What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 128 Stack Exchange communities.

Summary:

Since a 70% or higher Accept Rate is considered excellent, make the Accept Rate number turn green at 70%, not 80%.


There are a lot of questions on meta about the accept-rate.

It seems, to me, that a large amount of these can relate to the comment "please improve your accept rate", or similar.

Currently, the accept rate colouring is green for 80%+, red / orange at 35% or below and grey otherwise. This implies that only an accept-rate of 80% is acceptable.

The average accept rate for users with greater than 200 reputation is 77.3%. This doesn't test for questions with an answer as the server couldn't deal with the query, which means that the actual number will be higher than this.

Removing the restriction on reputation the average accept-rate falls to 67.6%. This includes the condition that a question must have an answer.

Looking at a graph of the number of users with a certain accept-rate and ignoring the obvious spikes at 25%, 33.3%, 50%, 75%, 80% etc that are caused by users with a small number of questions there isn't really any obvious pattern.

However...

To quote Jeff:

... any accept rate of 70% or higher is considered "excellent"

and the FAQ

In general, an accept rate of 70% or better is quite good

Backing up Jeff et al the median accept-rate for all users (thank you Stack Overflow) is 75%. This is, probably, heavily skewed by the number of users with 100% rates.

In order to reduce the number of questions on meta about the importance of the accept-rate or comments left about it and to reduce the number of comments on the main site could an accept-rate of 70% - or 75% if people prefer - be coloured green?

As green implies good this small UI change might, on its own, be able to affect a change in behaviour.

I like 70% because it's twice 35, which pleases my brain :-).

share|improve this question
16  
+1 Well-argued, sir. –  Ernest Friedman-Hill Jun 9 '12 at 15:09
1  
+1 brain pleasing is good :) –  Mottie Jun 9 '12 at 16:06
2  
I say 68%, because numbers aren't more magic just for ending in zero :P –  Ben Millwood Jun 9 '12 at 18:27
3  
And 68% is 2^5 less than 100. –  Mark Byers Jun 9 '12 at 20:53
5  
Sounds good enough to me, though I don't think you'll solve the "accept rate whining" with this...how many people complain because someone has a 70%+ accept rate? most of the complaints seem to be on rates well below 50%, often 0%. –  Ben Brocka Jun 9 '12 at 21:49
    
Actually, if the median is properly computed, it should not be skewed by users with 100%, because the median is insensitive to outliers / extreme points. And a median of 75% means that half the users have more than 75% accept rate, so a 70% rate doesn't seem that high. –  Mathias Jun 10 '12 at 6:08
    
sorry @robert. For my sins I work in BI, I really dislike the word "executive" :-) –  ben is uǝq backwards Jun 10 '12 at 7:16
    
@Matthias there are a lot of users with a 100% accept rate... –  ben is uǝq backwards Jun 10 '12 at 7:41
    
@Ben: And your accept rate is coloured green :P –  Somnath Muluk Jun 10 '12 at 7:48
    

1 Answer 1

My opinion on this may be colored by the fact that I am color blind (until I read this question I had never paid attention to it), but it looks to me that an obvious solution would be to simply remove the green coloring for "great accept rate".

It looks to me that the objective here is to drive people to accept good answers, not to have a high accept rate. Rewarding high accept rates is the primary source for the anxieties you list, and generates perverse side-effects: having people mark answers as accepted when they know they are not is detrimental to the whole of StackOverflow.

So rather than trying to find the magic level for "Great accept rate", drop that notion altogether, and simply flag unusually low accept rates as Red.

share|improve this answer
1  
+1 Interesting suggestion... –  thkala Jun 11 '12 at 15:53
    
Not sure I agree. I think it's reasonable to assume roughly 2/3 answers posted to the site should be answerable. The most complete answer that actually answers the question should be accepted, even if it's not perfect. Keep in mind that the asker will still be informed when new answers are posted, and can opt to change the accepted answer if they think the new answer is better. Also new answers can be up-voted by others to show their value, even if the asker is no longer active to change the accepted answer. –  raveturned Jun 14 '12 at 13:37
    
I don't know what to think... I mostly post on SharePoint overflow, and I often see or make question that finds no definitive answer (probably because no one knows it? it often comes to undocumented framework part, classes not working as intended and bugs that many experienced but for which no one ever had found a trigger). What should an user do then? My reasoning would be to vote answer that are at last usefull to provide some light to the problem but only accept an answer that resolve the problem. But that could lower your % to the red point, expecially if you avoid trivial questions... –  SPArchaeologist Dec 7 '12 at 9:39

You must log in to answer this question.

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