That's a wrap on the November 2024 drive! Thanks to all participants for their help; details on progress and benchmarks reached can be found on the drive page. Help is still needed on reviewing, so please sign up if you are interested! This update covers the changes since October so the totals are a little different.
Headline: We cleared 10,489 articles and are now officially below 72,000! For yourself and your fellow editors, please clap.
Minutiae: For anyone interested in a more detailed breakdown of the numbers - average was 114.8 articles; median 40; mode 39.
Highlights: the entire year 2008 is in the dustbin of history, along with January and February 2009! With the heroic work of editor in arms, @Turtlecrown, the stubborn, German river-centric category of September 2020 saw a 25.3% drop. And everyone's littlest friend, December 2023, further decreased from 48 to 37, a 22.9% decline.
Low-hanging fruit: Other than December '23, June 2024 is a skinny 115 articles, dangling precariously like a sinner in the hands of an angry God. Give it a nudge.
High-hanging fruit: Everyone's favourite BFC (Big Friendly Category), December 2009, is a lean, mean 9,568 articles as of this writing - bringing all categories to under 5 digits. The other high-hanging fruit are, still, the Frustrating Five (name open for revision): January 2013 (1,039), April 2019 (875), May 2019 (1,822), June 2019 (3,915), and September 2020 (1,020). This time, January 2013 had the lowest percentage of change between updates. Godspeed to anyone working on these.
Challenge results: February 2024 beat January 2024 264 to 270, and December 2015 sneaked a win from January 2016, 286 to 289. Live sports are just riveting, aren't they.
New challenge: No ties this time. Have a wonderful end of 2024 and I hope to see you all in 2025!
Announcements: Something discussed prior to the drive was building a regular, semiannual schedule for drives, which would put us on target for May 2025. I see this as a reasonable amount of time to prepare and make adjustments based on editor feedback about the strengths, weaknesses, and areas of opportunity for improvement. However, I see no reason to rush this as we are still tallying this drive's results. If this date is amenable to people, we can get started on a draft page in January. I would especially like to address the points raised by @Joe Roe about the prioritization of inline citations and how we can address this in the future.
I had a great time working on citations during this drive. Thanks to everyone working to organize these events. See you all in the field between now and the next drive! I will defeat December 2023!! Gnisacc (talk) 19:03, 5 December 2024 (UTC)[reply]
I take it back. Let's have as many backlog drives as we can in 2025 and clear the backlog. Why not schedule 3–4 next year? February, May, August, November. Cielquiparle (talk) 13:09, 7 December 2024 (UTC)[reply]
@Cielquiparle - I LOVE this attitude. Personally I'm open to it but I wonder if, by having (too) many drives, we run the risk of diminishing returns via lower editor engagement/participation. We cleared fewer articles in the November compared to the February drive but we also had a much larger backlog. February started with 111,643 and cleared 14,300 (12.8%); November started with 80,645 and cleared 8,511 (10.5%) - not much of a difference proportion-wise, plus we are probably clearing out the easier articles sooner and so what's left are the tougher nuts to crack. My concern is that if we did a drive every few months, would the proportion drop to more like 5-6%? How to do we make sure to encourage and sustain engagement throughout the drive? We definitely lost momentum by the halfway point of both drives, so I think making sure that we keep that up (somehow) for future drives is a concern to prioritize and I'm worried that might be difficult if we do 4+ drives a year. I think it's doable, but we have to be smart about it. Setting smaller goals that are likely to be met through the drive to keep momentum up is probably a good idea; at the same time, we can't have a party every time we go down 100 articles.
