On per-site metas, users are given a rank based on meta participation, and this ranking is seen on the default tab of the users page on each per-site meta. How is this participation numerically calculated to give a ranking? Is it by votes? Something else? Are community wiki posts included in this?
-
Are you referring to meta.stackoverflow.com/users? The default tab is based on reputation– Josh MeinCommented Aug 28, 2012 at 13:02
-
5@JoshMein No. Per-site metas, like <meta.askubuntu.com/users>, where rep is inherited from the main site.– nanofaradCommented Aug 28, 2012 at 13:03
-
Is the main meta site different for a reason?– Josh MeinCommented Aug 28, 2012 at 13:05
-
6@JoshMein - this site has it's own reputation for various reasons - historical being the main one.– ChrisF ModCommented Aug 28, 2012 at 13:08
-
2@YannisRizos How is each of those weighted? The answer below seems to suggest there is something weird going on...– nanofaradCommented Aug 29, 2012 at 12:17
-
1@JoshMein yep, see Why do I have a different reputation score on meta.stackoverflow.com and stackoverflow.com?– PopsCommented Sep 25, 2012 at 20:20
-
3Votes seem to form a potentially insane amount of "participation", a user with lots of meta votes but almost no posts, comments or received votes is sorted above me (with lots of everything but not as many votes) on Workplace meta– ZeldaCommented Sep 25, 2012 at 22:01
Add a comment
|
4 Answers
This is pretty much answered by the tooltip you get when you hover over the "participation" sort link:
Users most active on this site (combined number of posts, votes, comments and edits).
-- that's all it is. Take the sum of those four values and sort the users based on those sums. The only minor details that are not in the tooltip are things like "don't count deleted stuff", and the fact that "votes" includes answer acceptance and that "edits" excludes edits to your own posts.
answered Oct 2, 2012 at 11:11
-
1Is it just a sum, or is it weighted somehow? Do votes one your posts, your votes on others' posts, or both count? Commented Oct 4, 2012 at 10:20
-
5Just a sum. And other people's votes on my post cannot really be called "my activity". Commented Oct 4, 2012 at 11:28
Contrary to ire_and_curses's observations, whatever the secret sauce is seems to be working. I reviewed the participation tab for a site I moderate and the rankings make perfect sense. All the top users are clearly the most active folks on our meta, and I don't see anybody listed who is a non-participant.
The exact workings of the mechanism remain a mystery. I think we need an official voice to chime in with the backend details. Even though meta reputation is typically just a reflection of the main site, your posts on meta do have votes and the reputation number calculated from these used to be discoverable. I don't know where it is now, but it could well be a factor. Also the raw number of posts, comments, edits, votes etc may be a factor.
A good hint as to what is likely considered can be found on Election nominations. Some detailed metrics are shown under each entry. You can check the latest SO election for a sample:
meta participation
questions: 63 / +746
answers: 128 / +1,016
helpful flags: 1383
I think it is probable that the magic sauce you want to know about is a formula applied to that data with some kind of weighting applied.
I'm not so sure those users are being ranked. On the board and card games meta users tab, user Daenyth is placed 4th, despite a low main site rep, having never asked or answered a meta question, and with only 18 votes.
Compare this for example, to my 'position' (8th), with more participation in every category, and the ordering hypothesis doesn't seem to make much sense.
answered Aug 28, 2012 at 21:10
-
1Main site rep has nothing to do with Meta participation, hover over the participation link and read the tooltip. Participation includes posts, votes, edits and comments (on Meta only).– yannisCommented Aug 28, 2012 at 22:39
-
@Yiannis Rizos - I assume the same thing, but include it for completeness, since the algorithm is unknown. However, you are missing my point. The example user has no posts, no edits, no comments, and only 18 votes. Commented Aug 28, 2012 at 22:45
-
My observation is quite different. I wonder if raw visits to the meta site counts and perhaps that user is a lurker or has a script downloading the page regularly for them. Do they have a meta fanatic badge? What does their visit days counter show?– CalebCommented Oct 2, 2012 at 9:27
-
2@Caleb - I think this has changed recently. The user page I've linked to now has a completely different ordering from when I answered this a month ago. On casual inspection it does now look reasonable (my example user is nowhere to be found). Commented Oct 2, 2012 at 9:58
An unannounced change was released in July 2022 that added time periods to meta user lists.
As of when the change was made, Users > Participation shows Week, Month, Quarter, Year, and All views.
Before this change, the Participation page showed the most active users on Meta in the last 60 days, and it was impossible to see historical participation. Now the tooltip says
Users most active on this site (combined number of posts, votes, comments and edits).
The Quarter, Year and All views show data participation data for extended periods, in the same way as Voters and Editors tabs and the Reputation tab from the main sites.
While not announced, this change was noticed previously in a couple of bug reports on MSE:
- Searching for a username on a Meta site under the "Participation" tab throws a 500 server error
- Per-site meta "Users" tab filter is not functioning correctly with pagination
- Filter URL Parameter not considered when navigating between user pages on Metas
In particular, the second example has a comment from Catija explaining the change:
[...] the option to choose the time period was just added a month or so ago on metas, so there does seem to be something going on here... likely what happened is that, because meta had no filter by timeframe, they forgot to retain it in the URL when switching between the pages.
-
3@Joachim Nah, I got it added about a year ago because I was frustrated that you couldn't see users on Meta who hadn't been active recently and it was annoying. There's a bug related to the rollout here. There's a duplicate that I left comments on in this regard.– CatijaCommented Oct 4, 2023 at 6:56
-
2I had always intended to post something about the change here on MSE but it uh... escaped my memory.– CatijaCommented Oct 4, 2023 at 6:58
-
1Well the question here is how the participation is calculated. This is NAA, so I'm flagging as such, and voting to delete. Commented Oct 4, 2023 at 7:19
-
This is now a lot better than the rev. 1. While this might not be a complete answer, having buttons for periods is relevant about how the calculation of the participation is done since July 2022 (h/t @Catija) has to consider the time start date of each time period and the current date .– RubénCommented Oct 4, 2023 at 9:08
-
3At the point in time the question was asked, the 60 days limit actually did impact who showed up on the list. Someone who had been inactive for >60 days essentially had no meta participation based on the old UI - and that's why I got it changed. As such, I think that explaining the new feature and the old 60 day limitation is a valid aspect of this question worth answering. Since this has been edited since being deleted, I feel the current state is an answer, so I've undeleted it.– CatijaCommented Oct 4, 2023 at 15:40
-
@Catija Probably, the "conflict" is that the accepted answer was updated to show the current tooltip. Still, there was no explanation about the implications and no mention of the time periods. I'm unsure if the accepted answer should be rolled back or if this answer should be "merged" with it. Because you have added most of the details to all the other things you shared, I'm asking you, What do you suggest to do? (feel free, to delete this answer if you think that it's better to update the accepted answer or make it a CW, what I'm looking for is to have the info available for reference).– RubénCommented Oct 4, 2023 at 16:04