What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 134 Stack Exchange communities.

I've stumbled across this question:

update one table from the conditions of another table

Where people start complaining about the accept rate, as can be seen very often on Stack Overflow. If you check out this user's questions, it turns out that they have only asked 8 questions, some of which only recently and with a bad question quality (such that there simply isn't any good answer yet).

Maybe it might be a good idea not to display the accept rate so early? Or following a key like this (just an example):

  • If < 3 questions asked, then don't display it
  • If < 5 questions asked and with a rate = 0%
  • If < 7 questions asked and with a rate < 20%
  • If < 10 questions asked and with a rate < 40%
  • If < 15 questions asked and with a rate < 50%
share|improve this question
9  
The bad question quality is the biggest problem, not the accept rate. I'd very much prefer if we displayed that meter instead. –  Cody Gray May 24 '11 at 12:10
2  
I agree. Some early users get a lot of heat coming their way when they simply haven't spent too much time. Maybe a 1 month limit or something –  JohnP May 24 '11 at 12:10
1  
@JohnP, yes that might be even better! –  Lukas Eder May 24 '11 at 12:17

3 Answers 3

up vote 10 down vote accepted

Accept rate requires at least 4 eligible questions to be present. In fact, your example user only has 7 eligible questions.

This means, in order to have accept rate displayed at all, you need at least 4 questions that have been answered and have had answers for at least 3 days. If your questions are merely not getting answers at all (which often is paired with poor quality), then your accept rate won't be displayed until you start getting answers.

share|improve this answer
    
But how can you circumvent that, as a user? Delete the bad questions? It's not about accepting in this case, but about improving question quality. –  Lukas Eder May 24 '11 at 12:19
    
Argh, I was about to say that @Grace :P I will add that these conditions are enough for "beginners" on the sites. So experienced user should not have problem to maintain an acceptable ratio. –  M'vy May 24 '11 at 12:20
3  
@Lukas If this is about improving question quality, then why are we talking about accept rate? If you're asking decent questions, and engaging with users to actually get answers to your questions, then by the time you have 4 eligible questions you should ideally have a pretty fine accept rate. –  Grace Note May 24 '11 at 12:20
    
OK, I mean the user's problem is not the accept rate but the question quality. But the problem for Stack Overflow is still the accept rate, because it is displayed too early –  Lukas Eder May 24 '11 at 12:23
3  
@Lukas Any time it is displayed with poor stats early seems to indicate one of two things: poor questions overall, or someone who just doesn't realize that you can accept answers. These seem to be pretty good things to notice, so I don't see why this is a problem for Stack Overflow. –  Grace Note May 24 '11 at 12:25
    
I agree on those two things. But there are a couple of other reasons that might be the cause of a low accept rate, and my point here is that there are some ways to improve the stats, even if the data is still scarce, see also Robert's answer –  Lukas Eder May 24 '11 at 13:54
    
@Lukas Those are generally much rarer cases. I don't think we need to inhibit our ability to detect the vastly more numerous problem users in order to make the experience temporarily better for a select few. –  Grace Note May 24 '11 at 14:35
    
@Grace, I don't have access to the overall statistics. You may be right, maybe this isn't really a relevant issue, here... –  Lukas Eder May 24 '11 at 14:43

Right. IMHO accept rate isn't very useful. Doesn't tell anything about quality.

Much more interesting than measuring quantity of accepts, would be measure their quality. For example there should be some indication, if user frequently accepts the wrong answer, resulting in other answer getting the [populist] badge.

share|improve this answer
    
I think that would be very hard to quantify, wouldn't it? And often, answers are upvoted because they seem better, but in fact the best answer (from the question's perspective) is another one –  Lukas Eder May 24 '11 at 14:29
    
@Lukas: yeah, but populist badge is no just one or two votes more, it's an answer, that outscored an accepted answer with score of more than 10 by more than 2x –  vartec May 24 '11 at 14:32

"Accept rate" is a difficult thing. I've seen people chided for their low accept rate on ServerFault with people telling them they need to "work on it". What happens when people ask perfectly good questions and don't get answers that are worthy of accepting, then get chided about that when they ask new questions?

share|improve this answer
1  
That's another good point. Although, I think on Server Fault, there aren't as many users, so the chance to get a good answer in a short period of time is even lower –  Lukas Eder May 24 '11 at 13:55

You must log in to answer this question.

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