Maybe this is the point when we start organizing off-wiki. I also agree with your point that May might not be a good idea for exam reasons (excellent foresight!). Personally April is a wash for me but I could be game for June. Cheers, Kazamzam (talk) 16:52, 8 December 2024 (UTC)[reply]
Updates like these are the type of stats I love to see and make these things much more fun to participate in. Any chance that you have stats on the biggest decrease for categories by topic? Wozal (talk) 16:18, 8 December 2024 (UTC)[reply]
@Wozal - I wish :/ the way that the search function assigns categories by "topic" is pretty hit or miss. For example, if you try to filter the All articles lacking sources category for Biography, the first five results (by "relevance"): Fourth Dynasty of Egypt family tree, List of monarchs of Pontus, Single suiter, Quantitative notrump bids, and Entry-shifting squeeze. Not until you get past the first 20 hits do you get an actual biographical article of an actual person (Avgust Tsivolko, for anyone curious). So providing the stats on decrease by topic, let alone 200+ categories by topic, would be an incredibly time-consuming task. But this raises the larger issue of how the articletopic search function is basically useless; I think this is detrimental to potential WP members who might be really passionate about, say, Ancient Roman history or wetlands or Judaism or whatever but get frustrated and turned off by this filter that is supposed to be helpful. It's something I'm hoping we can address before the next drive, either to improve or replace it with something better.
I enjoyed the update very much, and always do. Thanks for all the hard work. And well done on the backlog drive, everyone!
My two cents on inline citations is that they make sense as a backlog drive task because of a need for what I'll call "meta-verifiability" - the ability to check each other's work and see if it does what it says it does. I've also seen a fair few articles in the wild that use only general references in which the article topic is not mentioned, and hunting down these resources, where possible, is time-consuming, and I don't want to assume intent but it can make content seem more reliable than it is. Maybe it pays in general to emphasise the difference between {{no footnotes}} and {{unreferenced}} though. Turtlecrown (talk) 23:10, 8 December 2024 (UTC)[reply]
Kazamzam thank you for the update. As a new editor, I really enjoyed working on the drive, and I learned a lot of things by reading through and adding references to a fairly broad selection of articles about different topics. I was always pleasantly surprised when it turned out that there was an interesting twist on an otherwise unprepossessing-looking article, and it was very satisfactory to be able to keep them in the encyclopedia. On the flip side, I was rather surprised to come across a few dreadful articles of almost no merit that had hung around for a long time; it was equally as satisfying to get those into the various deletion processes and bid farewell to them.
I was particularly pleased that we reduced the known backlog of unreferenced BLPs to zero, and I firmly think that, in time, (and maybe it is not as far off as we might fear), we can do it with the general unreferenced article base too.
On the issue of how the articletopic search function is basically useless, I note that our sibling project Unreferenced BLP Rescue have set up various Petscan searches to do this for Unreferenced BLPs. Maybe that is an approach that we could consider adopting, and I would be happy to do what I can to help set it up.
Would there be interest in organizing a year-round "marathon" for Unreferenced articles in 2025? It could run in parallel with the June 2025 Backlog drive (or take a break in June).
What would make it different from general project participation would be: We'd have a leaderboard and tally – just so we have visibility into volunteers who are active and what they have been up to. I think it would just help our loose community of year-round contributors feel a bit more plugged in...so you're not just having to constantly check Bambots to work out what articles are getting referenced.
It might also encourage more discussion or coordinated efforts to focus on specific categories or topics. Or is it too much overhead to set up and run...? Cielquiparle (talk) 04:02, 13 December 2024 (UTC)[reply]
I would definitely be game for that. Seeing monthly updates would be great motivation. It would raise the visibility of the project as well, especially if we are able to bring back the hashtags to include in edit descriptions. JTtheOG (talk) 07:04, 13 December 2024 (UTC)[reply]
@JTtheOG @Cielquiparle - the edit description that I've been using for a few years now is the following: WikiProject Unreferenced articles; you can help!
