Wikipedia talk:WikiProject Editor Retention/Archive 36

Editor retention trends
Hi all,

I have been poking around in the Wikipedia XML dumps lately. That has led me to generate the following graph of the proportion of editors who made their first mainspace edit in a particular year who were still making at least one edit after two, five, and ten years. I find this graph highly disturbing for reasons that I trust are obvious:

I knew things were bad but I guess I didn't really believe they were that bad. Anyway, I have two questions that I wonder if any of you fine folks might be able to answer: Many thanks for your time! -- Visviva (talk) 05:13, 7 August 2022 (UTC)
 * 1) It appears that the entering class of 2012 had an especially bad time of it (being the only one to post a retention rate of less than one in twenty after two years). I also notice that 2012 is the year this WikiProject was founded, and I wonder if it might have been spurred by the same causes, whatever they were. Were there any particular events happening on Wikipedia around 2012 that might have made things especially difficult for new editors in that year?
 * 2) Is there any way of looking at the above data that isn't a damning indictment of the changes in Wikipedia policy and practices, and in particular of those of us who joined in and before 2005, who were welcomed warmly into the project ourselves but promptly turned around and pulled the ladder up behind us? I have been mulling some possible alternative explanations, but I haven't come up with anything terribly persuasive.
 * There are numerous factors that can affect how attractive a project is to new participants. Projects evolve over time, causing their needs to change, and thus different processes or different skill sets in their members become more important. It's hard for any project (be it commercial or non-profit) to adapt to the changing environment, so for me, "indictment" is too strong a word. English Wikipedia's consensus-based decision-making traditions is one barrier, as it makes it hard to make any major changes, but that's not to say that other decision-making approaches would have fixed all problems.
 * The universe of available online pastimes has also expanded greatly. As someone who prefers the wikitext editor, and who writes HTML by hand, I get why long-time editors eschew more flashy interfaces. But the space that Wikipedia-as-a-contributor project competes in offers lots of engaging ecosystems to entice people away. Projects are typically more exciting in their initial expansion phase, and then become more routine later on. The outlier, from this perspective, is the first set of years. Trying to attract more long-term editors is still a desirable goal, but the reality is that Wikipedia may have to learn to accommodate more frequent generational turnover. isaacl (talk) 15:36, 7 August 2022 (UTC)


