Wikipedia:Wikipedia Signpost/Single/2023-07-17
Big bux hidden beneath wine-dark sea as we wait for the Tides to go out?
Wikimedia Endowment transparency – a year on, nothing seems to have changed
The Wikimedia Foundation has long promised that the Wikimedia Endowment – held by the Tides Foundation and managed as an opaque, organisationally completely separate entity by a board led by Jimmy Wales – would soon be transferred to a financially transparent 501(c)(3) organisation. These promises date back to 2017 (see Signpost coverage this year and last year).
In April 2021, Endowment Director Amy Parker and Director of Development Caitlin Virtue again said on Meta:
A full year has now passed since that 501(c)(3) determination letter (pictured) was received in June 2022. Yet the money has still not been transferred. This means that another year will have passed without public reporting on the Endowment's revenue and expenses, as it is organisationally separate from the Wikimedia Foundation (its revenue and assets are not included in WMF revenue and assets) and the Tides Foundation does not provide any such reporting either.
In response to an inquiry on the Wikimedia mailing list, WMF Chief Financial Officer Jaime Villagomez recently posted the following update on Meta:
In addition, WMF CEO Maryana Iskander and WMF board member Nataliia Tymkiv said on the Wikimedia mailing list:
The Wikimedia Endowment holds a significant proportion of all the funds the public has ever donated to the Wikimedia cause. Yet it does not follow the same standards of transparency that apply to other parts of the movement. For example, it would be unimaginable for any WMF affiliate to ingest over $100 million over the best part of a decade without ever publishing audited accounts detailing revenue and expenses.
Why should the Wikimedia Endowment be different? – AK
Wikimania scholarships
The Wikimania Scholarship outcomes were recently published on the Wikimania site. According to the summary provided there,
This means that less than one in six scholarship applications was approved. To one editor, at least, this seemed a "paltry" amount of funding:
A WMF spokesperson responded by saying:
User:Sdkb seemed unimpressed. – AK
Gitz6666 unglocked
In a rare reversal, User:Gitz6666 had his global lock overturned after lodging an appeal with the stewards. Gitz6666 had been indefinitely blocked on the Italian Wikipedia in May, along with another user, and then had his account globally locked by an Italian steward.
The underlying dispute concerned a sociologist's Italian Wikipedia biography that had attracted press attention for its alleged unfairness (see previous Signpost coverage). – AK
Brief notes
- New administrators: The Signpost welcomes the English Wikipedia's newest administrator, Firefangledfeathers.
- Articles for Improvement: This week's Article for Improvement (beginning 17 July) is Religious philosophy. It will be followed the week after by Act (drama). Please be bold in helping improve these articles!
Tentacles of Emirates plot attempt to ensnare Wikipedia
Agency's leaked emails provide rare glimpse into use of Wikipedia in a "smear campaign" financed by the ruler of the United Arab Emirates
An investigative article in The New Yorker, titled "The Dirty Secrets of a Smear Campaign", describes how "Sheikh Mohammed bin Zayed, the ruler of the United Arab Emirates, paid a Swiss private intelligence firm millions of dollars to taint perceived enemies". Most of the lengthy article (whose audio version runs 1 hour and 13 minutes) isn't about Wikipedia, but there are several paragraphs about how the firm ("Alp Services", founded by an investigator named Mario Brero) used it for their purposes alongside many other interesting tools (such as illegitimately obtaining phone call records or tax records of their targets, and planting stories in various news outlets).
The first part is about an American oil trader named Hazim Nada, founder of a company called Lord Energy:
This outside view from the victim's perspective is later matched to what the reporter learned from leaked/hacked internal emails of "Alp Services":
And:
And regarding another target:
On fr:Kamel Jendoubi, a "Controverses" section was added by a SPA in January 2019, and expanded by another SPA in August 2019. Most of it was deleted in April/May 2021, by an account with only one earlier edit, and then by an experienced editor. Around the same time, the English Wikipedia's article Kamel Jendoubi likewise saw an attempt by an IP editor to remove similar information, which was reverted as "Likely censorship of content"; although a July 2022 edit that provided a more detailed rationale for a more limited removal was successful.