I think it's pretty straightforward and to the point, plus it directs people to the project webpage. We could make something specific for the "marathon" but I would want something that casual editors can use to go directly to the URA page rather than just a hashtag for points tracking purposes. There's also a talk page banner that someone created way back when but I don't think it gets a lot of use - maybe we want to change that? Cheers, Kazamzam (talk) 18:20, 15 December 2024 (UTC)[reply]
It would definitely be helpful for me. I wasn't able to participate as much as I would have liked in the recent drive because of other commitments. So a more generalized tracker that I can do any time would be more beneficial. SilverserenC17:31, 13 December 2024 (UTC)[reply]
I like the idea of a year long marathon, with a leaderboard etc. I also saw that the Guild of Copy Editors hold week-long blitzes which might be worth considering as a complementary activity, for example based on a theme or other subset of unreferenced articles. Maybe we could more explicitly tie those in with other WikiProjects, for example a blitz on unreferenced women's biographies in tandem with WP:WIRED. Cheers, SunloungerFrog (talk) 08:24, 14 December 2024 (UTC)[reply]
@ARandomName123 Could the monthly leaderboard exist somewhere on the main URA page so as not to detract from casual participants (but also encourage such participants to sign up)? Or would it need to be separate? Cielquiparle (talk) 20:48, 16 December 2024 (UTC)[reply]
@Cielquiparle: Please see User:ARandomBot321/URAleaderboard, which should be counting every edit as of the new year (Jan1, 2025, 00:00 UTC). It currently updates hourly, assuming my code doesn't break. You will have to sort by number to see who has the most, but I can try to get that to be the default later.
That's awesome! Particularly removing the need for colleagues to sign up to anything. I had three minor comments:
I note that some of my scores at least just hit on removal of tag even when I didn't add a source, because when I looked at the article it already had a bunch of references. For example, this one: Special:diff/1266851923 (see edit summary). So I am being rewarded for not doing anything much. But I imagine that it would be a lot trickier to amend the logic to look for an actual added reference....
There might need to be a little de-duping of results - for some reason one edit of mine appeared twice in the list: Special:diff/1266877231
Could you also hit on removal of {{BLP unreferenced}} as well as {{unreferenced}}?
Yes, you'd be right. The way it currently works is by monitoring https://stream.wikimedia.org/v2/stream/recentchange for any changes to Category:All articles lacking sources. To check if the edit added an actual reference, I'd have to get the diff and check if any links or ref tags have been added. Regardless, I think it's helpful to track the net change to the category as well.
Oh, I forgot about that, thanks. Should be removed in the next update. iirc, I ran into some issues which caused the tracker to stop for a bit. To make sure I didn't lose anything, I rewound it to the timestamp of the last tracked edit, which happened to be yours. That caused it to count yours twice: once in the initial track, and again when I rewound it.
That makes sense. I doubt anyone is that excited to look at all the monthly diffs for each person. But it's good to have it there as a reference.
Yes, resetting yearly sounds good... Though I was kind of hoping the "totally unreferenced" articles would be extinct within 2 years or less? (We should take bets!) Cielquiparle (talk) 04:36, 29 January 2025 (UTC)[reply]
I was kinda hoping we'd be done in time for the 25th anniversary of Wikipedia, but my current prediction is sometime summer 2026 (May?). We are on track for completing it within 2 years though. ARandomName123 (talk)Ping me!04:45, 29 January 2025 (UTC)[reply]
@ARandomName123 @Cielquiparle not to get too down in the mouth (do people still say that?) but I don't think we'll ever be "done". Maybe we can get it down to a one-year backlog and call it good, but we should start thinking about what a "completed" project looks like for measuring if/when we can say hey we did it. Or maybe a number instead of a time period, i.e. a backlog under 100 or 1,000? Kazamzam (talk) 16:10, 29 January 2025 (UTC)[reply]
Well yes, but I think hitting zero at some point would be a pretty nice symbolic achievement, even though it's probably going to jump back up within minutes (like how we temporarily cleared Category:All unreferenced BLPs). I would say a completed project would have a stable article count. We are currently removing more from the category than adding, so once we hit the point where they're about equal, that's probably the limit.
