-5
$\begingroup$

After hearing a response from my own suggestion to improve our gauge of a member's usefulness to the community, I've since had a change of heart on the solution, go ahead and down vote it, my previous suggestion, yep its pretty bad. Now, I'd like to hear other people's suggestions to improving the reputation system.

So, What other ways beside rep can gauge the quality of a user's posts, and more generally their contributions?

$\endgroup$
  • 17
    $\begingroup$ To read the post, perhaps? $\endgroup$ – Did Apr 9 '15 at 20:58
  • 13
    $\begingroup$ Why do we need to gauge the quality of a user's contributions? $\endgroup$ – Milo Brandt Apr 9 '15 at 22:01
  • $\begingroup$ @Meelo considering we already do it with reputation, I'd think that having an alternative method might help. For instance there is a metric called "impact score", that was created on the stack exchange blog to measure contributions in a different way. $\endgroup$ – Zach466920 Apr 9 '15 at 22:06
  • 2
    $\begingroup$ Does anyone have a link to the SE blog discusiion on "impact score"? $\endgroup$ – Bill Dubuque Apr 9 '15 at 23:07
  • 1
    $\begingroup$ @BillDubuque This looks like the desired post. $\endgroup$ – Milo Brandt Apr 10 '15 at 1:13
  • 6
    $\begingroup$ @Meelo Thanks for the link. The proposed "impact factor" will be even more meaningless than rep. It is sad that SE is doing nothing to attempt to reverse the extreme scoring bias on FGITW quick answers. Due to that, many of the best (but slower) answers have far fewer votes than they deserve. They will only be further denigrated by adding more meaningless numbers with even higher FGTW bias. That's a great way to attract further gunslingers and cherry pickers. But not a great way to attract great teachers who spend the time needed to compose truly excellent answers. $\endgroup$ – Bill Dubuque Apr 10 '15 at 1:39
  • 1
    $\begingroup$ On participation tab you can find score from last 60 days which, in som way, combines number of posts, votes, comments and edits. See Participation for the main site and Hidden features of Stack Exchange?. $\endgroup$ – Martin Sleziak Apr 10 '15 at 5:47
  • 2
    $\begingroup$ I don't think Brian or André are qualified to be called FGTW, but somehow they have more rep than half the site combined (half? quarter is probably true). Reputation is about how much effort you put in, and how many answers you dish out. Regardless of their quality. Some users which are constantly adding low quality content have several thousand points. Not because they are fast, but because they are actively contributing to the site. $\endgroup$ – Asaf Karagila Apr 10 '15 at 8:09
  • 1
    $\begingroup$ @BillDubuque do you have a idea on what to do to fix the problem? $\endgroup$ – Zach466920 Apr 10 '15 at 14:06
  • 1
    $\begingroup$ @quid: Go to the users page. Note there are ~3800 pages of registered users, each with 36 users. Go to page 1900, the users are scored 15 points. Go to page 2500, they are all scored 1. This means that roughly the lower half of registered users accumulated only 370,800 points. And if you want to talk about earned points, then those 1 point users have earned nothing, so we shouldn't even count them. $\endgroup$ – Asaf Karagila Apr 10 '15 at 18:45
  • $\begingroup$ @AsafKaragila that's a good point, if we could eliminate people who just use the site for one or two posts or who have earned less than 10 rep in the last month, a large part of the problem might just disappear... $\endgroup$ – Zach466920 Apr 10 '15 at 18:53
  • $\begingroup$ @Asaf just as I had posted the second incarnation of my comment I realized that I read "quarte of the points of the site combined" where you likely meant (and now confirmed) "the points of the lower quarter of the site combined." (which is why I then selfdeleted it quickly but not quickly enough) The latter is true. But then as you explain the lower quarter of that list (btw no unreg users there) merely has around 30k together; thus many a user alone would have the points of the lowest quater. I change my objction to: I do not think this is very meaningful thing. $\endgroup$ – quid Apr 10 '15 at 19:01
  • $\begingroup$ @quid: Oh, in terms of reputation wealth, yeah, they are probably around 2-5% combined. But here the drive was the "class warfare" issue of high reputation vs. low reputation. $\endgroup$ – Asaf Karagila Apr 10 '15 at 19:20
3
$\begingroup$