 * Thanks for taking the time! I guess what had led me to reject the theory that the nature of the project has changed is the continuing high levels of retention in the early editor classes. If it's a matter of the nature of the work or the project having changed, you would expect to see higher attrition for older editors, compared to recent editors who have taken the project as they found it. But in fact it's the oldest editors who have stuck around most, and the newer editors who are leaving in droves. (As of 2021, 9% of 2004 joiners were still editing, higher than the two-year retention for any class since 2007.) And people are still trying to join us, by the hundreds of thousands every year. Given that policy debates these days are mostly about how many alligators we should have in our moat, rather than why an encyclopedia that draws its strength from openness would want to have moats and drawbridges to begin with, the cause of the trend seems fairly obvious. But as you say, many things have changed over these years, so perhaps I am not thinking about this the right way.
 * Based on the interesting differentials in dropoff for the early classes (and in the opposite direction for 2012), I've been thinking about this in a voice-exit-loyalty framework: if you've had a good first-months experience on the project, you are more likely to stick around and retain your voice, even when things are no longer terribly pleasant. But if you've had a bad early experience, you're (of course) more likely to exit immediately, but even if you stick around at first, you'll be more likely to exit later on, because that bad early experience is going to make your underlying connection (or "loyalty") to the project weaker. But since Wikipedia is almost completely insensitive to the exit signal compared to voice, it is precisely those people who had a good initial experience who have had the greatest voice in making that experience worse for those who came later. I guess that's why I see these numbers as such an indictment.
 * That's of course just a hypothesis, and I may be reading a little too much into these numbers. But as a 2004er myself, there are many times when I've asked myself "why am I still here?" I don't suppose I've ever had a terribly convincing answer, but I've stayed around in one way or another -- and I think the initial rush of being meaningfully welcomed into the work by people who were interested in and working on the same things has a lot to with why I keep coming back. If I hadn't had those positive early experiences -- something to be loyal to -- I might have stuck it out for a while, but I don't suppose I'd still be here. -- Visviva (talk) 17:03, 7 August 2022 (UTC)
 * What it means to be welcoming has changed over the years. For example, a lot of Internet communities today use tools like Discord to have real-time conversations. Old-timers on Wikipedia typically are less enthused about these communications methods, and so they'll be present in lower proportions on any channels that get set up, which affects newcomer impressions and the amount of guidance available to them. Editors who have stuck around for many years have better knowledge of the history that has led the project to adopt new guidance. The most obvious one is a strong emphasis on citations for everything now. New editors may not appreciate why guidance has changed from Wikipedia's inception and not stay around long enough to experience first-hand the need for it.
 * There are certainly plenty of areas where Wikipedia could improve, so I do agree that it's important to consider editor motivation. Is there something different about new editors today than editors in the past that make them less likely to adopt Wikipedia as a longer-term hobby? What can we do to improve the environment to help retain editors of all stripes? What are the characteristics of a promising new editor, and how can we recruit suitable new editors? And as I mentioned, if the reality is that we have to deal with a shrinking editor population, how can we make the editing environment for effective? Part of the challenge is that it's been difficult to get even a core working group of editors to participate in prolonged, focused investigation and discussion. This could be an area where a partnership with the WMF and with university researchers would be fruitful. isaacl (talk) 23:28, 7 August 2022 (UTC)
 * I suppose you're right; there are an unlimited number of possible explanations for any given numeric trend. We are all intelligent folks, I suppose, and can come up with an unlimited number of comforting just-so stories. But I will confess I find it difficult to take these alternative explanations very seriously, given that the changes in the community's own practices provide a much more parsimonious explanation. If there are any established editors who truly believe that Wikipedia is not a hostile environment, I would recommend the following simple experiment: create a new account and simply continue editing in good faith and creating new articles under that account, exactly as you normally would. The difference between the treatment you receive then and the treatment you receive under your established account, which will be substantial, is as good a measure as any of Wikipedia's current level of hostility (free of any confounding factors such as generational changes). Any such differential is difficult to excuse in a community that is, by its own very successful design, made up exclusively of randos on the Internet. -- Visviva (talk) 05:31, 8 August 2022 (UTC)
 * We'll just have to agree to disagree; I think the simpler explanation is that Wikipedia has left its land rush phase and has a lot more interactive competition nowadays. isaacl (talk) 20:41, 8 August 2022 (UTC)
 * I tend to agree, although I would also say the danger of numbers is they ignore human factors and realities. Many of the broad or basic subjects have been covered, leaving niche subjects and people who get tired of correcting the same drive-by changes disguised as improvements to established articles over and over again. It's a model which seems by design to create editor fatigue and loss of good will. And you simply cannot mandate good will or welcoming behaviors. I only edit occasionally because I find Wikipedia has become more concerned with its own processes than anything else. It also doesn't acknowledge its own limitations. I would also be wary of a partnership between the AMF (which has its own agenda that often doesn't seem to connect with the community) and university researchers (who also have their own agendas).
 * Just my totally devalued $.02. Intothatdarkness 21:41, 8 August 2022 (UTC)
 * That's why the community should get involved in helping to define research studies, so it can ensure that the objectives align with its interests. Surely there are some editors with links to prestigious university sociology or computer science departments to work on studies the community wants? The WMF involvement is needed to providing funding, since that's what universities are looking for. It can be beneficial for everyone involved.
 * I fall into the post-2005 population (just past the cutoff), so maybe a changeover had already been made from "super welcoming" to "unfriendly", but I've never found Wikipedia's environment particularly enticing for good-faith editors. The effort to deal with unco-operative editors compared with the amount of effort they need to expend has always been a discouraging high ratio. I've heard editors allude to the early days as being incredibly confrontational, though, so it's not clear to me that those were halcyon times. But in any case, I'm not concerned about indicting past behaviour, but examining what can be done to improve processes going forward. How can we design our processes such that desired behaviours are encouraged, and poor ones are losing strategies? If we can do that, selective pressure will automatically reduce the amount of poor behaviour. isaacl (talk) 22:09, 8 August 2022 (UTC)
 * I don't nessecarily agree that all the articles on important subjects have been written. Yes, a lot of them have been, but a lot haven't. A perfect example is a draft I'm working on right now: Draft:Racial segregation in Canada. I was kind of shocked there wasn't an article about this already in 2022. Same goes for Project Surname. I usually focus on Canadian-related topics and even there I see a lack of coverage in important things. I imagine there's a lot of really important topics that aren't covered globally, too. Clover moss  (talk) 19:04, 14 August 2022 (UTC)
 * Importance shifts over time, but the main large topics have been covered to at least some degree. World War II, for example. Your linked articles may well be important in Canada, but might lack that same importance in, say, Senegal. By the same token, there may be a subject considered vitally important in Liberia that doesn't raise any interest in Canada. Those are factors of regional importance, which are not the same as global importance. It doesn't mean the subjects are not important, but that their level of global importance will be different. It also means they will attract more niche editors. Intothatdarkness 20:49, 14 August 2022 (UTC)


 * I have discovered that I made a silly mistake in the above data (I did not correctly restrict it to mainspace), which I am in the process of fixing; will update. -- Visviva (talk) 05:31, 8 August 2022 (UTC)
 * Having corrected that mistake, I find with interest that the 2012 dip disappears, instead, the gradual decline continues through 2014, with an intriguing dead-cat bounce thereafter. Perhaps things have actually been getting slightly less bad? Poking around, it looks like 2012 may have been the year that we really started pushing people into AFC, which would explain the discrepancy; (some of) the users who bounced hard off of AFC would be recorded in the all-namespaces count but not the mainspace count. Of course that also makes me realize that these numbers will heavily understate attrition for users joining in at least the past decade, since they will omit all the poor lost souls who trust the guideposts and walk into AFC, never to be seen again. Quantifying that could be challenging, however. Anyway in lieu of another stacked bar chart, here is a line graph of the corrected data, in case it might be of interest to anyone:

-- Visviva (talk) 17:22, 8 August 2022 (UTC)


 * If anecdotal evidence helps, I think its not just being unfriendly to new editors, but also how policies apply differently to old ones and new ones.
 * I've recently had a situation where an experienced editor made five personal comments in a row on a talk page, which essentially said that I shouldn't be commenting or editing because my account is two months old. (Just to be clear, the rest of their comments' text had no intention to build a consensus, it was purely disruptive.) Yes, its relatively minor, but the result is I have no appetite to participate in that subject area anymore; there are things in life that are more fun.
 * Tolerance towards civility transgressions or uncooperativeness by the established editors really hurts: I get that the consensus of allowing an established editor to have more weight on what happens is not necessarily a bad thing, but I can't help thinking that the relative weight of contributions by such uncooperative editors could've been not as prominent were they not allowed to actively discourage new editors. PaulT2022 (talk) 16:12, 14 August 2022 (UTC)

Being more welcoming to new editors
I think one of the practical issues are rules (not the lengthy policies, but how they apply in practice) being, effectively, kept secret from the new editors.

I've listed what I experienced in the Wikipedia talk:WikiProject_Countering_systemic_bias post, as well as drafted a welcome text that condenses what I wish I would've known User:PaulT2022/Welcome.