The New Yorker article was published in April. Its findings were put into a much wider context earlier this month when various European news media collaborating in the European Investigative Collaborations (EIC) network reported on the results of an investigation dubbed "Abu Dhabi Secrets", revealing that
This investigation was based on a stash of "78,000 confidential documents obtained by the French online newspaper Mediapart", according to Middle East Eye, which summarized the modus operandi of the campaign as follows:
Many or most of the news reports emanating from the collective EIC investigation don't seem to have focused on the Wikipedia angle. Still, the Spanish publication Infolibre reveals some further details, quoting from messages where Alp's paid Wikipedia editors report about their efforts to their Emirati clients, in particular edits (presumably including these) on the English and Spanish Wikipedia to connect Mohammed Zouaydi (known as the "Al Qaeda's financier") to the Muslim Brotherhood. On the Spanish Wikipedia, they claim to have entered "an intense battle with pro-Muslim Brotherhood elements who wanted to censor information about the Brotherhood and its links to Al Qaeda" (translated back from Spanish).
At the French Wikipedia's "Projet Antipub", editors are currently looking into various other articles and accounts that may be connected to the campaign. – H
Ruwiki
The Telegraph reports (non-paywalled) on the founding of Ruwiki (see previous Signpost coverage):
According to The Telegraph, Ruwiki lacks specific content compared to the Wikipedia version:
In related news, on July 5 between around 2 am and 4 am Moscow time, access to Wikipedia and other "Western internet services" including Google was temporarily disrupted as Russian authorities tested the country's "Sovereign Internet system", as reported in a Twitter thread by Access Now staff member Natalia Krapiva. – AK, H
In brief
- Wiki wars: A BBC radio programme presented by Lara Lewington tries to explain how the Wikipedia sausage is made, touching on hoaxes and Things Gone Wrong like the Croatian and Scots Wikipedias. The programme contains two gross errors – it claims that IP editors lost the ability to edit biographies of living people after the Seigenthaler incident (ahem) and that paid editors are forbidden from editing articles directly (they are only "very strongly discouraged" from doing so).
- TEDx talk: Annie Rauwerda (of Depths of Wikipedia) gave a TEDx talk titled "Why an encyclopedia is my favorite place on the Internet". Watch the recording to find out which encyclopedia it is.
- Jimmy Wales interview: The "co-founder of Wikipedia" was interviewed by Lex Fridman on his popular podcast (video, transcript), for three hours and 15 minutes. One user compiled some excerpts they found especially interesting from a Wikipedian perspective, and journalist Stephen Harrison (known for his Wikipedia column in Slate) summarized several points from the interview in a Twitter thread, e.g. about Wales "push[ing] back on Wikipedia’s alleged left wing bias."
David Thomsen (Dthomsen8) and Ingo Koll (Kipala)
David Thomsen (Dthomsen8)
David died on November 25, 2022, at the age of 83. He was a prolific editor and a self-declared gnome who added to articles on Philadelphia, United States, and created many new articles. David lived in Fairmount, Philadelphia. After earning a degree at Lafayette College and serving for two years with the US Army in West Germany, he worked for 27 years as a computer programmer for Sunoco.
David began his post-retirement Wikipedia career on March 13, 2008; five years later, he had made 100,000 edits, mostly through wikignoming. As of this issue, he is the 52nd-most-active editor in Wikipedia. David was a member of the Guild of Copy Editors, and took part in its monthly drives and blitzes. He was also part of the Article Rescue Squadron, saving uncited-but-promising articles from deletion. David was openly proud of his Wikipedia-editing activities, and he even designed his own Wikipedia-themed caps, which he wore to Philadelphia wiki-meetups and gave away to his fellow Wikipedians. Outside Wikipedia, David was a member of the Historical Society of Pennsylvania.
Dthomsen8 has left an indelible legacy at Wikipedia, and the encyclopedia and editing community are poorer for his loss. Messages of condolence can be left on his talk page.
Ingo Koll (Kipala)
Wikimedia Tanzania and Jenga Wikipedia ya Kiswahili have published a tribute to User:Kipala, a much-loved German Wikimedian who – as a fluent Swahili speaker – was particularly active in the African Wikimedia community (see also Wikimedia-l thread).
