User talk:Tamzin/340/112/16: An RfA debrief

Thoughts

I so appreciate everyone who takes the time to write a debrief. This one seems particularly challenging to have written and so I am especially appreciative. I do have a few reactions:

The RfA process could be restructured to remove or greatly reduce the possibility of crash-and-burn failures (roughly making the social impact of a failed RfA no worse than that of a failed request for a sub-admin permission). Agreed. This restructuring would have to be major; Agreed. the first idea that comes to mind as potentially sufficient is banning replies to votes by anyone other than the candidate. I don't see how this helps with the loss of social standing. I hypothesize that the reason we don't see 80/20 rfas anymore is because many of the 20% have been socially pressured to not say silly stuff. So either something is serious enough to merit many opposes or it's trivial enough to get less than 10. Having a situation where more people feel OK opposing seems like it would give more people for whom there is a loss of social standing, not less. So I would love any more insight into how this last part helps with the earlier pieces.

I'm going to cheat slightly on this next one. The community, on a deep and pervasive level, enjoys RfAs like ours. Strongly agreed. RfA won't get better because the community doesn't want it to get better. I think this is slightly more ambiguous. RFA2021 was a failure but I'm not sure it's fair to say it's a failure because the community doesn't want RfA to get better.

the only way [RfA change] will happen is if people say publicly that they're not comfortable running under the current system, and stick to that. I think you're missing just how common this is already. See this discussion for a recent and concentrated grouping of people saying just that. In other words, it's not some new solution. People are already choosing, explicitly as in that discussion and implicitly in the declining numbers, not to run. I do agree that the longer this happens the more likely it is that there will be an eventual will to fix RfA but I think the actual not running matters far more than the people saying it publicly.

Thanks again for writing this and, despite my invitation above, please feel no pressure to actually reply. Best, Barkeep49 (talk) 15:43, 10 May 2022 (UTC)[reply]

@Barkeep49: Limiting replies to opposes was a thought going in, but I no longer think it would be sufficient. I think the only way forward is the scuttling and wholesale replacement of RfA as a system. And I agree that it's good that people have been publicly indicating their reluctance to run, but I think a more visible effort with a unified message would make a big difference. I would make a userbox or something, but I don't think it's my place, as someone who can't boycott RfA because she's already done it (and already committed to re-RfA under certain circumstances), to actually start that effort. -- Tamzin^{[cetacean needed]} (she/they) 20:37, 10 May 2022 (UTC)[reply]

I wonder if that might be more successful as an ongoing list, rather than a userbox, because I would expect some of the people whose declines carry the most impact for others might also be reluctant to sport such a userbox. And FWIW, I am not at all sure that such a "boycott" effort needs to be led by a non-admin to be successful. Best, Barkeep49 (talk) 21:34, 10 May 2022 (UTC)[reply]

Wikipedia:Admin hopefuls for RfA reform? theleekycauldron (talk • contribs) (she/they) 23:08, 10 May 2022 (UTC)[reply]

WP:post-2015 Admins for RfA reform Valereee (talk) 15:51, 2 October 2022 (UTC)[reply]

I agree with what Barkeep49 has said here. I really appreciated reading each of your (C, 5, F, R) perspectives on what occurred, and I thank all of you for sharing them. I know I said privately to TNT, as well as attempted to in my replies on the RfA page, that I was very worried about the emotional and physical cost that a lot of editors commenting seemed to forget about when they chose their words. I think any reform or remake of the process needs to address that, though perhaps that's a discussion for another place. Sideswipe9th (talk) 20:46, 11 May 2022 (UTC)[reply]

Second runs

My second RFA was a very long time ago, and maybe second runs have changed their odds without my noticing. But my experiences both as a candidate and as a nominator is that second RFAs are generally uncontentious, provided the candidate has waited a few months and addressed the major reasons why the first RFA failed. If anything, a first run RFA can neutralise an issue, as subsequent runs focus on events since the last run and whether the candidate has responded to the opposes. Ϣere SpielChequers 16:50, 10 May 2022 (UTC)[reply]