You (technically) ask different questions in the title and the body. On Stack Exchange, contributions $\neq$ reputation and posts. There are many, many, many other ways in which a user can help out:

  • Editing posts without making trivial edits
  • Completing review tasks without merely skimming the post at hand
  • Participating on meta without simply giving knee-jerk opinions
  • Adding (helpful) comments without making noise
  • Voting without blindly clicking a mouse

Here are some links with which you can explore the users who do each of these the most:

These are all valuable contributions. However, you can't just point to a user and think, "This guy is #$a$2 on edits, so his contributions must be really helpful!" because you have to look at the edits to judge how much they help a post (if at all). Don't be superficial.


Now, in your question body, you write

So, What other ways beside rep can gauge the quality of a user's posts,

Well, read them, as Did said! Some posts get ridiculously high amounts of upvotes when they aren't much better than average, because the question at hand becomes quite popular. The reverse can also be true: great answers can be given to questions with low views, and so they'll receive lower scores. I always read posts in their entirety before voting. Don't be a sheep and do what everyone else is doing.

Stack Exchange has the paradigm for users of "I'll use my vote however I want!" So votes aren't always the best way to measure contributions. Accepted answers are better, but remember that only one answer can be accepted per question, and answers to poor, off-topic questions aren't always worth as much as answers to good questions - unless the answers provide great assistance to the person asking the question and others viewing it in the future.

So reputation can be a good indicator of contributions, although churning out a bunch of sub-par posts can get you the same rep as a smaller amount of high-quality posts. I know this might sound odd on Mathematics, but you have to look beyond the numbers and read the posts themselves.

And look at other ways the user helps the site, though editing, reviewing, participating on meta, commenting, and voting - and some other helpful things, like being around in chat.

The rep system works pretty well, but don't judge a user purely by rep. Look deeper.


1 Sorry; I'm not too good at using Data Explorer.
2 Let's assume that a is really low.

$\endgroup$
  • 3
    $\begingroup$ About "posts get(ting) ridiculously high amounts of upvotes when they aren't much better than average", a phenomenon regularly mentioned is that, from time to time, flocks of users from other SE sites swarm a given question and upvote it massively, simply because the question was mentioned in the list of "Hot Network Questions". Personally I do not like the argument too much because it makes math.SE users feel that they are not responsible for erratic votes on math.SE (while they are, for the most part) .../... $\endgroup$ – Did Apr 12 '15 at 6:58
  • 3
    $\begingroup$ .../... but this swarm-from-outside-the-ken phenomenon, although it does not explain everything, should be kept in mind. $\endgroup$ – Did Apr 12 '15 at 6:59
  • $\begingroup$ My jaw dropped when I read "The rep system works pretty well". Really? Works pretty well for what? Measuring how quickly one can pick cherries? That's about the only thing rep seems to measure well. $\endgroup$ – Bill Dubuque Apr 12 '15 at 22:00
  • 1
    $\begingroup$ @BillDubuque I'll admit that the Fastest Gun in the West does better on Mathematics than on other sites. $\endgroup$ – HDE 226868 Apr 12 '15 at 22:08
  • $\begingroup$ @HDE226868 I have to agree with Bill here. If we just made a system that was better, we wouldn't have to waste time reading 10 posts to make a judgment about a user we probably won't even see again. Its obvious that reading the posts is the best way, it just sounds like beating around the bush to say that and then not admit rep has serious problems. Especially considering that at least 90% of users don't have the time or the care to read additional posts to form their judgments. If we improve the system 90% of users will have a slightly better way to judge other users contributions. $\endgroup$ – Zach466920 Apr 13 '15 at 14:09

You must log in to answer this question.

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