Wikipedia:Administrator elections/October 2024/Debrief

Please leave any comments about how the Wikipedia:Administrator elections/October 2024 went on this page. Please use bullets for small comments and level 3 subsections for large comments. After collecting comments for awhile, the next step will be surveys and RFCs (on a different page) to make changes to the administrator elections rules and workflows, and then an RFC to see if the community is interested in having a second administrator election. Please sign all posts.

Feedback from candidates

Queen of Hearts

Placeholder charlotte ^👸♥ 20:57, 4 November 2024 (UTC)

FOARP comments

General

• Thanks to everyone who worked, stood, and voted in this.
• Getting 106 oppose votes when no negative comments were raised in the entire process..... is what it is. My perspective is a pass is a pass. I think if this were a traditional RFA a lot of these people wouldn't have contributed. If I were any of the people who got over 50% but less than 70%, particularly Valenciano who got fewer opposes than me, I would be feeling hard-done-by right now.
• There seems to have been a block of voters either down-voting every single candidate, or opposing every single candidate whose name they didn't immediately recognise. It's hard to explain people getting so many oppose-votes when no negative comments were made at all about their candidature.

Proposals

• The real risk with this process is it ends up just cannibalising the admins gotten through the RFA process without resulting in more admins. To avoid this we need a process with a high through-put, especially if the pass rate is going to be less than 50%. For that reason we need to hold regular elections - at least every quarter, possibly every month. Any cap on the numbers of candidates need to be scaled to the regularity with which the election is held.
• I really am not sold, at all, on the voting process. I really don't think we've gotten a better process than even just 10 community members discussing someone's candidature in detail and then casting open support/oppose !votes. I would still prefer a more discussion-based process simply because I don't think this process teaches candidates anything - if you failed in this election with no negative comments made whatsover about your candidacy then just what is it you are supposed to do? The voting system also adds complexity to these elections that will hinder them being held regularly.

Feedback from voters

Novem Linguae

• Suggestions
  • max 10 candidates per election (I didn't mind the 32 candidates, but a lot of folks suggested less candidates on the talk page, and this would be an easy fix. I also think modifying the # of candidates is better than elongating the discussion period, because the latter would make the candidate experience worse)
  • lower support threshold from 70% to 65%. the candidates in the 65.00%-69.99% range of the October 2024 admin elections were good candidates, whom I mostly voted support for. would have been good for them to pass
  • put the discussion phase on the weekend, to help out voters with day jobs
  • alphabetize candidates in SecurePoll. i assume many folks research the candidates before arriving at the vote screen, so might as well make it easier to transfer the votes over.
  • change voter suffrage to extendedconfirmed, to match RFA and to greatly simplify the SecurePoll voter eligibility workflow
  • create a subpage that is linked from the main election page where links to voter guides can be added
• lessons / other notes
  • 616 votes total, lowest # of support+oppose votes for a candidate was 318. so there are absolutely no concerns about quorum. more scrutiny per candidate than a typical RFA.
  • candidate with lowest # of opposes was 74. this suggests that 74 of the 616 voters (12%) blanket opposed everyone.
  • candidate with the highest support percentage received 83.42%. this suggests that there is at least a -16.58% support penalty for doing AELECT compared to RFA. judging from ACE, it may be as high as -20%
  • elections are best pre-scheduled, not spontaneous when a candidate list fills up. we pretty much have to use pre-scheduling until enwiki gets local electionadmins, due to the WMF T&S election calendar not having that many openings and having months waits until the next opening.

–Novem Linguae (talk) 20:48, 4 November 2024 (UTC)

It's a minor point, but your suggestion about 74 voters blanket opposing everyone presumes that the candidate with the least opposes had no valid opposition, only blanket. I would not be surprised if the real number of blanket opposers was much lower. Giraffer (talk) 21:22, 4 November 2024 (UTC)

Maximum number of candidates

I think we should be careful with limiting the number of candidates too much, so to not take away opportunities from people. Reviewing this number of candidates was tough (took me >20h to investigate), and some maximum is needed, but I'd rather it be at 12-15 than 10, especially as some people will withdraw during the process. After the first 1-3 elections, we'll likely have much less pent-up demand, so it may be a self-resolving issue anyway. If we can do elections very frequently (say every 2 months), a number as low as 8-10 makes sense, but that will require a lot of scruniteering capacity / organisational capacity. —Femke 🐦 (talk) 21:01, 4 November 2024 (UTC)

Saqib

Thank goodness the admin election is over - now I can focus on the real elections tomorrow! On a serious note, I suggest we cap the number of candidates in future admin elections. This time, it was nearly impossible to vet everyone and many of us had to abstain - a wasted vote!! Also, it shouldn't be a first-come basis for nom; that only adds to the chaos. We should also reconsider how often these elections are held; once or twice a year seems far more sensible than more frequent elections, which might just dilute their significance, imv and I also think, we definitely need more than two days for discussions - this time wasn’t nearly enough to engage meaningfully with each candidate. Just my 2¢. --— Saqib (talk I contribs) 21:21, 4 November 2024 (UTC)