– AK
ABC for Fundraising: Advancing Banner Collaboration for fundraising campaigns
- Julia Brungs is the Lead Community Relations Specialist at the Wikimedia Foundation Advancement department.
At Wikimedia, collaboration is a pillar of everything we do. The Foundation is committed to building more of it into our efforts to raise funds to support our mission. The collaboration process for the 2023 English fundraising campaign is kicking off now, right from the start of the fiscal year.
2022 English Campaign
In December 2022, the English Wikipedia community ran a Request for Comment that underscored the importance of the Foundation’s fundraising team working with volunteers on banner messaging. The team kicked off a collaboration process that resulted in the campaign featuring more than 400 banners that came from the co-creation process with volunteers. The revenue performance of the banners declined significantly last year and resulted in a longer campaign with readers seeing more banners than previous years. The fundraising team learned a lot through the collaboration process and is eager this year to build on this work with volunteers to develop content that will successfully invite donors to support our mission. We aim to reach fundraising targets in ways that minimize the number of banners shown, to limit disruption and resonate with readers and volunteers.
Community collaboration in 2023 fundraising
The fundraising team took the community collaboration model created in the 2022 English campaign and expanded it in 2023 to Sweden, Japan, Czechia, Brazil, and Mexico. In addition to on-wiki collaboration spaces, the team created spaces to increase transparency around the fundraising program and worked with volunteers from each community on local language wikis, on virtual calls, and in other forums. Collaboration between communities and the Foundation early on in the fundraising process is critical. Each community gave their time to discuss their own unique context and input ranging from how to provide a strong local payment experience to the best ways to translate Jimmy’s new message from 2022 for a local audience. The focus for fundraising messaging in the 2023 campaigns was primarily centered on the quality of translations and localization. Affiliates often made up a key part of this process–working to provide localization expertise and also bring others to the table. The team is grateful to everyone who worked together with them on these campaigns, and wants to invite you to participate in the 2023 campaign.
What’s next?
Collaboration for the 2023 English banner campaign is starting now! The community collaboration page for the English campaign launched on Thursday (13th of July), before the start of the first pre-tests, to kick off collaboration right from the start of the year. The page provides information to increase understanding of the fundraising program, background on improvements around community collaborations that have been made since the last campaign, new spaces for collaboration, and gives messaging examples to invite volunteers to share ideas for how we can improve the campaign together in 2023.
Building from the campaigns that have run in the past six months, there will also be collaboration calls for community members to bring their ideas. Foundation staff will be at several in person and virtual community events throughout the upcoming months to open up collaboration opportunities where volunteers are gathering and are happy to come to other volunteer venues.
We will continue to iterate and improve this process in fundraising campaigns going forward. Be part of the conversation and help shape the 2023 campaign.
Background on annual planning
This change in approach to how we raise funds was accompanied by a change in how we spend funds in the annual plan for the 23-24 fiscal year. The Foundation made wider changes in its annual plan. Growth slowed last year compared to the prior three years. We also made internal budget cuts involving both non-personnel and personnel expenses to make sure we have a more sustainable trajectory in expenses for the coming few years. This year's Annual Plan provides more granular information on how the Foundation operates, and recenters Product and Technology work with an emphasis on experienced editors, aiming to ensure that they have the right tools for the critical work they do every day to expand and improve quality content, as well as manage community processes.
To learn more about the Foundation’s annual plan, please see a brief Diff post — or the full annual plan.
Are the children of celebrities over-represented in French cinema?
In December 2022, New York magazine did its cover story about the "nepo babies" of Hollywood. The French cinema industry is also known to be biased in favor of children of celebrities. French journalist Maxime Vaudano published a paper in Le Monde on the topic 10 years ago. More recently, Belgian humorist Alex Vizorek has suggested that there is now lots of children of celebrities in sport but "not as much as in French cinema". It's not difficult to find examples. Louis Garrel and Esther Garrel are the children of Philippe Garrel and Brigitte Sy. Chiara Mastroianni is the daughter of Catherine Deneuve and Marcello Mastroianni. Julie Depardieu is the daughter of Gérard Depardieu. Charlotte Gainsbourg is the daughter of Serge Gainsbourg and Jane Birkin. Mathieu Amalric is the son of Jacques Amalric, etc.
