The Signpost

Investigative report

Is the requests for adminship process 'broken'?

Contribute  —  
Share this
By The ed17

Related articles
Reforming RfA

Will the new RfA reform come to the rescue of administrators?
16 May 2024

Jimbo's NFT, new arbs, fixing RfA, and financial statements
28 December 2021

Editors discuss Wikipedia's vetting process for administrators
26 September 2021

Administrator cadre continues to contract
31 July 2019

The Collective Consciousness of Admin Userpages
31 January 2019

The last leg of the Admin Ship's current cruise
31 July 2018

What do admins actually do?
29 June 2018

Has the wind gone out of the AdminShip's sails?
24 May 2018

Recent retirements typify problem of admin attrition
18 February 2015

Another admin reform attempt flops
15 April 2013

Requests for adminship reform moves forward
21 January 2013

Adminship from the German perspective
22 October 2012

AdminCom: A proposal for changing the way we select admins
15 October 2012

Is the requests for adminship process 'broken'?
18 June 2012

RFAs and active admins—concerns expressed over the continuing drought
14 February 2011

RfA drought worsens in 2010—wikigeneration gulf emerging
9 August 2010

Experimental request for adminship ends in failure
13 October 2008

Efforts to reform Requests for Adminship spark animated discussion
23 April 2007

News and notes: Arbitrators granted CheckUser rights, milestones
6 February 2006

Featured picture process tweaked, changes to adminship debated
27 June 2005

More articles

Requests for adminship (RfA) is the process for creating new administrators on Wikipedia. RfA has been around since June 2003, when Camembert created the page after discussion on a mailing list. Reform of the process has been discussed continually since at least 2005, but the process has remained virtually the same as when Camembert created the page in 2003.

Brief history

The first major attempt to reform the RfA process was with discussions for adminship (DfA). Bureaucrats actually initiated many of the mooted changes before reverting to the old format pending discussion. The idea behind DfA was relatively simple—a week of discussion would precede the voting phase. This would combat voting that organizers felt was little more than "stab[s] in the dark" thanks to limited information about each candidate. They took pains to note that they were not looking to completely modify RfA, but they felt limited tweaks were necessary. Ilyanep emphasized the "ridiculously increasing high standards", "huge amount of instruction creep", and RfA's similarity to polling. Lar noted that while the process was currently working and "good enough", he believed "... 'better' is the enemy of 'good enough'. I think that even if this current process works pretty well, there may nevertheless be better ones out there."[1]

Also in 2006, Adminship renewal, proposing a term of office for administrators, was rejected. This idea is common; it had been proposed before and has been proposed many times since, which is why it is now listed on the perennial proposals page. In 2007, a large survey was conducted to attempt to find what the views of the Wikipedian community were. Around the same time, Wikipedia talk:Requests for adminship/Reform was created to attempt to improve the process. The page quickly ballooned before participation suddenly declined, leading editors to conclude that it was "dead" and trying to draw consensus for RfA reform from it would be flawed. 2008 saw a major request for comment, where many of the perennial proposals were debated, but no consensus was found for any substantive changes. Interestingly, the request spawned from an attempt to delete the main RfA page. Many of the arguments against reform were based in the number of administrators promoted in the previous month (34), which led to the belief that with so many successful RfAs, the process itself could not be broken. As Deskana stated, "Many people desire that RFA be changed, to improve Wikipedia. Change can be good, and it can be bad. What a fair chunk of the people that argue for the reform of RFA don't seem to appreciate is that it isn't 'broken'. That's not to say that it can't be improved. Broken would imply that it's not working at all, which it is."[2]

RfA reform 2011

The most recent and probably most in-depth attempt at reform was in 2011, with the aptly named "RfA reform 2011". The process was immediately spawned by My76Strat's RfA, which was widely viewed as a microcosm of the larger problems at RfA. In its aftermath, Jimmy Wales commented "RfA is a horrible and broken process", and the general feelings led to the creation of RfA2011 by Kudpung. Even this massive effort failed to break the RfA deadlock; although nearly all of the participants desired some sort of reform, they were unable to get their proposals adopted by the main RfA community. Worm That Turned (WTT), one of the coordinators of the initiative, blamed this on the inability to find a root cause of the problems at RfA: "Different people thought it was it was too hard, too easy, got the wrong candidates through, was too uncivil, had too many questions, could give votes without reasons, with [poor] reasons... there was a long list." Kudpung added that the participants were also discouraged by those who vocally voiced their opinion that the page would not accomplish anything.

