Commons:Requests and votes/Voting Approval Poll/Results discussion

From Wikimedia Commons, the free media repository
Jump to: navigation, search

The preliminary results look like 100 edits and 1 month for suffrage and the current 200 edits and 2 months for adminship. Some more checks may be needed, and some feedback are wanted before the results go official. 哦,是吗?(висчвын) 02:53, 31 March 2008 (GMT)

Odd how the adminship one hasn't changed at all, but shame how so many people would prefer to judge voters by their number of edits rather than by what they do. Majorly (talk) 15:04, 31 March 2008 (UTC)
Then again, have you checked for sockpuppets/double voting etc? Do the voters reach the current new eligibility rules? Majorly (talk) 15:07, 31 March 2008 (UTC)
Checking for sockpuppets... hmm... if someone points some out with a basis for a check, then that could be done but I wasn't planning on checking every voter... privacy violation. (this was me a few days ago ++Lar: t/c 19:50, 2 April 2008 (UTC) )
I would also recommend not doing checkuser...CU is not for fishing. 哦,是吗?(висчвын) 00:24, 03 April 2008 (GMT)
I never mentioned check user. One can easily see if someone is a likely sockpuppet by, for example, looking at the number of edits they have made and when they registered their account. Majorly (talk) 01:59, 3 April 2008 (UTC)

vote analysis[edit]

I think rather than seeing a percent of total, I'd like to try to figure out who all the unique voters were and annotate the totals with what the percentage of unique voters each item got. My suspicion is that each of the items that won got a large majority of the unique voters approval. I might be wrong. But I hope I'm right, because it would mean that the selected items have large support as being "acceptable" even if not the "favorite" of the unique participants... ++Lar: t/c 19:04, 31 March 2008 (UTC)

That sounds really complicated. I personally think the vote is flawed - who is to say that one user making a vote here knows anything of what they are talking about? There could be socks, but I have no idea, and see no suspects, but it'd be easy to make one. Also, it lasted only a week. A "change" like this should last a lot longer than that. From the result, we can see there's no actual change on requirements to running for adminship, only now we are barring people from voting, regardless of who they are, but by the number of edits they have and when they joined. Majorly (talk) 19:29, 31 March 2008 (UTC)
OK, see Commons:Administrators/Requests_and_votes/Voting_Approval_Poll_suffrageOnly (119 unique, highest choice was 47 votes) and Commons:Administrators/Requests_and_votes/Voting_Approval_Poll_adminOnly (117 unique, highest choice was 38 votes, both assuming I ran AWB right, check my work, eh?). This shows that even though this was structured as an approval poll and it was emphasized that you should vote for every choice you could live with, there appears to be wide variation here. (see the part up there where I said "my suspicion is..." ??? I was wrong) No choice got a majority in either case. Short of running the poll longer, and doing running totals as we go and encouraging people to vote for more choices, I don't see how else to get consensus out of this. Not sure what to do next, (reopen and run longer with stronger encouragement to pick more choices rather than starting over, or declare no consensus, no change, and crat discretion continues) but yes, I'm agreeing with Majorly, this vote didn't work out as I thought it would. Seemed like a good idea at the time. But you lot are so widely diverging it's frightening! :) (and terrifically cool at the same time that we get so much useful stuff done anyway) ++Lar: t/c 03:47, 2 April 2008 (UTC)

Is it just me, or did the results we got reflect current practice. Certainly in adminship requirements they did (200 & 2), and I don't think we had a suffrage requirement before now. What relevance does this comment have? I don't know. Just thought I'd bring it up. giggy (:O) 23:48, 3 April 2008 (UTC)

I noticed it too. Kind of ironic. Btw, we didn't have a suffrage requirement before. I really don't think suffrage requirements are needed - only in close cases do they matter, and it's the job of the bureaucrat to decide. Majorly (talk) 00:59, 4 April 2008 (UTC)
If you go with "the choice that got the largest votes wins" then no they don't, as there is now a suffrage requirement when before there wasn't. If you go with "no choice got a majority, therefore we have status quo ante" (same standards as before) then ya, the results reflect current practice, by definition. :) ++Lar: t/c 01:01, 4 April 2008 (UTC)
OK then, putting aside admin requirements (let's call that a consensus, for now), the two highest "scores" in the suffrage vote (41 and 47 votes, respectively) were for 50 edits & 2 weeks and 100 edits & 1 month. The 3 months somewhere option polled next highest, at 34.
So, what say we go with something along the lines of "75 edits & 3 weeks". I would propose an "or" one, but that didn't many votes, so how about this as a midpoint? Or is that a bit too complex? (whereas 1 month is more simple...:) Thoughts? giggy (:O) 01:54, 4 April 2008 (UTC)
I would be happy with 100/1mo. or 50/2weeks. Keep it simple. The only thing I don't really like is how there's a gap, however, I think many of those 0 requirement voters were thinking about it differently. They really don't want to let everyone vote, they just don't want it said. Many of the comments were "let the 'crats decide", "it doesn't need to be formal", or "I oppose this poll in general" ;) It's same people who say "it's not a vote" but the thing is, it is (with a touch of discretion when it's close). We're a democracy in denial. Anyway, I would also be ok with no change as there was no consensus. Rocket000 02:43, 4 April 2008 (UTC)

It would be wrong to expect an absolute majority (>50%) to appear in any poll like this where users are asked to vote for every option that they can live with, and running the poll for longer won't change that. I would go with

  • 200/2 for eligibility on the grounds that: (1) that got the most votes, and (2) that's the existing criterion.
  • 100/1 for suffrage on the grounds that: (1) that got the most votes, and (2) that got a decent majority: nearly 15% more votes (47) than the next most popular option (41).

