Wikipedia:Administrator elections

Administrator elections are a second process by which interested editors can become administrators on the English Wikipedia.

This process will run side-by-side with the old process

The first process, which has probably existed since around its page creation date of June 14, 2003, is called Wikipedia:Requests for adminship (RFA), and involves leaving public supports, neutrals, and opposes, usually with a rationale. These two processes will both be available to prospective admins. The old process is not being deprecated. It is possible both processes will be available forever, or someone may eventually launch a deprecation RFC for either Administrator Elections or the RFA process, but probably not until we are much more familiar with the effectiveness of this new Administrator Elections process.

Why this second process?

Administrator elections will involve leaving private supports, neutrals, and opposes via the SecurePoll software. Folks can see who voted, but not what they voted. The idea behind keeping this information private is that it will reduce the toxicity of the process for the candidates, resulting in an increase of candidates willing to go through the process to become administrators. English Wikipedia administrator promotion rates have been steadily declining since the golden age of administrator promotions around 2005–2008, where as many as 408 admins were being promoted per year. Modern promotion rates are closer to 12 per year.

List of RFCs related to this process

• Wikipedia:Requests for adminship/2021 review/Proposals#Closed: 8B Admin elections - Proposed by Worm That Turned. No consensus.
• Wikipedia:Requests for adminship/2024 review/Phase I#Proposal 13: Admin elections - Proposed by Novem Linguae. Consensus.

Exact text of the latest RFC

In addition to the existing RfA process, Admin elections are held every six months.

Candidates must sign up by a certain date, followed by two phases of debate:

1. Three days for discussion and questions - no bolded !votes.
2. If candidates choose to progress, a secret ballot for a full week. Voter suffrage would initially match Arbcom elections. Candidates who achieve 70% Support would pass and become administrators.

• Further explanation is at Wikipedia:Requests for adminship/2021 review/Proposals/Admin elections

Exact text of the close

(non-admin closure)

Fair warning, this is going to be long. However, given the number of participants and the questions that needed answered from the closer, I don't think it's avoidable. I apologize, but there is a bulleted summary at the end for everyone's reading convenience.

So, the easy part: there is consensus among editors that administrator elections should be an additional means of attaining adminship on the English Wikipedia alongside the requests for adminship process. For the avoidance of any doubt, this process will run parallel to RfA, and will not otherwise replace or effect it in any way.

For the sake of brevity, I will leave out a summarization of the support and oppose arguments in this closure.

Now, many editors presented concerns with the implementation details of this proposal. A common theme among concerns, though, was that the editor approved of the proposal in it's current state and stated such things could be ironed out later. These same editors, among others, expressed a desire for this proposal to be enacted as a trial run.

Looking at the raw numbers, there is a technical majority of editors who can be interpreted to have satisfaction with the proposal without changes or a trial run. However, most supports were interpreted in this manner due to not having any caveats or requirements with them, not because they stated that the proposal was fully functional as written - almost none did. A group of editors slightly smaller in size to the aforementioned majority have expressed a desire for a trial run or some kind of change. Moreover, such editors generally argued why a trial run, or changing things, was preferable and needed.

With all of the above taken into account, I find that there is sufficient support to implement this proposal for 1 election cycle, after which the community will discuss it's performance in Phase II.

I arrived at 1 election cycle, rather than 2, as many editors desiring a trial did not specify a length. While many of those that did indicated 1 or 2 cycles, considering the number of editors which indicated a desire for adjustments to the proposal either now or at "some point later", I find that the community would likely prefer to review the process after 1 attempt rather than 2.

In terms of the implementation details themselves, I find that because a trial run is occurring, that there is more than sufficient support to implement this proposal with the details as written. By this I mean the election will run as prescribed at Wikipedia:Requests for adminship/2021 review/Proposals/Admin elections (or whichever page such details will be moved too).

In conclusion (TL;DR), based on my best reading of consensus:

• This proposal is successful.
• The community supports trying this proposal for 1 election, after which it will be reviewed in Phase II.
• While there are concerns regarding the implementation details of this proposal, given this is a trial run, there is sufficient support to run the election as written.

I am welcome to any and all feedback. I understand this is a significant closure, and I am not opposed to scrutiny of my reasoning or experience. —Sirdog (talk) 07:30, 15 April 2024 (UTC)

Wikipedia:Requests for adminship/2021 review/Proposals/Admin elections

Phase in an alternative route to adminship of elections every six months. Candidates would sign up by a certain date, then would have a shorter period of 3 days for discussion and questions. There should be discussion only in this period, no bolded !votes. At the end of the period, candidates can progress to the next period, secret ballot (through SecurePoll) for a full week. Voter suffrage would initially match Arbcom elections. Candidates who achieve 70% Support would pass and become administrators.