Wikidata provides a fantastic database to test the assumption that the number of children of celebrities is higher in the cinema than in other fields. Using it, I found that a) French actors and French filmmakers have a higher probability to have parents with a Wikipedia article than people with another occupation, b) this pattern holds true for males and females but is more important for women than for men, and c) this pattern is specific to French actors and actresses.
Methodology
I first constructed a SPARQL query which takes all people born after 1970 with an article in Wikipedia in French by occupation and citizenship. I then checked if those people have a father with an article in Wikipedia in French or a mother with an article in Wikipedia in French. I haven't controlled for the occupation of the parents, and assumed that having an article in Wikipedia is a sign of celebrity regardless of occupation and country of citizenship.
My first query was the following:
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX schema: <http://schema.org/>
PREFIX wd: <http://www.wikidata.org/entity/>
PREFIX wdt: <http://www.wikidata.org/prop/direct/>
SELECT DISTINCT ?item ?itemLabel ?year ?father ?mother WHERE {
?item wdt:P31 wd:Q5;
wdt:P106/wdt:P279* wd:Q33999;
wdt:P27 wd:Q142;
wdt:P569 ?birthdate;
rdfs:label ?itemLabel FILTER(lang(?itemLabel) = "fr") .
?sitelink schema:about ?item;
schema:isPartOf <https://fr.wikipedia.org/>.
FILTER(YEAR( ?birthdate ) >= 1970 )
BIND(YEAR(?birthdate) AS ?year)
OPTIONAL {
?item wdt:P22 ?father.
?fatherlink schema:about ?father;
schema:isPartOf <https://fr.wikipedia.org/>.
}
OPTIONAL {
?item wdt:P25 ?mother.
?motherlink schema:about ?mother;
schema:isPartOf <https://fr.wikipedia.org/>.
}
}
All other queries are derived from this original query. All computations were done using the Observable platform, which makes it easy to visualize data using JavaScript.
Many more children of celebrities in French cinema than in other occupations
I collected data about people of French citizenship (country of citizenship (P27) is France (Q142)) born after 1970 with an article in the French Wikipedia for several occupations. I assumed that having an article in Wikipedia is a sign of celebrity.
So I chose to look at actor (Q33999), musician (Q639669), writer (Q36180), athlete (Q2066131), politician (Q82955) and film director (Q2526255). For each occupation, I computed the proportion of people whose father or mother, or both, had an article in the French Wikipedia.
I found that 8.0% of French film directors born after 1970 have a father with an article. This probability is slightly higher than for actors and actresses (7.7%) and much higher than in other occupations such as musicians (4.7%), writers (4.1%), politicians (3.8%) and athletes (1.8%).
Looking at the probability to have either a mother or a father with an article, I found similar results. The probability to have a father or a mother with an article is 8.6% for film directors, 8.3% for actors and actresses, 5% for musicians, 4.6% for writers, 4.4% for politicians and 1.9% for athletes.
Those numbers provide some evidence that children of celebrities in France are much more successful in French cinema than in other occupations.
Women in French cinema are more often children of celebrities
_with_a_wikipedia_article_by_gender_and_occupation2.png)
The probability of being the child of a celebrity is much higher for women than for men. 9.8% of actresses with a biography in the French Wikipedia have a parent with a biography. At the same time, 6.8% of actors with a biography have a parent with a biography. I found the same pattern for film directors: 11.2% for women versus 7.3% for men.
A distinctive feature of French cinema
_with_a_wikipedia_article_by_country_of_citizenship.png)
Once I had evidence that there is a huge phenomenon of children of celebrities in French cinema compared to other occupations, I compared the number to other countries. The methodology is slightly different since I look at people with at least one Wikipedia article (whatever the language) and not only an article in the French Wikipedia.
Among French actors and actresses born after 1970 who have a Wikipedia page, 8% have a father or mother with a Wikipedia article. This proportion is much higher than for Italian (2.4%), Spanish (1.8%), Belgian (2.3%), Swiss (1.8%) or Canadian (2.0%) actors and actresses. This confirms that there are disproportionately more children of famous people in French cinema.