Why has reform failed?

With these many attempts at reforms, why is there still no consensus that RfA needs to be changed? Perhaps the answer lies in something simple: the natural conservative tendencies of Wikimedians, as most notably illustrated by Ironholds in the Signpost and on his blog. This is not to say most Wikipedians hold conservative political views, but that they are resistant to most forms of change. As Ironholds stated on both pages, "Wikimedians actually tend to put a fairly small amount of stock in changing things to boost the community or the social aspects of the movement. Whether it's WikiLove, help reform or any other project to ameliorate the less pleasant aspects of the projects, the same refrain comes from an annoyingly large chunk of the community ... people don't like change, and ... existing editors are largely comfortable with the current situation". This mindset is seen in the RfA archives with comments like "maybe it's better we stick with what we've got, and try and tweak it to perfection!" or "It's the worst system except for all the others." [3]

Current opinions

With the demonstrable issues with RfA's process and culture—its central two tenets—many strategies, some mentioned above, but most not, for addressing them have been raised. Perennial proposals may sum it up best: "While RfA is our most debated process and nearly everybody seems to think there's something wrong with it, literally years of discussion have yielded no consensus on what exactly is wrong with it, nor on what should be done about that." Reformist editors are therefore swimming against a strong current to even stay afloat, much less find concrete proposals that may garner support. Despite acknowledging these difficulties, there are those who still attempt to reshape RfA. The aforementioned RfA2011 was successful in implementing a editor review-style process which gives candidates a chance to catch possible problems before the public process of RfA. It was also able to put an edit notice above the main RfA page warning inexperienced candidates of their RfA's likely conclusion, and contributed large amounts of research into RfA which is still available for other editors to read through. The Signpost asked WereSpielChequers, Worm That Turned (WTT), Dweller, and Kudpung what changes to process and culture it would take to bring RfA to a level where it could help maintain the administrator core while not driving disheartened editors away after a bad experience.

WereSpielChequers' full views on RfA are available in his userspace, but he believes that RfA is 'broken' and that action needs to be taken to halt the decline in active admins—more than 300 since its peak in 2006. His solutions for RfA vary:

A counterpoint is provided by WTT, who says that his research conducted during RfA2011 has led him to believe that the process is not 'broken' per se, but "it's keeping the right people out and letting the right ones through. If people stopped considering it as a hell-hole, I'm sure they'd realise it isn't one. ... adminship is "no big deal", even if RfA is." Buttressing WTT's argument, only one editor has passed RfA with less than 3,000 edits since 2009, but he had over a million edits to other Wikimedia projects. Dweller agrees in that he does not think RfA is 'broken'. However, in his view the standards for becoming an administrator have risen—in some cases too high. The impact of this has been limited, thanks to the increase in highly-capable bots, but the "ordeal" of the process itself has a detrimental effect on the current efforts to retain editors: "Even more than good admins, we must value our editors, and bad experiences threaten ongoing participation and RfA is an area where bad experiences can happen." Kudpung takes a middle ground, saying RfAs have "become such a rare event, it's not possible to be able to say whether it has become calmer or not."