On a side note, we should probably have a discussion at some point (maybe when we're under 10k or hit zero once?) on bot-tagging untagged articles. ARandomName123 (talk)Ping me!16:21, 29 January 2025 (UTC)[reply]
@Cielquiparle: On second thought, we might have to organize it differently, or add requirements. Currently, the January leaderboard has about 570 unique users. If we assume another 300 unique users each month (based off of ~380 entries with only one diff), by the time it's the end of December, this table will be tracking 3870 users, meaning it'll have 3870 rows. ARandomName123 (talk)Ping me!16:39, 29 January 2025 (UTC)[reply]
Hi all! I was doing some work on January 2010 and found this article, Puccio Pucci (politician), which was edited as part of our November drive. I'm concerned about two things in this edit - one, the use of Geni as a reference and two, the lack of a references section. I'm less concerned about the points of the drive and more concerned about ensuring that what we do as a Project is a net positive - not that this article is worse off for the edits, but imho it wasn't really improved. I'm wondering if it's possible to make a filter for reviewers to ensure that people are not getting points for adding "references" from sites like Geni, Goodreads, IMDb, etc. I think we can work this into the next drive (June 2025?) but I'm wondering if it's a bot-problem or something that needs human eyes and reviewers. Thanks, Kazamzam (talk) 16:40, 18 December 2024 (UTC)[reply]
@Kazamzam: Do you mean something like a bot that checks all submissions against RSP (or whatever list we decide), and provides a feed of only potentially-problematic references for reviewers to look through? ARandomName123 (talk)Ping me!20:35, 18 December 2024 (UTC)[reply]
Well, I could modify the review tool to highlight or arrange separately edits with unreliable sources, though it might take the tool longer to load. Alternatively, we could just have a bot periodically post a list somewhere on-wiki. ARandomName123 (talk)Ping me!23:20, 18 December 2024 (UTC)[reply]
@Kazamzam: Another suggestion, if this has not already been addressed, would be clear and well-publicized reminders about Wikipedia policies on WP:UGC. I edit many biography articles, and many of these articles still have citations to genealogy web sites and Find-A-Grave in support of facts about birth and death. I don't know how many active editors still cite UGC, but with the prevalence of UGC cited, it's easy to see how many editors would perceive this as an editing norm. Thanks to everyone herding us cats. Best regards, Oldsanfelipe2 (talk) 22:58, 18 December 2024 (UTC)[reply]
Does anyone know if there a project oriented around removing citations of UGC? (quick search only turned up an AI cleanup project). I could be interested in that. Relatedly, I know that the editor will stop you from using some banned urls and such, does that not catch Imdb/geni/goodreads? Gnisacc (talk) 23:39, 18 December 2024 (UTC)[reply]
Maybe something to consider as we get closer to clearing the backlog (ie under 40k) would be to run a filter sweep like this and see how many articles are referenced solely with UGC. The reward for finishing the work? It’s more work. Kazamzam (talk) 04:54, 19 December 2024 (UTC)[reply]
As discussed above w.r.t. to inline vs. general references, I think this project should maintain a tight focus on eliminating unreferenced articles. An article with a reference to an unreliable source is literally speaking no longer unreferenced; it has moved into the broader category of Wikipedia articles with sourcing issues. I think this project has been successful in large part because its aim is incremental and achievable – we've tried "WikiProject Fix All Sourcing Issues at Once" before, and it doesn't work. One thing at a time. – Joe (talk) 10:13, 4 January 2025 (UTC)[reply]
@Joe Roe - I appreciate that perspective but this is specifically in the context of drives. If someone "references" an article with their own Facebook page, I am not inclined to award that person points for the purpose of the drive, separate from the fact that I am also going to revert it. The issue is not that there are unreliable references but editors adding new unreliable references during a friendly competition of providing, per the guidelines, reliable references. Kazamzam (talk) 12:56, 4 January 2025 (UTC)[reply]
I would not call that a constructive revert; a citation to an unreliable source is a reference and one reference is an improvement over no references, so your revert would be degrading the quality of the article. We have many templates to call attention to problems with sources (for a Facebook post, any of these could fit: {{unreliable source?}}, {{primary source}}, {{self-published source}}, {{better source needed}}). When it comes to points, of course you can give them out as you wish, but the broader question is whether you want to encourage to make any improvement they can, however incremental, or whether you expect them to get everything right at once (i.e. an properly formatted inline citation to a reliable, secondary source). In my experience the former approach has proved far, far more successful in building an encyclopaedia. – Joe (talk) 13:27, 4 January 2025 (UTC)[reply]
@Joe Roe - If a citation to an unreliable source is a reference and is therefore an improvement, we could knock out the entire backlog in a few hours by adding the first result from Google. If you would like to improve the 70,000+ articles in the backlog by adding a reference to any site mentions the article's topic, regardless of quality or verifiability, please be bold and go for it.
