-7

I suggest to use an h-index-style to indicate SE users performance. As with publications, it gives a direct indication of activity level and users interest.

The h-index for answers (for example) is calculated by arranging the answers in descending order and stopping at the answer that has an equal number or higher of votes as its order. For example, a user with 15 answers with the following votes: 50, 43, 42, 20, 15, 14, 14, 11, 10, 10, 8, 4, 2, 2, 1 will have an h-index of 10, while a user with 1 answer of 246 votes will have an h-index of 1 (although both has same number of total votes).

7
  • You'll have to elaborate on what you mean. How would your proposed h-index be computed? In what way would it be different from the current reputation score, and why would it be better/more useful for this website?
    – ff524
    Mar 14, 2016 at 5:52
  • @ff524 I added some explanation, I hope that helps.
    – jak123
    Mar 14, 2016 at 7:18
  • 1
    Yes, thanks. Obligatory warning: voting is different on meta. Downvotes on this post most likely indicate disagreement with the proposed feature request, not that there's something wrong with the post.
    – ff524
    Mar 14, 2016 at 9:03
  • @ff524 as long as it doesn't affect my poor reputation it's ok :)
    – jak123
    Mar 14, 2016 at 9:26
  • Anyone know how to write a query for the data explorer? data.stackexchange.com/stackoverflow/query/45194/…
    – StrongBad
    Mar 14, 2016 at 19:25
  • @StrongBad Try this one: data.stackexchange.com/academia/query/66081/what-is-my-h-index
    – ff524
    Mar 14, 2016 at 20:40
  • @ff524 much better. It probably should be an answer ...
    – StrongBad
    Mar 14, 2016 at 20:50

2 Answers 2

5

Citation metrics are flawed for...lots of reasons. Especially when they're pegged to something like "indicating performance", presumably with the usual rankings and the like that come with it.

Just using StackExchange related problems:

  • It utterly ignores other contributions, such as questions (never underestimate the importance of good questions to a site), comments, etc. As an example, whether you like it or not, JeffE's "Run, don't walk" comment is a cultural item on this site, and an "h-index" wouldn't really see that at all.
  • External votes become a thing. "Hot Network Questions" get lots of traffic, even if they're not particularly good question. Which means that, by strategically answering those, you could boost your "h-index". At the same time, diligent, workman-like answering of lots of questions that are likely to be helpful, but not flashy, and get a few upvotes and an accept, would be discouraged. That's the opposite of a healthy community.
  • We already have reputation
3
  • 1
    1. The reputation system also ignores comments etc... Questions and answers can be part of the h-index system. 2. The voting system does not address the source of the votes. 3. Same argument applies to citations with publications, h-index is a combination of quality and quantity (see the data explorer provided by @ff524 above, a user with 50k reputation has higher h-index than a user with 100k)
    – jak123
    Mar 15, 2016 at 8:20
  • 1
    @jak123 That the reputation system is not fully useful does not suggest a system duplicating many of it's downsides is a good idea. One of the flaws is assuming that "Lots of upvotes" is a reliable signal of quality. I suspect it's much more likely a signal of "Lots of people looked at this question."
    – Fomite
    Mar 15, 2016 at 8:40
  • Same argument applies to publications, it's annexation rather than duplication (as is the case with badges). Quality (as with publications) means many people find it useful. Quality as an absolute concept is, sadly, irrelevant.
    – jak123
    Mar 15, 2016 at 9:01
3

This has been proposed before: Modified h-index for questions and answers?

To me it does not seem all that useful. If you want to look at users based on their H-index, you can use this data explorer query

You must log in to answer this question.

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