Still, even if we are ensuring that only qualified candidates are applying, then there is clear evidence that the number of qualified candidates is falling. The administrator corps is currently in decline through attrition and a lack of new blood (see table, below). Whether RfA is 'broken' or functional, it seems to not be fulfilling its intended purpose of at least maintaining the number of administrators: there are 705 active administrators as of this writing, down from a peak of 1,021 in mid-2008.[4]

Successful requests for adminship on the English Wikipedia
Month\Year 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Totals
January 2 13 14 44 23 36 6 6 3 1
February 2 14 9 28 35 27 9 7 9 3
March 8 31 16 34 31 22 13 2 9 1
April 6 20 25 36 30 12 14 8 3 3
May 10 23 17 30 54 16 12 8 6 1
June [5] 24 13 28 28 35 18 12 6 4
July 3 11 17 31 26 31 16 10 7 4
August 4 9 12 39 26 18 12 11 13 1
September 0 17 29 32 22 34 6 8 6 4
October 0 10 16 67 27 27 16 7 7 3
November 3 9 27 41 33 56 11 13 4 2
December 1 15 25 68 19 34 9 6 1 4
Total promoted
Total unsuccessful
Total RfAs, including by email


  0–5 successful RFAs
  6–10 successful RFAs
  11–15 successful RFAs
  16–20 successful RFAs
  21–25 successful RFAs
  26–30 successful RFAs
  31–35 successful RFAs
  36–40 successful RFAs
  41–50 successful RFAs
  51–60 successful RFAs
  More than 60 successful RFAs

See also


  1. ^ "Here's what you haven't figured out...," Wikipedia talk:Discussions for adminship.
  2. ^ "View by User:Deskana," Wikipedia:Requests for comment/Wikipedia:Requests for adminship.
  3. ^ "Current system," Wikipedia talk:Requests for adminship/Archive 92.
  4. ^ Wikipedia:List of administrators: 19 June 2012 versus 28 February 2008.
  5. ^ 33 had been appointed in early 2002.
  6. ^ Early RFAs were done by Email and only the successes are known.
  7. ^ 2004–2011.
  8. ^ Unsuccessful for 2002 to 2003 are not available.
In this issue
+ Add a comment

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.

I tried once. Not likely to try again. They want flawless and super-active people, good luck finding them. Maybe admin rights should be given to every user who has shown themselves to be trustworthy by contributing in a positive way and not having been blocked for a certain period (e.g. 1 year). Make it similar to applying for rollback. If a user abuses their rights just take them away. Or else have a modding system like on Slashdot where established users can mod down an edit and if it gets too much of a negative vote it gets high up in a queue for admin reviewing. Targaryenspeak or forever remain silent 21:21, 19 June 2012 (UTC)[reply]

As someone involved in some of the discussion 5-6 years ago, I guess the recent outcomes aren't too much of a surprise. My observations then were that calls of it being broken etc. were largely driven by individual cases where person X should have been promoted or person Y shouldn't. Maybe it's more general now and it is a belief more people should pass, I'm quite out of touch with that side of wikipedia, but I guess at the bottom of it the concrete reasons as to why weren't really forthcoming, beyond the same material listed in the RFA itself. i.e. it's more or less to do with simple disagreement about who will/won't make good admins. It's pretty much the same reason then why reforms are hard to define, a way to remove the opinion part is to provide solid criteria, however defining those solid criteria requires the input of the same set of opinions which will undoubtedly differ. Other methods such as boards to create admins, lead to suspicion that it'll end up in "group think" type situation whereby those selected are all of the same ilk, rather than being the diverse sets of the past. As said though wikipedia is ultimately quite conservative as that is part of the core principals - outcomes with no consensus default in retaining the current status quo in pretty much all forums. -- (talk) 22:19, 19 June 2012 (UTC)[reply]

Have to agree with Targaryen. The large number of failed RFAs and the dropping numbers of attempted RFAs are a good sign that people are being driven off by the process, not that our process is working to guarantee better admins. Nyttend (talk) 02:11, 20 June 2012 (UTC)[reply]

The voting system should be eliminated. Non-substantive arguments should be discouraged and ignored, with the final decision going to the bureaucrat. Like an AfD, if the arguments for a certain outcome are superior, that outcome should win, even if the other side gets more "votes". That would disempower those few nasty, petty individuals that make RfA such a hellish process. --JaGatalk 04:31, 20 June 2012 (UTC)[reply]

