered. You have to register to ask now, so these anonymous edits really are just random internet passersby who show up from Google and then click the edit link. – Adam Rackis Jan 31 '12 at 17:34
@adam only on Stack Overflow due to extreme volume – Jeff Atwood Jan 31 '12 at 17:44
@Jeff - I didn't know that - thanks. – Adam Rackis Jan 31 '12 at 17:44
Data Explorer reports 3932 accepted on StackOverflow and 2305 rejected (SELECT COUNT (*) FROM SuggestedEdits WHERE RejectionDate IS NULL AND OwnerUserId IS NULL; / SELECT COUNT (*) FROM SuggestedEdits WHERE RejectionDate IS NOT NULL AND OwnerUserId IS NULL;) – Matt Jan 31 '12 at 18:48
@Matt Ah, excellent. That's not too awful. – Matthew Read Jan 31 '12 at 18:51
@John That raises another issue I've been thinking about -- if that was a registered user, the edit would have been rejected and they'd be told to post a separate answer. Similarly you might accept a too-minor edit from an anon because you can't educate them about making extensive improvements to posts (unless you don't want to bump the post). That latter case could even be used by registered users -- just sign out first and your minor edits get approved! I find it interesting, anyways. – Matthew Read Feb 1 '12 at 0:55
up vote 12 down vote accepted

You can see the numbers using the following query on Data Explorer:

For all time:

Approved Rejected Good    Site Name 
-------- -------- ------- --------- 
2635     4568     36.58%  StackOverflow
351      330      51.54%  SuperUser 
167      149      52.85%  ServerFault
7        12       36.84%  Cooking   
12       5        70.59%  Game Developers
173      213      44.82%  Gaming    
11       5        68.75%  GIS       
126      50       71.59%  Mathematics
13       6        68.42%  Photography
24       15       61.54%  Statistical Analysis
28       16       63.64%  Web Apps  
19       19       50.00%  Webmasters
125      48       72.25%  Apple     
76       5        93.83%  Theoretical Computer Science
52       74       41.27%  English Language and Usage
31       9        77.50%  TeX - LaTeX
244      97       71.55%  Ubuntu    
17       3        85.00%  User Interface
28       22       56.00%  Unix and Linux
32       22       59.26%  WordPress 
77       53       59.23%  Programmers
12       52       18.75%  Android Enthusiasts
19       8        70.37%  OnStartups
25       23       52.08%  Physics   
15       9        62.50%  IT Security
19       5        79.17%  Electronics and Robotics
17       8        68.00%  Science Fiction
11       8        57.89%  Code Review
8        8        50.00%  Code Golf 
28       10       73.68%  Skeptics  
40       14       74.07%  Drupal Answers
8        8        50.00%  Parenting 
24       8        75.00%  SharePoint
9        5        64.29%  Jewish Life and Learning
39       2        95.12%  Japanese Language and Usage

(edited to remove any sites with ~15 or less anonymous suggested edits, because sample is way too small to even look at)

So yeah, we get a high proportion of crappy edits, however also get a significant amount of good edits.

Anonymous sucks as an editor, she does not edit as much as we initially thought and has a high miss rate.

However, dealing with rejected edits is trivial, we ban anon really early so she can not wreak too much havoc, and she does help out quite frequently.

Overall, I see no reason to make any changes here.

share|improve this answer
Great, thanks. 44% accepted overall. TCS and JLU seem to have it good :P – Matthew Read Jan 31 '12 at 23:20
Blech. Android is way down near the bottom. From my own experience I'd say that matches. I think I maybe have accepted 1 anonymous edit suggestion. Rejected plenty of others. – Al E. Feb 1 '12 at 3:22
@AlEverett I looked through the issue on android, lots of "I do not know how to log in, so will edit my own stuff as anon" issues. I did not see much malice going through the stats. – waffles Feb 1 '12 at 3:31
@waffles: No, I wasn't suggesting that at all. "Never attribute to malice what can adequately be explained by stupidity." – Al E. Feb 1 '12 at 3:32
@AlEverett my only slight concern is that some information is lost with the rejections (which are valid)... but it really is only drops in the ocean – waffles Feb 1 '12 at 3:36
Well, now I feel sort of dumb for having made this just now, but it's certainly been an education in SQL. – a cat Feb 1 '12 at 20:06
Could you or someone else please update this data? – curiousdannii Apr 14 '15 at 3:09

Well, here's some stats from Gaming (apparently we get a lot of views per question):

  • 387 anonymous edits total

  • 173 approved (19 of those improved)

  • 214 rejected.

So ~45% were helpful and ~40% were approved as-is. That's not abysmal, but really a healthy portion of it is just vandalism or an attempt to comment.

share|improve this answer
Thanks. Where can one view these stats? I can see the day-to-day total on the admin analytics page of my sites but not the number approved/rejected. – Matthew Read Jan 31 '12 at 17:39
@MatthewRead Possibly diamonds have a more convenient way, but I just used the suggested edit stats page available to 10k users from the queue (tab: anonymous, timespan: all). They're sorted by whether they were approved/rejected, and there's 50 per page, so I just counted them up (Ctrl+F "Community♦" for improved edits). – a cat Jan 31 '12 at 17:43
I expect it'll be a bit worse on gaming due to the subject matter (entertainment) – Jeff Atwood Jan 31 '12 at 17:47
Ah OK, thanks :) – Matthew Read Jan 31 '12 at 17:47
@JeffAtwood Yeah, that's probably right. I expect a system administrator researching fingerprint authentication or somesuch is not very likely to casually call everyone a fag and try to deface the accepted answer. – a cat Jan 31 '12 at 17:54
StackOverflow's stats: 36.4% approved from anonymous edits. – vcsjones Jan 31 '12 at 20:06

You must log in to answer this question.

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