20
$\begingroup$

Is it currently possible to get a mean score for either questions or answers for yourself, or other users? The motivation behind asking, is that sometimes total reputation is not always a clear indication of the quality of the average post by a given user.

In some instances for example, a user has offered so many bounties, so that their total reputation is unrepresentative of the quality of their usual post. Similarly, some MO users have a relatively low score on MSE owing to their infrequent contributions. The same principle applies to some new users.

$\endgroup$
  • 4
    $\begingroup$ SE has added an "impact" score which is supposed to represent a lot of different factors, such as quality of questions answered. It more or less represents the number of people who may at least have seen something you have done and benefited, or something like that. But for a quick look at a high-rep user (or one that is very active in answering questions), I still like to look at the ratio of rep to number of answers. $\endgroup$ – Ron Gordon Dec 9 '15 at 15:37
  • 1
    $\begingroup$ @RonGordon ah, I see what you mean now. I think I am asking something slightly different, but thatnks for the pointer. $\endgroup$ – martin Dec 9 '15 at 15:41
  • 7
    $\begingroup$ @RonGordon: That's not indicative at all. There are quite a number of high-reputation users who do nothing but answer simple questions, such as basic limit questions, and hence also have a high ratio of reputation to answers, higher than many who answer harder questions that few even read. $\endgroup$ – user21820 Dec 10 '15 at 0:05
  • 1
    $\begingroup$ @user21820: We here at M.SE do not differentiate between "hard" and "easy" questions at times like this. Every question is equal. If you want to differentiate across tags, that's a whole layer of complication. I never claimed complete accuracy, just a rough measure of the effectiveness of the average answer of each user. I assert that this helps differentiate those who simply are rep farming vs those who supply truly useful answers. $\endgroup$ – Ron Gordon Dec 10 '15 at 0:32
  • 2
    $\begingroup$ @RonGordon: And I disagree with that assertion, because it is easy to give answers to questions that you know will garner the most reputation, and I've seen a number of such users who did just that in order to reach the moderator level (rapid reputation climb to 10000 and then plateau). $\endgroup$ – user21820 Dec 10 '15 at 0:47
  • $\begingroup$ @user21820 Is that really a problem? The questions are getting good answers. For some people it's also easy to give answers to hard questions and reach 10,000 quickly. They answer their hard questions, and the people with less expertise answer the easier questions, and the people with even more expertise answer harder questions, and the people with even less expertise... $\endgroup$ – Matt Samuel Dec 12 '15 at 1:37
  • 1
    $\begingroup$ @MattSamuel: You should know by now that many of all those simple questions are homework questions. Of course people who want their homework done for them will upvote work done for them. You cannot assume that higher votes means better answers. $\endgroup$ – user21820 Dec 13 '15 at 2:31
  • 3
    $\begingroup$ SE "reputation" is little correlated to anything - esp. quality. $\endgroup$ – Bill Dubuque Dec 14 '15 at 15:59
  • 1
    $\begingroup$ @user21820: Actually, many of the questions that might well be homework questions don’t get uprated by the asker, because the asker hasn’t enough reputation to uprate an answer. They may not even get accepted, because the asker may not understand the system. $\endgroup$ – Brian M. Scott Dec 19 '15 at 2:56
  • 1
    $\begingroup$ @BrianM.Scott: That's only one part of the story. The asker only has at most one vote and at most one accept. But other users contribute all the other upvotes, and unsurprisingly questions that plenty of students have (due to homework) get more upvotes than questions that few are interested in (because it does not affect their grades). I don't think it's arguable that such factors have a large influence for a large proportion of questions. $\endgroup$ – user21820 Dec 19 '15 at 6:23
  • 2
    $\begingroup$ @user21820 mainly I think it is users tend to upvote things they read and understood and appreciate a bit. This is not in itself unreasonable. But leads to short answers to things that are relatively simple for the predominant population of the site getting many votes. Whether they are homework in a literal sense is tangential in my opinion. $\endgroup$ – quid Dec 19 '15 at 14:56
  • $\begingroup$ @user21820: I think that it is at most an extremely minor effect. My experience suggests that quid’s observation is quite correct. Oh, and the bulk of low-level questions that might well be homework do not, in my experience, attract large numbers of upvotes. We do occasionally see clear instances of several people, possibly from the same course, asking the identical question, but this is noticeable in part precisely because it is not the norm. It may be that my experience is atypical, since I tend to concentrate on topology and combinatorics and don’t look at most of the calculus ... $\endgroup$ – Brian M. Scott Dec 19 '15 at 18:03
  • $\begingroup$ ... calculus questions, but I don’t remember things being substantially different when I was spreading my efforts around a bit more. $\endgroup$ – Brian M. Scott Dec 19 '15 at 18:04
13
$\begingroup$

If you are looking for statistics about this site then data.stackexchange.com is the place to go. You can query the SE database about practically anything. If you don't know SQL you can run some of the other queries that people have made.

Here are some relevant queries for your question:

What is my average question score?

What is my average answer score?


To have something to compare against; below is a histogram of the average answer score (pure votes, not including accepts) for all high rep. users (arbitrarily defined as $>50$k reputation) enter image description here

$\endgroup$
  • $\begingroup$ thanks, now i know that my average answer score is quiet terrible ;-) $\endgroup$ – tired Dec 14 '15 at 14:27
  • 1
    $\begingroup$ @tired Nah, that's quite normal. Most of the high. rep. users have something between $2$ and $4$. Also note that these queries do not take the 'Accept Score' into account so the score should be a bit higher (that adds $1.5\frac{p}{100}$ to the score). You can find your accept percentage here. $\endgroup$ – Winther Dec 14 '15 at 14:41

You must log in to answer this question.

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