Given that we haven't done this over the 18 years that this project has been active, that we continue trying to find reliable sources that meet the guidelines, and that articles are regularly deleted or taken to AfD given the lack of reliable sources, the consensus, both within and without the URA, seems to be that we will follow the guidelines for reliable sourcing, verifiability, and the importance for inline citations. This is the policy that the URA have extended to drives with the goal of improving the encyclopedia. Best, Kazamzam (talk) 15:59, 4 January 2025 (UTC)[reply]
I think it is important to distinguish between the kind of edits we should make from the kind of edits that need to be policed when made by others. Personally, in nearly two decades of editing, I've never added material not supported by an inline citation or knowingly cited an unreliable source, but I can also recognise that there is no policy that says you must not add citations to Facebook and I'm not going to chastise a good faith editor for doing so. What is ideal and what is acceptable are two very different standards and if you are trying to motivate other editors to help with something it is probably more effective to lean towards the latter, is all I'm saying. – Joe (talk) 16:09, 4 January 2025 (UTC)[reply]
@Joe Roe - I appreciate the clarification since that's quite different from what you originally said. The point of the drive is to motivate editors to edit, which I believe we have successfully done given our two recent drives cleared nearly 20,000 articles from the backlog. The point of the guidelines of the drive is to motivate editors to edit well so that the drive is a net-positive for the encyclopedia and does not become a burden to other editors. Happy to discuss further as we get closer to planning the June 2025 drive, but personally I'm not going to advocate for a standard that I myself wouldn't follow. Cheers, Kazamzam (talk) 15:44, 5 January 2025 (UTC)[reply]
@Cielquiparle - I’m showing my age/status as the elephant who never forgets but it’s definitely not the first non-chronological clearance. Back in 2022, I got rid of July 2007 and then went back for the rest of 2007. I believe there’s a mention of it in Archive 4 but I can check the spreadsheet when I’m not on my phone. Either way, very exciting! I wonder if it’s appealing to people to pick a category from the middle of the pack rather than just old - maybe? I know that December 2023 was listed as a clean up project which is why it got so much attention. Now we can just add December 2009… Kazamzam (talk) 10:39, 11 January 2025 (UTC)[reply]
Bonjour! Does or has anyone kept an eye on the Category:All articles lacking in-text citations? Its backlog is bigger and older than ours, and I wonder if it doesn't get much visibility. I wonder if it would be worthwhile to put it on the URA homepage, either in addition to or instead of the Articles needing additional references section. It's getting a bit crowded, and maybe there's a larger discussion to be had about organizing the page so it's easier to read, but I think this is an important category that I've personally neglected and hopefully we can see some of the similar progress we've had in our backlog extend. Cheers, Kazamzam (talk) 15:48, 5 January 2025 (UTC)[reply]
I looked at the category. It is perhaps a pity that the category is populated by two maintenance templates: {{No footnotes}} and {{More footnotes}}. I kind of feel like URA should be more concerned about {{No footnotes}}. The Petscans below select on that, which about halves the number of affected articles. There's a part of me that worries slightly that work on the backlog of completely unreferenced articles would slow down if we added this actively.