One of the commenters suggested that this Wikiproject could be a more appropriate venue, so wanted to share this here as well. PaulT2022 (talk) 16:18, 14 August 2022 (UTC)
 * Thank you Paul for sharing and raising an issue that has been at the forefront of most of the 10 year history of this project. I think you will find friendly ears here for any suggestions on how to improve the treatment of new editors. Perhaps you and others can re-kindle an "Editor Retention" conversation that, admittedly, has become dormant over the past few years. &#8213; Buster7  &#9742;   03:52, 15 August 2022 (UTC)
 * How to improve is a tough question and I suspect if easy answers existed someone would've came up with them.
 * Informing new users about rules in a concise form would help a lot. I could only figure them out with an external advice to get started with BRD etc, and then had to spend a few days reading policies and guidelines. (Because what happened in my case was that someone saw a red username, instantly said "WP:SYNTH" and refused to explain anything.) I don't have an opinion how exactly its best to do so, but the current popular welcome templates are all almost useless.
 * Some nudging would help: i.e. a notification if your edit is reverted for the first time, explaining how to discuss - and encouraging to do so. I've seen new editors jumping into edit wars without apparent bad intent; leaving Wikipedia after first revert is probably a lot more common, judging from the people I know who tried to edit and quit instantly.
 * One thing that seems counter-intuitive to me is lack of voluntary KYC/identification. Similar to how dating apps do it, allowing to voluntarily confirm identity to a third-party provider without disclosing it publicly.
 * The reason why its important is a lot of community effort seems to be fighting all sorts of vandals and undesirable editors creating new accounts. I suspect that acceptance of rough editors by administrators comes from lack of authentication: its impossible to protect articles from bad editing if you can't just revert a bad edit without proper explanation because anyone can make a new account and such edits. I think having verified accounts would've helped a lot to differentiate between likely good-faith editors and less trustworthy ones.
 * The biggest problem IMHO is the ability to engage in confrontation without repercussions, but policing it right now would result in Wikipedia being destroyed by unchecked vandals in a month. PaulT2022 (talk) 05:19, 15 August 2022 (UTC)
 * A problem with all written instructions is that the target audience has to be willing to read them, and "read this before you do anything" instructions are typically not read. I agree though that the one message I would try to get all newcomers to learn is to use article talk pages to discuss changes, either in advance of an edit if you are uncertain, or after being reverted. Presumably once a new editor engages in discussion, they'll have more motivation to follow up on reading any guidance to which they are pointed. I think the assigned mentor for new editors (as part of Growth Team features) gives them a good starting point for a support network. (Currently, due to scalability concerns, only 10% of new editors are shown an assigned mentor.)
 * Jeff Atwood, as part of his thoughts on designing Stack Overflow, wrote about the importance of having a pipeline of new editors, so they can graduate to an intermediate level editor who can help out new editors in turn, and any intermediate editors who may have tired of helping new editors can relinquish that role without regrets. It's kind of a catch-22: we need new editors to stay around long enough to ensure we get more new editors. Alternatively, we need to entice a greater number of experienced editors to continue to be patient with new editors and bring them up to speed.
 * On a side note regarding not having a user page: I haven't had any specific issues because of this. I have not edited any particularly controversial areas, though. isaacl (talk) 06:59, 15 August 2022 (UTC)

What can we tell new anon users who might be hesitant to register about WMF email policy?
Knowing the answer, could help us persuade anon users to register. Your feedback would be appreciated at this discussion at WP:Village pump WMF. Thanks, Mathglot (talk) 04:45, 14 October 2022 (UTC)

How could a WikiProject be more welcoming to new editors?
The WikiProject Vital Articles has a lot of members, but most of them are experienced editors and there isn't a lot of new blood. How should the project be modified to be more welcoming to new editors? CactiStaccingCrane (talk) 12:53, 19 October 2022 (UTC)
 * Do you think the project has failed to be receptive to new participants, such as you when you first started participating? How do you think your experience could have been improved? What is the ideal initial engagement in your view? Given the available resources, what approximation of this ideal do you think could be achieved?
 * As I'm sure you've observed in the social groups you've seen, there are various reasons why newcomers to a group can have difficulty integrating. Do you see evidence of this occurring in the vital articles WikiProject? Specific issues are probably best discussed on the vital article WikiProject talk page. I've participated a little bit on the talk page so I have some exposure to the project. I'm sufficiently experienced though that I don't require anything more than what occurred: some editors responded to me, and we discussed matters. isaacl (talk) 00:56, 20 October 2022 (UTC)
 * I think that the project have failed to attract members, which has significantly stuttered the project. The project ideally should encourage editors to edit vital articles and organize efforts to do so. Here's some of the problems that I think could not be easily fixed just by discussions at the project itself:
 * Even though lots of people know and complain about the list's idiosyncrasies, most are not bothered to so something about it. I suspect this is because of the complex system and unwritten rules at Vital articles about how an article should be swapped/removed/added to the list, and that really drives a lot of people away. That's also the reason why most members of the projects are on Wikipedia for 1 yr or older.
 * People at WT:VITAL imho is really unwelcoming to change and don't want to actually improving the articles, which is kinda crazy considering the whole purpose of the list is to focus editors' efforts. It has gotten bad enough that I've used the the WikiProject's talk page far more often than it should have been.
 * There is no resources about how to write and improve vital articles. Like above, this is because there isn't an organized effort to improve them until now.
 * CactiStaccingCrane (talk) 06:05, 20 October 2022 (UTC)
 * OK, that's a different (though potentially related) issue than the one you initially raised: a group can still be welcoming yet fail to attract new members. But it sounds like you want to change how the project operates, and that will have to be discussed with the project, as projects are just groups of users with common interests who have agreed to work together on initiatives. As with any social group, there isn't a way to compel others to agree with your suggestions. Remember just as no one can decide how you're going to spend your volunteer time on Wikipedia, you can't make those decisions for others, either: you can only rely on the goodwill you've built up to influence other editors.
 * One good way to gain social capital is to work on a related set of tasks that can be done alone, or can be spread out across multiple interested persons. If you can work on that steadily and make tangible improvements, you may be able to get more people interested in helping. Some editors are attracted by progress boards and seeing tasks getting (literally or metaphorically) checked off as they are completed. If you show other project members that you are vested in the project's goals and have contributed to its success, they'll generally be more receptive to your ideas. Good luck! isaacl (talk) 06:44, 20 October 2022 (UTC)