Conclusions
This data exploration shows strong evidence supporting the nepo-babies hypothesis in French cinema. It would be useful to go further, and look at the occupation of parents, as well as look at more occupations and more countries.
All feedback, improvements and complements to my analysis are welcome.
References
What automation can do for you (and your WikiProject)
Over the years, people have designed a variety of tools to save you time and headaches. Most deal with centralizing information in some way so you don't have to look for "all the discussion related to topic X" yourself, but can instead make use of centralized lists. Some are my ideas. Others are from, well, other people. Here is a summary of three of the biggest ones out there.
Article Alerts
Ah Article Alerts (or WP:AALERTS)... this is by far the dearest and closest tool/project to my heart. People who already use it can probably fathom why. Prior to 2008 or so (see previous Signpost coverage), if you wanted to know if "Topic X" had proposed deletions, you would have to stroll Category:Proposed deletion, and manually inspect every article out there. Let's say you are interested in dance. For some topic, like the Miani Sahib Graveyard, you can fairly easily tell that it's unlikely to be related to dance. But Gustave Geffroy? Are they a physicist? An athlete? A ballet dancer? A Simpsons character? You have to read the article to know for sure. This takes time. Repeat that for the dozens of articles PRODed... Congratulations, after 20–30 minutes, now you've compiled a dance-related list of PRODed articles. That no one else has access to. That will be outdated tomorrow. For one workflow/discussion venue.
And that's the tedium Article Alerts is designed to tackle. AAlertBot will cross-check all the articles (and other pages like templates) in a WikiProject's scope against all the discussion venues on Wikipedia and create a daily report for the WikiProject. WP:AFC, WP:DYKN, WP:FAC, WP:FAR, WP:GAN, WP:MERGE, WP:PROD, WP:RFC, WP:TFD... it covers them all, though projects have a wide variety of customization options. So if you're curious about dance, head over to WikiProject Dance and look for "Article Alerts", "AALERTS", "News" or similar somewhere on that page.
WikiProject Dance's current Article Alerts listings |
|---|
Did you know
Articles for deletion
Proposed deletions
Categories for discussion
Good article reassessments
Articles to be merged
Articles to be split
Articles for creation
|
The same will apply for any other WikiProject. The full list of Article Alerts subscriptions is available here if you want to browse things directly. If your project isn't subscribed to Article Alerts, it's very easy to do so. Technical help is always available at WT:AALERTS, though most people can probably figure things out themselves.
If your project doesn't advertise its Article Alerts subscriptions on its front page, it's probably a good idea to start a discussion on the talk page to ask what's up with that and if it should be added. And while you can regularly check the mainpage of a WikiProject for the most recent alerts in most cases, putting the Wikipedia:WikiProject .../Article alerts page on your watchlist is what most people should do. For WikiProject Dance, that would be Wikipedia:WikiProject Dance/Article alerts. Lastly, if your project has a standard shortcut, like WP:DANCE, it's a good idea to create a shortcut like WP:DANCE/AALERTS so you can easily point to it during discussions, like a talk page message welcoming a newcomer to the project.
Hats off to Hellknowz for coding that bot.
Recognized Content
Similar to the Article Alerts tool above, which focused on finding active discussions, Recognized Content (or WP:RECOG) is all about finding articles that have achieved some kind of recognition somewhere on Wikipedia. Want to know if your topic has anything listed at WP:FA? WP:FL? WP:GAN? WP:DYK? Well, inspired by the success of Article Alerts, I thought it would be nice to have a bot – in this case JL-Bot – do the hard work of collecting these for you and give you a nicely formatted page with all that information. Using this time WikiProject Bhutan as an example:
You can also have lists of DYK blurbs, this time using WikiProject Berbers as an example:
WikiProject Berbers DYK listing | ||
|---|---|---|
Transcluding 10 of 19 total |
The full list of customization option is available at WP:RECOG. If you're not sure how to set it up, just look at a listing that you like, and you can generally copy-paste what they did, changing WikiProject Foobar to whatever is appropriate. Just as with Article Alerts, most WikiProjects advertise these lists of recognized content somewhere on their front page (search for "Recognized content", "Featured content", "Showcase" or similar). If your project has such lists, but isn't advertising them, I suggest starting a discussion on the WikiProject's talk page on how to best address that issue. You can browse Category:Wikipedia lists of recognized content to find individual listings, which again, you really ought to put on your watchlist.