Ha further to my last message - because I looked at a few of them - it seems as thought the category gets populated even when one of the templates is added referring to a section, not the whole article. So it scoops up e.g. Book of Enoch and Japanese language even though both articles have tons of inline citations. If we can't weed out the articles where it's only individual sections missing inline citations, then it will be tricky to find the unreferenced needles in the backlog haystack. Cheers, SunloungerFrog (talk) 16:31, 6 January 2025 (UTC)[reply]
@Kazamzam - One of my subpages is for tracking Category progress and includes the "All articles lacking in-text citations". I only check a couple times monthly, usually when adding the "Random" button & "Topic filter". Regards, JoeNMLC (talk) 15:09, 28 January 2025 (UTC)[reply]
Less than 1%
Probably not a greatly interesting milestone, but the number of tagged unreferenced articles has just gone bellow 1% of the total articles (69,445/6,944,585 at the time of writing) for the first time since at least before Apr 2008 (based on ARandomName123's above graph) Cakelot1 ☞️ talk17:18, 25 January 2025 (UTC)[reply]
Hi URA-ites! The bimonthly update will be on 4 February 2025 but I wanted to let anyone who's new or didn't know that the annual update of the historical data (here) will be updated by long time editor Altamel on 11 February. I think these data are really interesting and are great indicators of progress - you can see how many categories were cleared between 2021 and 2025, plus how ENORMOUS some of these categories used to be. I groan when I see a category with 500+ articles but in August 2009, almost every category was in the thousands. Really puts things in perspective and is a fun piece of Wikihistory. Cheers, Kazamzam (talk) 16:19, 29 January 2025 (UTC)[reply]
February 2025 update
First update of the new year!
Headline: We cleared 2,963 articles and are now hovering barely over 69,000 (69,036 to be exact)! For yourself and your fellow editors, please clap.
Minutiae: For anyone interested in a more detailed breakdown of the numbers - average was 35.3 articles; median 12; mode 11. The average category declined by 6.6% and of the 189 categories being tracked, 149 (78.8%) had a decline of at least 10 articles and over 2% of their starting number compared to previous update (I consider this a growth metric that we're being comprehensive in clearing categories across the board). The smallest decrease was September 2019, which only decreased by 3 articles (I'm not mad, I'm disappointed). And the amount of time it takes to update all the categories individually is exactly the length of two (2) Simon & Garfunkel albums.
Highlights: Along with March and April 2009, December 2023 is in dustbin of history! This is the first non-chronological category clearance since 2022 (with the removal of July 2007). We also had a successful DYK submitted by @Cielquiparle for the show The Befrienders which had been unreferenced since 2006! And as editor in arms @Cakelot1 pointed out above, the unreferenced backlog is now under 1% of the total articles for the first time since April 2008. Huzzah.
Low-hanging fruit: We have a number of small fry thanks to the robust efforts of regular editors and drive participants. The infamous September 2019 is a humble 101 articles, dangling precariously like a sinner in the hands of an angry God. Give it a nudge.
High-hanging fruit: Everyone's favourite BFC (Big Friendly Category), December 2009, is a finger-lickin' 9,108 articles as of this writing, after a decrease of 460. The other high-hanging fruit are, still, the Frustrating Five (name open for revision): January 2013 (1,012), April 2019 (842), May 2019 (1,760), June 2019 (3,805), and September 2020 (1,004). This time, September 2020 had the lowest percentage of change between updates (1.57%). Godspeed to anyone working on these.
New challenge: No ties this time, but @ARandomName123 has set up a monthly leaderboard to encourage everyone's inner blood lust. Go forth.
Announcements: Per above discussion, the consensus seems to be that June 2025 will be the best time for our next drive. If anyone is inclined to start a draft, please be bold! Thanks, as always, for the amazing work. We're doing it! I'm calling it now - under 50,000 by the end of the year, if not sooner. All the best, Kazamzam (talk) 16:28, 4 February 2025 (UTC)[reply]
Hooray! Just wanted to quickly acknowledge everyone who contributed in January 2025 – every reliable citation is appreciated – and recognize the Top 10 editors adding references to completely unreferenced articles last month (more or less). Some impressive stats here:
Also - the next update will be in May 2025 so not to interfere with a) my vacation in the Japanese alps and b) the June drive. Hopefully we will knock through the better part of 2009, and then we can start to figure out how to scale Mount December 2009. Kazamzam (talk) 18:46, 4 February 2025 (UTC)[reply]
I'm aiming to kill off December 2024, I've been spending all my time working through there mostly alphabetically. So we'll hopefully have another non-chronological wipeout soon! SilverserenC23:15, 4 February 2025 (UTC)[reply]