Discussion of opaqueness of core policies
I've stumbled upon the conversation Wikipedia_talk:Verifiability and wanted to share it here as I think the way the discussed core policies are written makes them very hard to understand and follow.

I understand the sentiment A problem with all written instructions is that the target audience has to be willing to read them, and "read this before you do anything" instructions are typically not read expressed by User:Isaacl previously, but I would argue that apparent contradictions (apparently explainable by the scope differences) between WP:ONUS/WP:BURDEN/WP:QUO/WP:NOCON are hard to grasp even for experienced editors. PaulT2022 (talk) 11:42, 14 December 2022 (UTC)
 * For better or worse, English Wikipedia's consensus-based decision-making traditions make it hard to modify policy pages in a substantive manner. We accumulate guidance pages by accretion, with new ones being added and gaining approval independently, without reconciling the effect on previously existing pages. It's also just hard to rigourously define community intent, as each situation has its particular nuances that the participants in a conversation will take into account.
 * Regarding the specific pages to which you referred: generally speaking, is saying that it's not enough for a piece of information to be verifiable to be included in an article. As with everything, if there is a dispute on what to include, the community consensus should be determined.  is saying that article content should be based on appropriate sources.  is behavioural guidance not to edit war.  is the ambiguous policy, regarding what to do if there if a consensus view cannot be determined. By the letter of the policy, if some current content in an article is sourced appropriately but no consensus was achieved to keep it in the article, then it should be removed. But it can be hard to get enough editors to weigh in to determine a broad-based consensus view, and there's no magic way to know when there have been enough editors participating, so those disagreeing with removal can continue with discussion indefinitely. The deference given to the concept of implicit consensus is a concession to the reality that it's operationally impossible to get the whole community to provide their viewpoints on every change. Consensus-based decision making breaks down, though, when the community isn't strongly aligned in goals, and it's hard to know if all those not weighing in are implicitly in favour of the current content or not. isaacl (talk) 17:10, 14 December 2022 (UTC)
 * My impression is Reverting is not followed when overriding considerations of WP:BURDEN/WP:ONUS take place.
 * With regards to the latter, I found an interesting discussion at Village_pump_(policy)/Archive_162, which largely eliminates the concept of implicit consensus/status quo. In the light of this RfC and how policies are applied by editors and admins in practice I think WP:NOCON wording and spirit of the guidelines like WP:BRD seriously misleads editors, provoking edit wars, as the reality is any controversial content is (usually) removed before discussing, new or longstanding, with edit warring or not.
 * See also this note in WP:BURDEN (emphasis mine): "Once an editor has provided any source they believe, in good faith, to be sufficient, then any editor who later removes the material must articulate specific problems that would justify its exclusion from Wikipedia (e.g. why the source is unreliable; the source does not support the claim; undue emphasis; unencyclopedic content; etc.). If necessary, all editors are then expected to help achieve consensus, and any problems with the text or sourcing should be fixed before the material is added back."
 * It's completely open-ended and essentially allows any editor to remove material as long as they believe a policy-based reason to do so exists and this material cannot be added until a new consensus is reached.
 * I don't find By the letter of the policy, if some current content in an article is sourced appropriately but no consensus was achieved to keep it in the article, then it should be removed. to reflect the existing practice (or other policies) at all: removal usually happens first. PaulT2022 (talk) 17:57, 14 December 2022 (UTC)
 * I don't want to get too much into the weeds regarding these specific policies, since there's not much that can be done on this talk page regarding them. Implicit consensus is what holds for unchallenged changes. It's not supposed to hold for challenged changes, but operationally it's understood that specific circumstances have to be taken into account. Regarding removing changes for which no consensus was achieved, I was referring to how this end result is supported by policy. I wasn't referring to what happens during the interim while discussion is occurring.
 * Going back to the broader issue: the key message we need to convey to new editors is that decisions are made by individual discussions amongst the editors who participate in them at the time. Although this theoretically means the process can be very responsive to changes in viewpoints, it also means it can be inconsistent and highly variable based on who participates. Policy, guidelines, and other guidance pages outline the viewpoints upon which these discussions will be based, but they have ambiguities which have to be resolved for each situation. It's much like collaborative work in any other sphere of your life. Some participants will have well-reasoned arguments. Some will have flaws in the lines of reasoning, but often there are still valid aspects of their argument. The only way to move forward is to work with everyone there to reach a common understanding on an approach everyone can live with or produces the least amount of dissatisfaction. isaacl (talk) 19:51, 14 December 2022 (UTC)
 * Thank you! I raised it on the Reverting talk page, as I think WP:QUO creates a false expectation in this regard and somewhat masks that It's not supposed to hold for challenged changes as you're rightly saying. PaulT2022 (talk) 20:28, 14 December 2022 (UTC)
 * As it says at, no one can be compelled to follow the advice. I appreciate why people spend time arguing over the interim state (*), but often it's a waste of time. The point of avoiding reverting during discussion is stop spending time on the current included content, no matter who reverted first or if the status quo is no longer in place. Everyone's time is better spent discussing the final state of the article. (And sometimes challenges are without justification, so I don't feel a hard rule to "remove the material immediately on challenge" is the best way forward.)
 * (*) If there aren't enough people willing to weigh in and continue to participate throughout the entire discussion, the interim state can become the de facto final state. As dissatisfying as that is, on the whole it's still undesirable to spend time edit warring over the interim state. isaacl (talk) 21:36, 14 December 2022 (UTC)
 * Well said! I wish these words would've been in WP:QUO. Would you like to add them? Much more useful advice than what the last paragraph of it says now. (First implying that not favoring status quo is edit warring and then reminding that restoring status quo is edit warring.) PaulT2022 (talk) 21:50, 14 December 2022 (UTC)
 * I don't have a strong viewpoint on the last paragraph so I'm not planning to make a proposal to change it. (On a side note, personally I don't feel that saying action X is one way to prevent edit warring implies not-X is edit warring, and the last sentence says edit warring is editing warring, no matter the reason why.) Although naturally I like the view from which I see things, I'm not convinced it's any more illuminating than the existing wording. Beyond the broad strokes, there is so much variation in actual situations, making it hard to draw specific lines. Biographies of living persons is one of the prime exceptions: editors will be much quicker to reset a biography to a state that avoids making a disputed claim about the subject. The desire to avoid having inaccurate info appear at the top of search engine results motivates editors to establish a specific interim state. isaacl (talk) 23:23, 14 December 2022 (UTC)