If you want to start with reform at rfa then you may want to consider listing the reform options up at rfa for a year, then letting candidates choose the reforms they want to test in their rfa, then draft proposals based on that approach. Allowing the candidates to test the reforms would provide actually usable data for any reform attempt so that we could move forward with facts and not opinions, fear, guesses, etc. TomStar81 (Talk) 09:34, 20 June 2012 (UTC)[reply]

Make it less of a "big deal" to get, and less of a big deal to suspend (either by oneself or forcibly). And introduce a three- or six-month probation period. It's one of the few reforms in which Jimbo could take the lead, but understandably, he sees problems whichever way one goes. Tony (talk) 10:11, 20 June 2012 (UTC)[reply]
The Signpost "Discuss this story" section is transcluded from Wikipedia:Signpost/Template:Signpost-article-comments-end. The "Discuss this story" and the "In this issue" section are transcluded from the same source, so its impossible to remove the white sidebar (which is technically the "In this issue" section) without substituting the template. Michaelzeng7 (talk) 22:00, 20 June 2012 (UTC)[reply]
What is the link to the "In this issue" sidebar in the comments section? I can't find it. I can fix the width if I can find it. --Timeshifter (talk) 12:48, 21 June 2012 (UTC)[reply]
Yes, there needs to be some kind of representative democracy involved in admin recruitment and selection. The current method of direct democracy has become too unwieldy for such a large organization as English Wikipedia. --Timeshifter (talk) 18:09, 20 June 2012 (UTC)[reply]

I've started up an RFC on this topic. Please feel free to comment at Wikipedia talk:Requests for adminship#Request for comments about whether the RFA process should be changed. Targaryenspeak or forever remain silent 18:31, 20 June 2012 (UTC)[reply]

I suspect that it might be more acceptable if there was a second intake stream in which a representative cabal of experienced admins invited existing non-admins to join their ranks. Highly undemocratic, but totally sidestepping the inquisition stage which many non-admin abhor. Even lowly editors with only a few thousand edits can identify very quickly a few names that they would trust to do the job, but persuading those names to be put forward at RFA is like drawing teeth. It isn't that RFA isn't working, it is that many consider the pain not worth the very minimal gain.  Velella  Velella Talk   21:57, 20 June 2012 (UTC)[reply]

Some of us consider RFA not to be working for that very reason - too few qualified candidates are willing to run. ϢereSpielChequers 23:00, 20 June 2012 (UTC)[reply]
Yeah, sure...many non-admins are just chomping at the bit to have the admins pick their own favorites to rule over them in the way that User:Velella describes. That's exactly what I have been sensing lately.- UnbelievableError (talk) 02:50, 21 June 2012 (UTC)[reply]
I suspect that the great bulk of editors wouldn't care either way. Only a handful are seriously interested in where admins come from. Nick-D (talk) 03:16, 21 June 2012 (UTC)[reply]

Brilliant article, and much needed. Every admin action remains logged and reversible -- adminship should be no big deal. Clearly some improvements are needed to how RfA is handled and how adminship can be removed if an admin doesn't work out. I like a number of the suggestions above:

Thanks for this writeup, and for the great comments. – SJ + 13:17, 21 June 2012 (UTC)[reply]

I don't see a lot of evidence that this is anything but a symptom of the changing (read: shrinking) active user base. Yes, I'd admit that the unofficial requirements (>3000 edits, not automated or automated depending on the phase of the moon, sainthood, etc.) restrict the pipe a bit but the nozzle doesn't matter if there isn't any water pressure behind it. by and large new editors become admins. This is more true now than it was before, but it has always basically been true. If the supply of new editors who stick around for 3-9 months and care enough about the site to want to be administrators dries up, so to will the pool of admin candidates, regardless of what RfA looks like. Protonk (talk) 07:49, 22 June 2012 (UTC)[reply]