Lastly, just as with Article Alerts, if your project has a standard shortcut (e.g. WP:BHUTAN or WP:BERBERS), it's a good idea to create shortcuts like WP:BHUTAN/RECOG or WP:BERBERS/DYK so you can easily point to them during discussions.
Hats off to JLaTondre for coding that bot.
Cleanup listings
This tool I had no part in its development or design. However, like the tools above, CleanupWorklistBot is designed to collect all cleanup-related information for articles within a WikiProject's scope. This one is a bit less straightforward to setup, but luckily most WikiProjects already have been integrated. All you have to do is to browse the list of cleanup listings and find something that you care about. Cheese perhaps? Or maybe human rights?
These listings, unlike the two previous tools, cannot be embedded directly on Wikipedia. Instead, most WikiProjects use {{WikiProject cleanup listing}} to advertise their cleanup listings on their front page, though alternatives exist. You can also put those on your own user page if you want.
The listings can be viewed alphabetically, by category, downloaded in a .csv file, and the 'History' link shows a graph of the number of cleanup tags over time for the project. The listings are updated weekly on Tuesday, so if you seriously tackle one cleanup category, or systematically go through a set of related articles, you can actually see the difference you're making from week to week!
If you use the box above, you don't need to create new shortcuts for Cleanup Listings. In the case of Wikipedia:WikiProject Human rights, with the standard shortcut WP:HR, you can just use WP:HR#Cleanup listings and you will be taken to the section where the box is listed.
Hats off to Bamyers99 for coding that bot.
Final thoughts
There are many other tools out there. Some are bot-assisted, like TedderBot's New Page Search, HotArticlesBot's Hot Articles, or JL-Bot's Journal Cited by Wikipedia. Others are user scripts-based like my own Unreliable/Predatory Source Detector, SuperHamster's Cite Unseen, or Trappist the monk's HarvErrors. I plan to cover those in follow up Tips and Tricks columns, but there are other tools I've never used or heard of I'm sure! In the comments, I'd like people to put what tools they use to facilitate WikiProject-wide collaborations or which are otherwise helpful to their editing. Those can be the tools I've already mentioned, so others know they've got widespread endorsement, or tools I've never heard of so people can discover them!
Tips and Tricks is a general editing advice column written by experienced editors. If you have suggestions for a topic, or want to submit your own advice, follow these links and let us know (or comment below)!
Wikipedia-grounded chatbot "outperforms all baselines" on factual accuracy
A monthly overview of recent academic research about Wikipedia and other Wikimedia projects, also published as the Wikimedia Research Newsletter.
Wikipedia and open access
- Reviewed by Nicolas Jullien
From the abstract::
Why does it matter for the Wikipedia community?
This article is a first draft of an analysis of the relationship between the availability of a scientific journal as open access and the fact that it is cited in the English Wikipedia (note: although it speaks of "Wikipedia", the article looks only at the English pages). It is a preprint and has not been peer-reviewed, so its results should be read with caution, especially since I am not sure about the robustness of the model and the results derived from it (see below). It is of course a very important issue, as access to scientific sources is key to the diffusion of scientific knowledge, but also, as the authors mention, because Wikipedia is seen as central to the diffusion of scientific facts (and is sometimes used by scientists to push their ideas).
Review
.png)
The results presented in the article (and its abstract) highlight two important issues for Wikipedia that will likely be addressed in a more complete version of the paper:
- The question of the reliability of the sources used by Wikipedians
- → The regressions seem to indicate that the reputation of the journal is not important to be cited in Wikipedia.
- → Predatory journals are known to be more often open access than classical journals, which means that this result potentially indicates that the phenomenon of open access reduces the seriousness of Wikipedia sources.
The authors say on p. 4 that they provided "each journal with an SJR score, H-index, and other relevant information." Why did they not use this as a control variable? (this echoes a debate on the role of Wikipedia: is it to disseminate verified knowledge, or to serve as a platform for the dissemination of new theories? The authors seem to lean towards the second view: p. 2: "With the rapid development of the Internet, traditional peer review and journal publication can no longer meet the need for the development of new ideas".)