ANI improvement stuff
I think that people in this WikiProject might want to take a look at my thread about making ANI less toxic here: Wikipedia talk:Administrators' noticeboard/Archive 16. CactiStaccingCrane (talk) 16:58, 24 December 2022 (UTC)

Feedback requested on retaining/regaining Wiki Ed student editors
The Education program manages a platform which organizes thousands of university students who contribute to Wikipedia as part of their regularly assigned university coursework. (Sample Wiki Ed course page here.) These Wiki Ed student editors typically contribute for several weeks during the second half of their semester course, and then disappear, no doubt busy with college ccourses, graduation, and real life. I've made a proposal at how Wikipedia might attempt to regain some of these former student editors at a later time, perhaps after graduation. Your feedback would be welcome at WP:ENB. Thanks, Mathglot (talk) 22:57, 30 March 2023 (UTC)

WikiProjects and editor retention
It seems like most WikiProjects have essentially become inactive over the years. I was wondering if this declining participation in wikiprojects has had an impact on editor retention? I think it's likely just because places where people can easily collaborate and improve content together seems to be pretty much the ethos of the project. Clovermoss 🍀 (talk) 21:22, 12 March 2023 (UTC)
 * WikiProjects that are successful and have longevity pertain to specific topic areas that can attract a focused group of like-minded editors to work as a team to improve articles, usually thru discussion and editing. Editor Retention's focus was on the editors that worked in the Wiki Arena. Early on, its goal was to act as a forum where relevant issues about the act of editing might be discussed, where guidance and support might be available. And where recognition might be dispensed. It was more about actions rather than editing. Molding good acceptable behavior and resisting controversy and contention was its specific topic rather than Sports or National Parks. Eventually controversy and contention took over. In the end recognizing good performance was all that was left as a tool for retaining editors and, now, has become an end onto itself. &#8213; Buster7  &#9742;   01:27, 13 March 2023 (UTC)
 * I don't know if anyone has studied it, so not sure if there are any data-based views that can be provided. It might be interesting to ask participants in formerly active WikiProjects for their views on why they have become less active. For two of the sports-related WikiProjects I follow, I think their talk pages are less active now because much of the specific guidance that many editors are interested in has been established. The remaining guidance to be determined is either contentious, and so there is no agreement to be found, or there aren't enough editors interested in that specific area to establish agreement. I'm just guessing, but I think there may be fewer editors editing in those areas because player articles have been written for most of the players where it's relatively easy to do so. I don't think there is a causal link to the activity on the WikiProject; if anything, I think it goes the other way around. isaacl (talk) 01:47, 13 March 2023 (UTC)
 * The main reason I was asking this question is because I was thinking that suggesting wikiprojects that fit with a new editor's interests may be encouraging. But at the same time, if a substantial percentage of these wikiprojects are inactive to begin with, maybe that path of action would actually be a bit disheartening and not really have the impact I was hoping for. Clovermoss 🍀  (talk) 02:00, 13 March 2023 (UTC)
 * Wikipedia talk:WikiProject Editor Retention/Archive 28 is a good read on what the intention of Editor Retention was.
 * &#8213; Buster7  &#9742;   02:00, 13 March 2023 (UTC)
 * As long as you're suggesting active WikiProjects, I think it's a good idea. It can help with building a network. isaacl (talk) 02:06, 13 March 2023 (UTC)
 * Any suggestions for reviving inactive wikiprojects while we're on the subject? Not for new editors in this case but because there's at least two wikiprojects that I'd like to be more lively. Clovermoss 🍀  (talk) 02:14, 13 March 2023 (UTC)
 * WikiProjects are just groups of editors sharing a common interest that are using a WikiProject page to co-ordinate. So I'd look at it from the opposite point of view. If you find some editors with a common interest that you share and that need to co-ordinate on some initiative or to tailor some guidance to the specific needs of the area in question, try co-ordinating amongst yourselves on an appropriate WikiProject talk page. isaacl (talk) 06:33, 13 March 2023 (UTC)
 * It should be said that there is also some animosity toward the very concept of WikiProjects as inherently encouraging "canvassing." Gnomingstuff (talk) 05:05, 4 April 2023 (UTC)

Exit interviews – obtaining first-hand testimony
I think it would be useful to gather information from departing Wikipedians whenever possible, to see what we can learn from their experience here, and from their decision to leave, if they're willing to share that, and what we might have done differently that might have resulted in their staying on. Also, any thoughts on what they think we could do better in the future, in order to retain editors going forward.