@WereSpielChequers: The sample size of second+ RfAs is even smaller than RfAs overall, but my main concern with any analysis thereof is a selection bias. People run a second time after NOTNOWs and NOTQUITEYETs, after "not enough content creation"s or "you've had an edit-warring block in the past year"s. Do they run a second time after RfAs that cross over into critique not just of suitability for adminship, but of suitability to be an editor or indeed the candidate's overall moral fitness? Not a rhetorical question. I'd be curious to see the numbers, assuming there's some semi-objective way to differentiate these categories. -- Tamzin^{[cetacean needed]} (she/they) 20:45, 10 May 2022 (UTC)[reply]

I'll admit to some sample bias in my perspective on this. I succeeded on my second attempt, and I was the nominator for another person's second amd successful run. I wouldn't describe either of us as having lost our first runs on a "notquiteyet" basis. However I see the logic that you'd expect an RFA that failed on "character" issues to be harder to overcome than one that failed due to inexperience. It is about a decade since I did a trawl through unsuccessful RFAs, but I remember that one thing I noticed was that a lot of people respond to an unsuccesful RFA by leaving the project. I've tried since then to be clearer in my oppose rationales as to what I'd expect to see to flip me into the support camp in a subsequent run. Ϣere SpielChequers 08:21, 11 May 2022 (UTC)[reply]

Well, the leekycauldron way is to throw statistics at the problem. So, here you are:


Attempt no.	Successful	Unsuccessful	Success rate
All RfAs	2,185	2,792	43.9%
2nd	243	376	39.3%
3rd	47	106	30.7%
4th	9	38	19.1%
5th	5	11	31.3%
6th	1	4	20%
7th	0	1	0%

With exception for the 5th RfA, which seems to be a lucky number of some sort, it looks like each successive RfA (not counting reason) is less likely. The possibility for success on the first try has to be higher than the total, so that the average balances out... wasn't able to differentiate based on failure reason. theleekycauldron (talk • contribs) (she/they) 08:40, 11 May 2022 (UTC)[reply]

Here might be a way to look at it: Is there a correlation between word count in an unsuccessful RfA (incl. talk page) and whether a candidate runs again? Proposing this as an indicator of failure type because "Oppose. Candidate has never contributed more than 100 words to an article" tends to generate fewer responses than "Oppose. I have it on good authority they kick puppies." Or could do an XY plot of word count, number of S+O+N, with the marker for a data point representing the boolean "Ran again?" -- Tamzin^{[cetacean needed]} (she/they) 09:17, 11 May 2022 (UTC)[reply]

okay, so some very rough back-of-the-napkin programming pegs the median bytesize of a failed first RfA+talk (which is followed by a second RfA) at around 20,000 bytes. This was pretty shocking to me, but yeah, lots of second RfAs followed pretty clear immediate failures for first RfAs. Comparison value forthcoming... theleekycauldron (talk • contribs) (she/they) 10:39, 11 May 2022 (UTC)[reply]

the overall median (for both those who run again and those who don't) is somewhere around 50,000... theleekycauldron (talk • contribs) (she/they) 10:51, 11 May 2022 (UTC)[reply]

I wouldn't put too much weight on that raw number of bytes. Two things that will make a big difference are that many people withdraw if their RFA is not heading for success, and the inclusion of a plug for live RFAs in the watchlist notice has dramatically increased participation in RFAs - the most recent 2% of RFAs and the 2% with the highest votes have a huge overlap. Ϣere SpielChequers 16:30, 11 May 2022 (UTC)[reply]

I think you're right, but I also think it's not so impactful on the data—omitting the 2% maybe changes the median from 20,000 to 18,000, from 50,000 to 45,000. And if most people withdraw early, that means their RfA wasn't so contentious, right? As I understand it, the question is "do users who fail a first, contentious RfA run a second time", and I think the answer is that they're less likely to do so. theleekycauldron (talk • contribs) (she/they) 18:34, 11 May 2022 (UTC)[reply]

Thanks Leeky Cauldron, any chance of splitting the second RFAs between those that were within 3 months of the first RFA and those that weren't? I have seen at least one overhasty second run that crashed immediately. We've also had second RFAs that were reconfirmation or runs after the person had been desysopped - I expect that the success rate of second RFAs where the first was successful could be lower than the success rate for second RFAs where the first was a success. As for the rest, I suspect they include a bunch of RFAs from people who didn't learn from the reason for their first RFA failing. Ϣere SpielChequers 11:15, 11 May 2022 (UTC)[reply]