Hmm. "Caring about the site" and "wanting to be an admin" don't necessarily equate. [Reinstated comment which seems to have been accidentally deleted] Pesky (talk) 09:34, 22 June 2012 (UTC)[reply]
I don't see a need, save for pathological specificity, to disambiguate the two. If you wish, please insert "care to be admins". Protonk (talk) 00:55, 23 June 2012 (UTC)[reply]
Not sure that is quite true. I have been around for 7 years+ with over 30000 edits but have no wish to submit myself to the inquisition. Sure having a few extra buttons would be great, but I, and quite a number of similarly placed editors, are quite content to allow others to press the buttons at our request. I would be interested to know the evidence behind your assertion that by and large new editors become admins , I am not convinced that this is true.  Velella  Velella Talk   08:37, 22 June 2012 (UTC)[reply]
With a little research... - you find that the average tenure of a successful candidate has gone up from 2.4 years in 2009, to 3.4 years in 2011 (and 4.5 years so far in 2012). My analysis implies that Protonk has that wrong. WormTT(talk) 08:55, 22 June 2012 (UTC)[reply]
Your facts about tenure are correct, but your interpretation is inverted. Average tenure of admin candidates is rising but so is average tenure of long term editors (rising nearly in lock-step). If we were seeing the same number of admin candidates with the tenure changing I'd be inclined to agree with you. But if we see the tenure rise AND we see the RfA rate plummet then I'm hardly going to accept that as a counterclaim. When I say "by and large new editors become admins" I mean that in a general sense and (depressingly) in a a specific sense. In the general sense, long term editors make a choice (sometimes vocalising it, sometimes not) as to whether or not they want to be admins. Those who delay that choice tend to suffer rockier candidacies and are likely not to try again. This is the sweet spot for RfA success. Enough edits to establish credibility but not enough to have pissed someone off. That spot moves over time (usually to the right) but it exists. In the specific sense you can overlay the new admin rate against a time series of new "active" accounts and see a pretty good lagged relationship. We're now 3 years after editor growth started to plateau and we're seeing the results of that in a process which operates mainly on newer editors. None of this means that RfA isn't broken or doesn't need fixing. But we (as editors) tend to focus on all the internal fiddly bits and ignore the large scale changes which are driving most of the dynamic. Protonk (talk) 17:28, 22 June 2012 (UTC)[reply]

RFA also has the problem that it privileges editors who are not willing to edit in controversial areas. If you frequent, say, the fringe theories noticeboard, and deal with the problems raised there, you're going to have enough people who hate you for not letting them promote their views on Wikipedia that you will never pass RfA. RfA seeks people with great experience in things like AfD, but throws out people who work hard to support five-pillar policies such as WP:NPOV, because they'll have upset too many people.... and so gets admins who have absolutely no knowledge of any of the real problems of Wikipedia. (talk) 16:42, 22 June 2012 (UTC)[reply]

I can only remember one RFA where the candidate's problem was that they'd edited a controversial article and a bunch of people from the "other side" derailed their RFA, there may be more, but it is rare. My experience of RFAs is that overzealous deletion tagging, lack of tenure, poor communication skills, answering the questions without first rereading the relevant policy and lack of content contributions are all frequent causes of RFA failure. Editing controversial subjects may get you a bit of scrutiny, but if you are doing so civilly and neutrally and citing reliable sources you should be OK. If a couple of outraged editors from the other side of an editing dispute turn up and oppose over it, then provided you keep your cool and they are simply taking a content dispute to RFA you will probably be fine. ϢereSpielChequers 23:58, 22 June 2012 (UTC)[reply]
Bingo. Someone give a prize. This is precisely the reason I don't support RfA. It elects candidates who have no real experience dealing with actual problems and encourages potential RfA candidates to avoid getting into any disputes in order to get elected. Why anyone would support this kind of system boggles the mind. Viriditas (talk) 10:22, 23 June 2012 (UTC)[reply]
This particular issue is the known problem of low RfA activity (and, as a consequence, very much varying population of voters). If 300 users would vote, 10 fringe theory freaks would just not be visible. This actually surprises me. I remember for instance from Russian Wikipedia that for an arbcom member it was almost impossible to not vote at an RfA, because everybody wanted to know what their opinions are. Here, there are arbcom members I have never seen at RfA, ano some more whom I have seen one or two times.--Ymblanter (talk) 10:53, 23 June 2012 (UTC)[reply]

Things sure have changed!