Recently I had some good interactions with, who I regret to say has decided to move on as of now (although we'd love to see them back), and they have graciously agreed to help us before they disengage, by adding their thoughts here about these issues. (See previous discussion at their UTP.)

Thanks very much for agreeing to respond here, User:Antimoany, that's really very kind of you, and I hope and believe it will be helpful to this project. I'm not sure how to start, other than to just ask you, what made you decide to stop editing Wikipedia, and can you name a couple of things that we could work on to do better? (And don't feel shy about changing your mind and coming back; ultimately, that's the result we'd like to see.) Thanks, Mathglot (talk) 08:47, 24 April 2023 (UTC)


 * hello! i do agree exit interviews are a good idea for a project like this, though i imagine most users don't declare they're leaving and then hang around to see the invitation to one.


 * quick context: while i'm quite new to wikipedia, i'm familiar with the markup thanks to my participation various (often relatively small-scale) projects that have copied it. i'm not familiar with how wikipedia specifically does things, especially in an organisational sense.


 * when i first started editing, it was in response to the various issue banners i've seen strewn about. well, my very first edit was in response to a friend complaining about the state of The Mask (comics), where i removed one particularly out-of-place paragraph. i wanted to clean the page up further, but wasn't sure how much it would have been acceptable for me to remove, so i left it at that.


 * but all my other edits were in direct response to issue banners, including my comment on that article's talk page.


 * my response to issue banners worked on the assumption that these could only be put into place (and later removed) by specially authorised accounts. on banner blindness (let's pause to appreciate the irony), i did tentatively remove such a banner after i felt i'd eliminated the issue it warned of. i did so with the expectation that a bot would be along shortly to put it back, as i felt certain i wasn't authorised to remove it. i wanted to see how wikipedia handled users taking actions they aren't permitted to.


 * even after learning that banner removal appeared to be something anyone could do, i still assumed there was greater structure at play. perhaps, i thought, edits that remove such banners are patrolled, and my removal of it alerted a human who would then check to confirm the issue was resolved.


 * it didn't even occur to me to try placing a banner myself to see what would happen. i didn't consider and then reject the idea, it never entered my mind. so assured was i that there was a system in place policing their addition and removal.


 * i did then gain some confidence in attempting to remove issue banners and took a couple off pages i felt didn't need them. i continued to assume someone was invisibly checking these removals.


 * i was, of course, wrong. i discovered through a talk page that anyone can place any issue banner on any article. the system is not patrolled, neither by humans nor bots. there are no checks, automatic or otherwise, to ensure that editors well-versed in the issue the banner describes have agreed the article suffers from it (and later, agree it is resolved).


 * i assumed these checks were in place, and my confidence in editing came entirely from that assumption.


 * should i have made some effort to confirm these assumptions? probably. i'm so familiar with heirarchical structures in volunteer projects like this that my assumptions were so well-rooted i never had the presence of mind to question them.


 * to repeat something i said on my talk page:


 * a user doesn't need to have been granted permissions to add a page to any given "needs work" category, or to remove it. anyone can do that, and they can be wrong.


 * i can't handle a system where edits have to be justified after the fact. i need the confidence that what i'm trying to do is something that someone should be trying to do. that someone with authority has decided is needed.


 * i expected a simple, linear, "job needed >> job done >> job confirmed" structure. without that, i can't contribute. i have too much anxiety.


 * emphasis on "i have too much anxiety". in projects like this, i rely on others with more knowledge (and, honestly, more energy for beurocratic discussion) to make the big decisions. i need someone else to say, "this article is too long, make it smaller" or, "this article is too short, add more detail". and i need them to, if not be right, then at least have had some significant authority to make that decision.


 * i haven't been blindly following these banners. i've looked at quite a lot of articles in the backlog and either struggled to see the issue, struggled to understand the issue, or found that while i agreed the issue was definitely in place, i lacked the skills to resolve it. i edited pages where i felt i understood the problem, and had at least some idea of how to fix it.


 * i wanted to say more, but this is already much longer than i expected and i think i'm repeating myself. i struggle to decide which information to include or omit in situations like this.


 * i hope this was helpful. Antimoany (talk) 11:30, 24 April 2023 (UTC)


 * got lost in the weeds of describing my own thought processes and completely forgot to answer this question:


 * and can you name a couple of things that we could work on to do better?


 * with specific respect to my need for structure and heirarchy in order to have confidence editing, i'm not sure. my main issue here is anxiety, and that's not something wikipedia can or should try to alleviate with structural change.


 * my working assumption before i learned otherwise was that adding or removing any issue banners from a page would alert a patrol group, who would then check that the banner either belonged or didn't, as appropriate.


 * i don't know how feasible it would be to create such a patrol group, or whether it would be possible to populate it with sufficient active patrollers (i'm well aware of issues of user inactivity in projects like this, especially boring or stressful jobs like authorisation patrols)


 * i think something like "issue banner patrol" might be worth trying. i'm wary of suggesting it, because i'm aware of a lot of potential issues if such a system were introduced, such as:


 * * difficulty finding enough users with the needed knowledge, and skills, and time, and willingness to do this patrol.
 * * patroller inactivity
 * * false sense of security: if there aren't enough patrollers, or if they're selective in which edits they look in on (very likely, not everyone understands every possible issue), a lot of banner-related changes might go entirely unpatrolled. everyone would assume someone's looking, but no-one is.
 * * patrollers would likely have to get involved in various discussions on talk pages, which can be time-consuming and stressful.