This is super interesting. But I would also suggest most of this data comes from an era of RfA radically different from our own and so it's hard to say how 2nd runs are really received in a contemporary context so WSC could be correct, we just don't know. Best, Barkeep49 (talk) 13:51, 11 May 2022 (UTC)[reply]

OK, so I've now had a look at the second and subsequent RFAs from the last two and half years, not a valid sample but what there is is in accord my thinking. People running again after a long gap since their unsuccesful RFA do very well if they have addresssed the issues in that first RFA. Typical gaps were so many years that I can't judge if the old 3 month minimum rule still applies. Former admins running again may have been less successful but again the one who had clearly addressed the issue for the desysop was fine. Ϣere SpielChequers 16:22, 11 May 2022 (UTC)[reply]

Yeah, that's roughly what I would have guessed. But again my point is about the negative space here, who doesn't run again. That's the real metric of failure of RfA as a process. (Yes there are some failed RfAs where a candidate could reasonably say, even moving past any self-flagellation, "Wow I just really don't have the cops to ever be an admin", but I think most failed RfAs are of people who could be viable candidates eventually.) -- Tamzin^{[cetacean needed]} (she/they) 16:31, 11 May 2022 (UTC)[reply]

I agree that almost everyone who has failed an RFA could potentially pass RFA in the future. I don't agree that the real metric of failure as a process is the lack of people running a seond or third time, I think the worse metrics are the number of people who don't run at RFA because of its risks and reputation and the number of unsuccessful candidates who leave the project. I have various ideas as to how we could improve the process. But to me the heart of it is that we haven't agreed a criteria for adminship, and the reason why we can't agree a criteria and why the process can get so toxic, is that it covers two hot button issues which divide the community: Deletionism and Civility. A lot of RFA Opposers are there to stop candidates who they think could be heavy or lighthanded with the block or delete button, and some have nothing to gain from reform as the current system means you don't need even close to 50% in order to have a de facto criteria for failing RFAs, but you might be nowhere near getting consensus for adding that to the RFA criteria. Ϣere SpielChequers 15:47, 18 May 2022 (UTC)[reply]

Comment from Bilorv

Moved from User talk:Tamzin

– -- Tamzin^{[cetacean needed]} (she/they) 10:20, 14 May 2022 (UTC)[reply]

My main takeaways: RfA won't get better because the community doesn't want it to get better. The community, on a deep and pervasive level, enjoys RfAs like ours. I shouldn't act holier-than-thou as I can be drawn into this myself, but I do agree. A lot of people write that RfA is broken, but few say so with their actions.

That you drafted a leaving message is quite a bombshell. We knew that, if the 'crats found no consensus, that would just be the end of things for us on Wikipedia. People who voted oppose should be under no illusions: they voted for you to leave the community. It is hardly as if there is no precedent for what happens after RfAs like yours fail.

You have some interesting advice: I do think it establishes a clear lesson for others, though: Don't RfA. You don't know what random thing will turn out to be the single aspect of your work that hundreds of people judge you by. ... But the only way [RfA will be replaced] is if people say publicly that they're not comfortable running under the current system, and stick to that.

I do feel that I reached some similar conclusions in the 'crat chat commentary: Each oppose has the effect of not just opposing this one person becoming an administrator, but opposing all people becoming new administrators, and of opposing this one person continuing their outstanding volunteer work on this website. ... I cannot [run for RfA] unless I am prepared to treat the exercise as a joke, in which whether I pass is a matter of randomness and I am scrutinised based on things I do not believe or that do not accurately reflect my contributions to this website (and in which this paragraph will inevitably be quoted back to me in a hostile manner).

I guess your suggestion to me would be to refuse to run for RfA, something I've been achieving for a fair few years now. But I am more pessimistic as many Wikipedians seem in denial of the fact that the website is currently on fire and facing existential threats in several different directions. The community, however, is incredibly resistant to all change. RfA seems not to be special in this regard.