Breakdown

Alternative route

This would be an alternative route to adminship, which could run in parallel with RfA. It should start every 6 months and designed to not clash with other elections so as to avoid election fatigue.

Format

The proposed format would be on a 10 day cycle. Election dates would be published ahead of time, and once the cycle begins, it would be publicised through watchlist notices. The cycle would have two periods. The process would be managed by the bureaucrats, initially in concert with the WMF to set up SecurePoll. In the future, a new role could be created to manage these elections, but there is no urgency in doing this.

Period 1 would be a discussion period and would last three days. During this time, questions can be asked and issues can be raised, as can positive feedback. Candidates are encouraged to participate in the discussion period, answering questions and responding to feedback. During this time, no bolded !votes should be cast, it should be a clear discussion.

Period 2 would be an secret ballot, using SecurePoll, which should last the full 7 days. During this period, discussion should be closed, and while candidates may be asked direct questions on their user talk pages, the intent is that they should not be required to watch their discussion page, nor the election for the full period, to reduce the "stressful" nature of the election.

Voter suffrage should initially match the Arbcom elections i.e. registered over 1 month before election, 150 edits by election, 10 edits in the year running up to election, not sitewide blocked during the election, not vanished, and not a bot - though the suffrage decisions could be changed in future.
Vote tally should be calculated by Support / (Support + Oppose) for each candidate.

At the end of Period 2, votes can be scrutinised then tallied, results can be announced and new admins can be appointed. As this is a clear vote, the pass rate would be 70%, which is the middle of the discretionary range.

Possible numerical changes

There are 4 numerical values, which should be static at the time of this proposal but can be updated in the future, depending on how well the system works.

• 2 elections per year - depending on the uptake, we could increase or decrease the number of elections per year.
• 3 days discussion - 3 days was chosen as the vast majority of discussion at an RfA happens in that period. During the initial Phase of this admin review we discovered Many editors believe it would be unpleasant to have so much attention focused on them, and reducing the time that the candidate is focussed on the process is a key part of this proposal.
• 7 days voting - This allows editors who edit sporadically to be involved in the actual voting process.
• 70% pass rate - This was chosen as the middle of the discretionary range. However, historically, we have found that secret ballot does give lower support percentages than named voting, so it may be that this number should be changed in the future.

Will there be another RFC before implementing this?

RFA2024 proposal 13 closed as The community supports trying this proposal for 1 election, after which it will be reviewed in Phase II. While there are concerns regarding the implementation details of this proposal, given this is a trial run, there is sufficient support to run the election as written.

So the current plan is to move forward with no further RFCs and exactly one election cycle, making judgment calls on details such as scrutineering. Then after the first election, we can hold RFCs on whatever the community would like to change, and an RFC on holding additional admin elections.

Areas that may need more discussion

• Is 3 days too short for pre-vote discussion?
• How will SecurePoll scrutineering work?

Todo list

• Link this document in other places, including WP:RFA, WP:ADMIN, Wikipedia talk:Requests for adminship/2021 review/Proposals/Admin elections, and Wikipedia talk:2024 administrative elections proposal
• Reach out to WMF to get SecurePoll software set up.
• Inform WP:BN of the plan. Make sure we have bureaucrat buyoff / willingness for bureaucrats to promote whoever wins the elections.
  • In fact, the proposal says The process would be managed by the bureaucrats. Start a discussion with them on how involved they want to be in the process.
• Set dates for 1st election
  • last day to sign up
  • day that discussion closes, day that SecurePoll opens (+3 days)
  • day that SecurePoll closes (+10 days)
• Create subpages for each candidate's questions and answers. Probably just copy the RFA process. Have nomination statements, then optional questions and answers. Maximum two questions per !voter?
• Watchlist notice
• Do it!

Administrator Elections will likely be harder to pass than traditional RFA

Looking at recent Arbitration Committee Election results, top candidates tend to receive around 80% support rather than 100% support. This suggests that secret voting enables and encourages folks to oppose more often than with public voting. This suggests that someone may receive a -20% reduction in supports by using Administrator Elections instead of a Request for Adminship. So if someone might receive 80% support at RFA and pass, they might receive 60% support and fail using an Administrator Election.

Please keep this "gotchya" in mind. Administrator Elections are likely to be less dramatic, but harder to pass, than traditional RFA.

Perhaps this will be remedied later with an RFC to lower the Administrator Elections pass threshold from 70% to 50%.

Suffrage requirements (who can vote)

This proposal copies the Arbitration Committee Election suffrage requirements.

• registered over 1 month before election
• 150 edits by election
• 10 edits in the year running up to election
• not sitewide blocked during the election
• not vanished
• not a bot

We may need an RFC later to simplify these requirements if they become too much of a burden on the WMF and the SecurePoll software. I hear that generating things like "10 edits in the year running up to the election" is time-consuming.