 * i'm learning more and more that wikipedia has infinitely less organisational structure than i ever imagined, yet also several interconnected labyrinths of beurocracy. the beurocracy i expected, but i expected that it came with far more structure than seems to be in place, and i didn't expect nearly as much beurocracy as seems to be in place.


 * i am deeply stressed by the kind of beurocratic discussions that happen on talk pages (the ones attached to encyclopedia articles, not ones like these). wikipedia's talk pages are like staring into the abyss. i usually avoid looking at them.


 * is that a personal issue? probably. it's also intrinsically connected to my need for structure. for someone else to participate in this beurocracy so i don't have to. i'm sure i'm not alone in this. i'm probably not even the first to say it.


 * on a matter not directly related to issue banners, i asked for advice on the Teahouse page (which i found a link to...somewhere within wikipedia's laberynthine back-of-house pages) about how to respond to two conflicting instructions:


 * an issue banner attached to an image file declared it should be transcribed as text into its included articles. the only included article was an archived talk page, with a large notice not to edit it.


 * my question was: which of these instructions should i respond to? how should i respond?


 * the answer i got only raised further questions. when i asked for clarification, the response was so confusing i had to ask a friend what it meant. my friend observed that it was "a very wordy shrug", which i admit was roughly my impression too.


 * i'm not upset at the user who tried to help me on the teahouse page. there's a reason i haven't linked directly to the discussion or named them. i do sincerely believe they were trying to help. unfortunately, i found their responses unhelpful, and i didn't want to frustrate them by asking for additional clarification a second time.


 * i think there were two separate issues at play here:


 * issue one: regarding my request for help.


 * i find pages like the teahouse suffer from what i call "solved-problem avoidance". as an active participate in help forums, i find i often suffer from this too and sometimes have to actively force myself to avoid it.


 * i define "solved-problem avoidance" as this: a question asking for help has at least one response that, at first glance, doesn't look totally worthless. the problem is solved, or if the discussion appears to be ongoing, the user who first replied is still in the process of solving it.


 * it makes sense. there's no need for 12 people to all give roughly the same advice with different verbiage.


 * but, this can often prevent someone from getting an answer that is helpful to them.


 * issue two: regarding the reason i asked for help.


 * i've called wikipedia's back-of-house pages labyrinthine twice. i'm going for thrice. it's a mess. i find most help articles by googling "wikipedia [hopefully the right keywords]". most of the time, especially for simple issues like formatting a table or citation, this gets me to the page i need to be on. i find wikipedia's own search not-great, but this isn't about search engines specifically. also searching isn't very helpful for more complex questions anyway.


 * there aren't any simple, easy-to-read indexes of stuff an editor would want to find. i can find myself reading a very lengthy article about some policy or rule or something, and not understanding most of it. most of it isn't even related to the reason i went to that particular page.


 * i appreciate that a page about a given policy has to be thorough, but often i find this thoroughness comes at the cost of...not explaining the policy very well.


 * a lot of policy pages have a small box at the top, "this page in a nutshell".


 * i find "in a nutshell" too oversimplified and utterly unhelpful, but the entire article too long and difficult to understand.


 * furthermore, that linked policy page doesn't seem to contain a list of all policies, which is what i expected to find. there's a section at the bottom that appears to list some policies, but i'm pretty sure there are a lot of rules that aren't on there. it does contain a list to a different policies page but that likewise does not seem to hold a full list of policies.


 * ironically, the policies and guidelines page contains some guidelines on what a policy page should look like. in my opinion, most policy pages fail even the first step, "Be clear".


 * i'm not really sure what change to suggest here, as i'm sure the policy pages are in their exact state for good reason and after protracted discussion. i don't necessarily think anything included in a given policy page should be removed. but, my difficulty reading the policy pages has led to me not understanding most policies. i can't find a policy page about "what to do when two policies conflict". i can't find a list of all policies.


 * my inability to understand wikipedia's various policies significantly impacted my decisions on which backlog pages to look at, and subsequently which articles to even ask myself if i can improve.


 * i'm sorry this is so long and probably unnecessarily so, especially when i'm now complaining about unnecessary length. Antimoany (talk) 13:03, 24 April 2023 (UTC)
 * Thanks for sharing Antimoany. The door is always open. Some other reasons can be found at WikiProject Editor Retention and WikiProject Editor Retention/Discovered reasons given for leaving Wikipedia. &#8213; Buster7   &#9742;   14:14, 24 April 2023 (UTC)
 * i can't handle a system where edits have to be justified after the fact. i need the confidence that what i'm trying to do is something that someone should be trying to do. that someone with authority has decided is needed. – I think it's a genuine cultural problem, mostly affecting contentious topics tbh, where challenging other editors is much more prevalent than a constructive talk page discussion.
 * It's seen as a form of reaching consensus, which it is, but I think such narrow view ignores the unfortunate result that instead of creating encyclopedic content, editors tend to focus on whether a certain POV is represented and how it's represented instead of considering how the resulting article reads. If you've seen any articles that have a lot of seemingly disconnected facts but lack the bigger picture completely, that's how they're usually written.
 * (I've mostly quit editing certain topics because of this.)
 * i'm not really sure what change to suggest here, as i'm sure the policy pages are in their exact state for good reason and after protracted discussion. – Sometimes there is no good reason. Once conflicting statements are in the policies, it becomes impossible to reach consensus to align them if the community is split. Sometimes it's just poorly worded. I've managed to prompt Consensus policy edits for something not really controversial but causing problems.
 * I think there are somewhat misaligned incentives, where the better command of informal interpretation of arcane policies one has, the better leverage one gets in the discussions. My impression is there's effectively a consensus for more ambiguous policies being preferred as they can be flexed depending on the requirements of the individual article or discussion. PaulT2022 (talk) 18:21, 24 April 2023 (UTC)


 * i feel this explains a lot about the issues i've had with wikipedia as an encyclopedia. i've seen a lot of that disconnected, fragmented content you mention that seems to be missing chunks or specifically focused on details and not really covering the overall subject well.