--MichaelMaggs 08:48, 4 April 2008 (UTC)

Only thought about this recently, and nobody's brought it up yet. Does anyone think that the introduction of SUL had any impact on voting...since it was introduced midway through the poll (IIRC)!? Certainly, my votes didn't take it into account. (rant extends here). giggy (:O) 10:44, 4 April 2008 (UTC)
You know, I didn't really give that a lot of thought until after voting... It does change things, but maybe not as much as it first seems (at least to me). I always took the time requirement as activity not account age. I've seen a lot of accounts that were made years before the user did anything more than a few edits. So in either case it's still about Commons edits, although trust is built faster when the user has the same name on all projects. Rocket000 03:55, 12 April 2008 (UTC)

<-- Breaking the results into IMHO obvious groups, (note:my additions get 315 in both groupings and checking the overall count not the 293 as listed);

by edits
  • 56 nil
  • 122 200-250
  • 101 300
  • 7 400
  • 29 500+
by time
  • 56 no time
  • 60 2wks - 1 month
  • 106 2 mths+
  • 81 3 mths+
  • 12 4 mths +

What I get from this is that to become an admin, what the community looks for statistically is about 200-300+ edits spread over about 3 months, at those points basically 75% of editors have reached some point where they'll consider the RfA candidacy as realistic. With suffrage for voting seen as only being known whether here or on another project. Gnangarra 11:13, 4 April 2008 (UTC)


Suffrage Requirements
Qualifications # of votes  % of votes
0 edits and 0 months 14 5.38
0 edits, 0 months, 200 edits and 1 week somewhere else 32 12.31
0 edits, 0 months, 3 months activity somewhere else 34 13.01
0 edits but the account must have been created before the RfA/B 23 8.85
0 edits and 1 week 9 3.46
50 edits and 2 weeks 41 15.77
100 edits or 1 month 10 3.85
100 edits and 1 month 47 18.08
200 edits and 2 months 21 8.08
200 edits and 3 months 29 11.15
Total 260 100

Admin Eligibility
Qualifications # of votes  % of votes
0 edits and 0 months 29 9.90
0 edits, 0 months, admin elsewhere 11 3.75
0 edits, 0 months, 1000 significant edits on another project 16 5.46
200 edits and 14 days or 1000 significant edits on another project 36 12.29
200 edits and 14 days 7 2.39
200 edits and 1 month 12 4.10
200 edits and 2 months 38 12.97
200 edits, 2 months, and 30 uploads 15 5.12
200 edits, 2 months, and 100 image contributions 9 3.07
250 edits and 0 months 5 1.71
300 edits and 2 months 19 6.48
300 edits and 2 months with edits elsewhere or 500 edits and 3 months here 25 8.53
300 edits or 3 months 7 2.39
300 edits and 3 months 37 12.63
300 edits, 3 months, and 100 image contributions 13 4.44
400 edits and 3 months 7 2.39
500 edits and 3 months 9 3.07
500 edits and 3 months and 30 uploads 6 2.05
500 edits and 4 months 7 2.39
700 edits and 4 months 1 0.34
1000 edits and 3 months 2 0.68
1000 edits and 4 months 0 0
1000 edits or 6 months 2 0.68
1000 edits and 6 months 2 0.68
Total 293 100

Declaring results[edit]

Given that the poll was very public don't you all think that the results should be equally as public. If you still don't think that there is a consensus run a straight first past the post run-off between the top two options from each of the polls.

As I see it you have two overlapping bellcurves, each representing a population of voters, that is one population that favours lower and one population higher, suffarage and admin qualifications. You,re getting two peaks, one where the lower tolerable limit of those favouring higher qualifications reinforces the bell curve of those favouring lower qualifications, and a second peak where the inverse is true. A run off of these two peaks should satisfy the greatest number of voters.KTo288 12:52, 5 April 2008 (UTC)

Median voting[edit]

Has it been considered that median voting (if you know a better name for it, please tell me) could be used instead of a regular approval voting with all its problems. Median voting is simple, sincere and swift, and it has been used with success in the Finnish Wikipedia (an example giving the result "70%"). Samulili 19:01, 9 April 2008 (UTC)

How does it work? I fear calling for a third vote though. Better to have some sort of declaration of outcome. ++Lar: t/c 02:52, 14 April 2008 (UTC)
You simply take the media value. Benefits: 1) It's simple - one vote is enough (or two if you want two values, time and number of edits). 2) Compared to average voting, there is no point in trying to manipulate the result by voting for an extreme value instead of your preferred one. 3) In general, there is little room for tactical voting.
Here is an example of a vote in Finnish Wikipedia which resulted in 1000 edits, and 100 days. Samulili 06:21, 14 April 2008 (UTC)