(Feel free to move my comment here—or just remove it—if you don't want it on your main talk page.) — Bilorv (talk) 09:35, 14 May 2022 (UTC)[reply]

@Bilorv (feel free to give this a better section heading): My suggestion would be more than not to run for RfA, but, rather, to affirmatively and visibly say that you will not run for RfA until the system has undergone drastic change. Not just more 'crat clerking. Not just some agreement to enforce CIV more strongly, which will fail for the same reason civility paroles always fail, namely that civility enforcement is an equilibrium state and can't be forced higher or lower. Something more like marking the whole process as historic and coming up with something new. Depending on order of events, does that maybe mean some span in the middle where there's no way to become an admin because the community hasn't come up with a new procedure?

Possilikely (a mix between possible and likely). But maybe that'd just make for a better incentive. -- Tamzin^{[cetacean needed]} (she/they) 10:20, 14 May 2022 (UTC)[reply]

Yes. I did commit myself to running under the PROD-style adminship proposal of 2021, and my comment might read flippantly but it wasn't a flippant decision to make that statement. RfA needs to be changed through conscious experimentation: pass thresholds like the current 65–75% range aren't deduced axiomatically, but discovered through practice. I will support almost any seriously designed proposal to change RfA, and offer myself as a test dummy wherever I can. Though I don't quite have confidence that me saying "I will not run for RfA until..." would be seen by most Wikipedians as an "oh no, we're missing out" rather than a "thank god" or a "who are you?". — Bilorv (talk) 14:02, 14 May 2022 (UTC)[reply]

I've been asked twice if I wanted to run for admin (these were independent queries by different editors). The first time, I declined because while I had written a fair amount, none of it had been formally stamped with a "GA" or "FA", so I figured I'd have a rough experience. The second time, I made what I hope was an appreciative remark but still felt extremely dubious; I had an article officially designated Good by that point, but my impression of the process had only grown worse. Now, I've probably torched my chances of ever crossing that bridge by being vocally burned out, repeatedly, for weeks at a time. But for what little it's worth, I will totally sign on and declare that I will never run for admin under the current system. XOR'easter (talk) 01:58, 18 June 2022 (UTC)[reply]

Though I don't quite have confidence that me saying "I will not run for RfA until..." would be seen by most Wikipedians as an "oh no, we're missing out" rather than a "thank god" or a "who are you?". I've seen you previously and likely therefore am not most Wikipedians, but in general my immediate inclination toward any random editor (with a reasonable contribution history) wouldn't be "who are you". Rather, just as I trust most Wikipedians (past the extended confirmed mark, perhaps) to be civil and cognizant of the basic policies, I would similarly trust them with admin tools. At least, I think misuse would be rare enough for the community to address via noticeboard. – Anon423 (talk) 02:52, 19 June 2022 (UTC)[reply]

Given how many people agree the current RfA system is broken, the chance of getting a critical mass to publicly state they're not comfortable running might be higher than you think.

What are the other existential threats, by the way? – Anon423 (talk) 02:54, 19 June 2022 (UTC)[reply]

Congrats buuut

Thanks for sharing but many many people were hoping youd address the concerns, could you please? This kinda comes off as tonedeaf.Tooscaredtospeak (talk) 20:05, 26 May 2022 (UTC)[reply]

Are there specific concerns you'd like to ask about that I didn't answer in the RfA or in this debrief? -- Tamzin^{[cetacean needed]} (she/they) 07:35, 27 May 2022 (UTC)[reply]

General thoughts

I missed this whole RfA and personally looking at it now I'd probably be neutral on the whole candidacy. Given that I indirectly prompted the comment which led to the question that blew up your RfA, I'd say there's an undercurrent of hypocrisy in this entire wiki relating to how we treat extreme left political views and how we treat extreme right political views. I got put on blast for bringing up a candidate's current political views by the same people who said that it's your right to oppose a candidate based on political views. The difference was I was opposing someone who I believed to be a leftist and you were discussing opposing someone who was on the right. That being said, you weren't hypocritical at the RfA in that regard (you did say support for Stalin/Mao is dq'ing) and I don't know enough about you to have an opinion on whether you'd be a good admin; I believe everyone should have the ability to impose their virtue tests at RfA because the alternative is having a central authority define what opinions are and are not OK.