 * the prefence for ambiguous policies is odd but i suppose understandable. if the rules are vague you don't get a "no man of woman born" situation. unfortunately, that does make them harder to understand or follow until this "better command of informal interpretation of arcane policies" happens. Antimoany (talk) 00:46, 26 April 2023 (UTC)
 * Thanks, User:Antimoany, much appreiated. Mathglot (talk) 05:49, 27 April 2023 (UTC)

Graphs for different wikis
Some of you might be interested in this tool: https://retention.toolforge.org/enwiki by @Danilo.mac. Whatamidoing (WMF) (talk) 20:19, 28 May 2023 (UTC)
 * Thanks for the link. What immediately stands out is that (roughly speaking) 2006 is an outlier, with editors who started in that year keeping up their editing activities for longer. Given that was English Wikipedia's peak activity period in terms of articles created per month, that seems to make sense. isaacl (talk) 21:10, 28 May 2023 (UTC)
 * There is an interesting data we can get from that graph and the graphs of users with at least 10 and at least 100 edits in the month, those show only users with a more robust collaboration, and we can see something happened in march 2020 that significantly increased the user retention and kept that higher user retention for some months. That was the pandemic effect. Now the user retention is lower than it was during the pandemic, but it seems to be still a little higher than it was before the pandemic. Maybe there is something besides the pandemic that increased the user retention. That is the reason I created that tool, we can theorize about what impact the user retention. Danilo.mac (talk) 22:57, 28 May 2023 (UTC)
 * Very interesting data! I noticed the pandemic bump as well. I'm curious if it coincided with any initiatives, perhaps from Growth. I know that the revamp of Welcome was April 2020, so perhaps that had a small effect. &#123;{u&#124; Sdkb  }&#125;  talk 00:58, 29 May 2023 (UTC)
 * I don't think Growth or other initiative that was also applied in other wikis in the last years had a significant impact, because other wikis, for some reason I don't know, didn't have the same impact from the pandemic like enwiki did, and they don't show the little increase in the user retention like enwiki does when compared to before pandemic. For example, we can see frwiki and dewiki had very little impact from the pandemic in the user retention, and they don't seem to have any increase after the pademic. The Growth was applied in those wikis too, so if that have had some significant impact in those wikis the pandemic effect would be too small there to hide that impact. The Growth team probably won't like my analysis, but that is what I can see in the data. I think that little increase in enwiki user retention can be some kind of long-lasting pandemic effect or something that happened only in this wiki. Danilo.mac (talk) 01:34, 30 May 2023 (UTC)

List of previous conversations
Regarding [//en.wikipedia.org/w/index.php?title=Wikipedia_talk:WikiProject_Editor_Retention&diff=prev&oldid=1157697197 this edit]: Buster7 created the section currently at the top of the talk page with the heading starting with "Previous conversations about newbies..." It was gradually expanded by multiple edits made by Buster7. I don't see a lot of value in trying to sign all of the additions that were made, particularly since the section was intended as a reference list. I feel, though, it is misleading to attribute the section to Legobot, which simply fixed some obsolete HTML markup within the section. Thus I propose removing the attribution to Legobot. isaacl (talk) 11:27, 30 May 2023 (UTC)
 * If there are no comments, I plan to proceed with the proposed change. isaacl (talk) 23:42, 5 June 2023 (UTC)
 * All Clear! &#8213; Buster7  &#9742;   02:10, 6 June 2023 (UTC)
 * I have implemented the proposal. isaacl (talk) 01:15, 9 June 2023 (UTC)

Many blocks shouldn't be indef
There is a discussion that may be of your interest in Administrators' noticeboard. Regards, -- Thinker78  (talk) 04:46, 1 October 2023 (UTC)


 * Indefinite time blocks, are tyranic. Kamil Hasenfeller (talk) 12:24, 22 October 2023 (UTC)

scope should expand to retain Administrators
Administrator cadre continues to contract rootsmusic (talk) 23:22, 1 November 2023 (UTC)
 * As there are a lot of other places where users discuss the state of the administrator corps, this project has historically focused on new editor retention. If you have specific initiatives in mind, you can raise them, but given the relatively low activity on this page, you might garner more responses in another venue. isaacl (talk) 23:55, 1 November 2023 (UTC)
 * Good @Isaacl, I'm not aware of any place where users discuss the state of the administrator corps (though I suppose administrators have their internal discussions). I wasn't even aware of this WikiProject, because it's not listed in WikiProject Council's Directory. rootsmusic (talk) 00:10, 2 November 2023 (UTC)
 * The current Signpost issue, for example, has a discussion on the decrease in active admnistrators. The topic is often discussed at Wikipedia talk:Requests for adminship (there are threads right now). isaacl (talk) 00:23, 2 November 2023 (UTC)

Helpful collection of reflections from experienced editors
See User:Clovermoss/Editor reflections. I feel like many of us understand well what most newcomers are like, and what most experienced editors are like, but less what newcomers who will ultimately become experienced editors are like. This page gives some insights on that. &#123;{u&#124; Sdkb  }&#125;  talk 18:43, 4 December 2023 (UTC)

Turkey master degree students
I think issues like this may be of interest to participants of the editor retention project. -- Dustfreeworld (talk) 19:22, 8 January 2024 (UTC)

Nomination for deletion of Template:Welcome-turnaround
Template:Welcome-turnaround has been nominated for deletion. You are invited to comment on the discussion at the entry on the Templates for discussion page. PleaseStand (talk) 22:16, 1 February 2024 (UTC)