"but the process has remained virtually the same as when Camembert created the page in 2003.".

I'm pretty sure that the process has changed. I recall that edit-counting was discouraged when I became an admin. I also distinctly recall (therefore) being able to read *all* of a person's edits and questioning them about them.

I maintain that anyone who can be trusted with the tools should have them, especially now that admins have been nerfed and can't delete the wiki anymore. We could give the buttons to 80% of the regular editors or more.

Ceterum censeo RFA editcountitis esse delendam.

--Kim Bruning (talk) 15:57, 22 June 2012 (UTC) ps. I was the nominator for the first WP:100 RFA (source). I now refuse to nominate people on the grounds that I don't do unto others what I would not have them do unto me[reply]

That's interesting. Thinking outside the box, it's probably much safer to give the buttons to people who'll probably never use them and didn't want them in the first place than it is to risk giving them to someone who sees the "power" aspect of the job rather than the litter-picking, pest-control, and loo-cleaning aspects of it. Pesky (talk) 17:17, 22 June 2012 (UTC)[reply]
The most recent RfA failed because some of the voters suspected (contrary to the statement of the candidate) that the candidate actually does not want to have the tools, and swiftly voted him down.--Ymblanter (talk) 18:46, 22 June 2012 (UTC)[reply]
That's just ancient common sense, Pesky! ;-)
Ymblanter, hmm, yeah, I'm beginning to get the impression that the problem might not lie with RFA, but with people's ideas about what an admin is or should be.<scratches head>
If we unbundle the buttons entirely (which is now quite possible), we would incidentally end up abolishing RFA at the same time. That wouldn't necessarily be a bad thing!
--Kim Bruning (talk) 20:08, 22 June 2012 (UTC) By splitting responsibilities more like this, I can imagine this could lead to (further) esperanzification of certain sub-areas of Wikipedia (like CSD or 3RR noticeboard). But that's the next problem, let's take things one step at a time. --Kim Bruning (talk) 20:08, 22 June 2012 (UTC)[reply]
I've always been in favour of unbundling the tools, and giving them out (like rollback) to people who would use that particular tool in a trustworthy fashion, and who ask for it and are obviously sensible people. There would be nothing stopping there being some kind of mini-RfA for the more "dangerous" tools, but with some unbundling there wouldn't be such an expectation of people having to be multi-talented, able to do anything and everything, and so on. One wouldn't have to give the delete button, for example, to the vandal-fighters (who need the block button but may not be the best judges on content matters), or the block button to content-specialists who are good at knowing what can and should be deleted, but whose personal skills may leave something to be desired ... and so on. I've had rollback for over a year; I think I may have used it twice. Possibly three times. Pesky (talk) 05:55, 23 June 2012 (UTC)[reply]
(de-dent) I agree with KB concerning editcounitis. somewhere along the way it's gone from: Well, it seems these days rfAs don't pass unless the candidate has x # of edits and been registered for z length of time; to: Well, you should withdraw as NOTNOW because you don't have x # of edits and been registered x length of time.
It was innocently done, but it's seemingly happened. Maybe we should see about "undo"-ing it? - jc37 21:04, 26 June 2012 (UTC)[reply]

What year?

Signpost editor, please note: the second paragraph under "Brief history" begins "In the same year . . ." - What year?? Textorus (talk) 21:56, 25 June 2012 (UTC)[reply]

Done, nice catch. Ed [talk] [majestic titan] 21:58, 25 June 2012 (UTC)[reply]


The Signpost · written by many · served by Sinepost V0.9 · 🄯 CC-BY-SA 4.0