The reason why I'm bringing this up now is because the substance of this debrief is that you're mad that people opposed you and imposed virtue tests on you, when the entire nexus of this dispute is your answer saying you would impose a virtue test on others. To quote "C" (who hasn't been listed as a "minority opinion" so it seems like they speak for most of you): "All this even though most of those personal attacks and misrepresentations made no attempt to tie back to the question of suitability to serve as an administrator. Because RfA isn't about suitability to serve as an administrator. It is, and has always been, a virtue test."

C says this, but you still feel you can impose what you openly acknowledge is a virtue test that nobody else would be OK with. F says that "You don't know what random thing will turn out to be the single aspect of your work that hundreds of people judge you by." yet you judge people by what many people consider to be a single aspect of their personality in regards to RfA. 5 says that "Things started to stabilize a bit, particularly as people began to defend our right to object to support for our own oppression." but again go on to criticize the very system that gives you (and everyone else!) the right to object to support for your own oppression.

This is the reason why RfA can't be "fixed". RfA gives everyone in the Wikipedia community the ability to oppose or support based on what they personally feel is legitimate; and this necessarily leads to conflict when people's ideas of what are "valid reasons" differ significantly. Everyone hates this conflict, but there's no way to resolve it without creating and endorsing a rubric for grading admins; but to do so would take away everyone's ability to idiosyncratically judge candidates according to their particular standards so nobody seems to want that to happen.

I'm certainly in the minority here but I don't believe there's anything wrong with RfA. I grew up having never known a world where Wikipedia has not existed as the sum total of human knowledge. The majority of what I have learned is from Wikipedia. This is true for most English-speaking people nowadays, so it's impossible to call RfA a WP:NOBIGDEAL when it's the confirmation process for the judges and arbiters of what is the most widely used source of knowledge in the world. When you are an admin, you are granted a massive amount of power with endless discretion and not much oversight on how that power is used, power that extends far beyond the clique of those who edit Wikipedia, power that is more than a fair chunk of elected public officials.

You said yourself you want to be involved in WP:ARBIPA; that would make you responsible for policing the edits of a community with 1.64 billion people in it. [1] Many of those people speak English, given that those are the official languages of India and Pakistan.

I would say it's a fair price to put admins under the microscope in a public trial by fire, given the damage an actively hostile admin would do to the project. In 2015 an admin abused their power to create thousands and thousands of unnecessary redirects. It took 2 and a half years + the creation of a special deletion criteria to clean those redirects up. [2] Imagine how much worse it would be if it was on an issue that actually mattered. Chess (talk) (please use {{reply to|Chess}} on reply) 06:16, 20 July 2022 (UTC)[reply]

@Chess: Unfortunately, our opinion of the status quo is irrelevant. It literally does not matter whether RfA is "fair" or not, nor whether people will agree for it to be fixed. The attrition numbers are totally unsustainable; many processes such as RfD are routinely backlogged by multiple days, and it is only going to get worse. In another five or ten years, we'll have to start using meta:Steward requests/Miscellaneous just to keep the wiki running (except I imagine they'd move it to meta:Steward requests/English Wikipedia because our workload would overwhelm the rest of the page). Either we replace RfA now, in a calm and considered manner, or we run out of admins and then replace RfA in a blind panic. --N Y Kevin 00:47, 24 July 2022 (UTC)[reply]

Great debriefing

I missed your RfA (was on an unintended break) but I have been encouraging people to debrief for quite a while now in hopes of convincing the community that we have a problem. Many non-admins think we have a specific need to make RfA "difficult" -- if you can't deal with RfA, how will you deal with adminship? -- and don't see that it's not just difficult but is often an incredibly hideous experience that can actually cause lasting harm. Many admins who RfA'd more than seven years ago think it can't possibly feel as bad as some of us are telling them it is because "I've been through it and it wasn't that stressful." Like they think we're either exaggerating or we're just not made of the same tough stuff as they are. So thank you for this. Valereee (talk) 16:11, 2 October 2022 (UTC)[reply]