- The solidity of the paper's conclusions
- The authors said: "STEM fields, especially biology and medicine, comprise the most prominent scientific topics in Wikipedia [17]." "General science, technology, and biomedical research have relatively higher OA rates."
- → So, it is obvious that, on average, there are more citations of Open Access articles in Wikipedia (than in the entire available research corpus), and explain that open access articles are cited more.
- → Why not control for academic discipline in the models?
More problematic (and acknowledged by the authors, so probably in the process of being addressed), the authors said, on p.7, that they built their model with the assumption that the age of a research article and the number of citations it has both influence the probability of an article being cited in Wikipedia. Of course, for this causal effect to hold, the age and the number of citations must be taken into account at the moment the article is cited in Wikipedia. For example, if some of the citations are made after the citation in Wikipedia, one could argue that the causal effect could be in the other direction. Also, many articles are open access after an embargo period, and are therefore considered open access in the analysis, whereas they may have been cited in Wikipedia when they were under embargo. The authors did not check for this, as acknowledged in the last sentence of the article. Would their result hold if they do their model taking the first citation in the English Wikipedia, for example, and the age of the article, its open access status, etc. at that moment?
In short
Although this first draft is probably not solid enough to be cited in Wikipedia, it signals important research in progress, and I am sure that the richness of the data and the quality of the team will quickly lead to very interesting insights for the Wikipedia community.
Related earlier coverage
- "Quantifying Engagement with Citations on Wikipedia" (about a 2020 paper that among other results found that "open access sources [...] are particularly popular" with readers)
- "English Wikipedia lacking in open access references" (2022)
"Controversies over Historical Revisionism in Wikipedia"
- Reviewed by Andreas Kolbe
From the abstract:
This brief study, one of the extended abstracts accepted at the Wiki Workshop (10th edition), follows up on reports that some historical pages on the Japanese Wikipedia, particularly those related to World War II and war crimes, have been edited in ways that reflect radical right-wing ideas (see previous Signpost coverage). It sets out to answer three questions:
- What types of historical topics are most susceptible to historical revisionism?
- What are the common factors for the historical topics that are subject to revisionism?
- Are there groups of editors who are seeking to disseminate revisionist narratives?
The study focuses on the level of controversy of historical articles, based on the notion that the introduction of revisionism is likely to lead to edit wars. The authors found that the most controversial historical articles in the Japanese Wikipedia were indeed focused on areas that are of particular interest to revisionists. From the findings:
The paper establishes that articles covering these topic areas in the Japanese Wikipedia are contested and subject to edit wars. However, it does not measure to what extent article content has been compromised. Edit wars could be a sign of mainstream editors pushing back against revisionists, while conversely an absence of edit wars could indicate that a project has been captured (cf. the Croatian Wikipedia). While this little paper is a useful start, further research on the Japanese Wikipedia seems warranted.
See also our earlier coverage of a related paper: "Wikimedia Foundation builds 'Knowledge Integrity Risk Observatory' to enable communities to monitor at-risk Wikipedias"
Wikipedia-based LLM chatbot "outperforms all baselines" regarding factual accuracy
- Reviewed by Tilman Bayer
This preprint (by three graduate students at Stanford University's computer science department and Monica S. Lam as fourth author) discusses the construction of a Wikipedia-based chatbot:
The paper sets out from the observation that
The researchers argue that "most chatbots are evaluated only on static crowdsourced benchmarks like Wizard of Wikipedia (Dinan et al., 2019) and Wizard of Internet (Komeili et al., 2022). Even when human evaluation is used, evaluation is conducted only on familiar discussion topics. This leads to an overestimation of the capabilities of chatbots." They call such topics "head topics" ("Examples include Albert Einstein or FC Barcelona"). In contrast, the lesser known "tail topics [are] likely to be present in the pre-training data of LLMs at low frequency. Examples include Thomas Percy Hilditch or Hell's Kitchen Suomi". As a third category, they consider "recent topics" ("topics that happened in 2023, and therefore are absent from the pre-training corpus of LLMs, even though some background information about them could be present. Examples include Spare (memoir) or 2023 Australian Open"). The latter are obtained from a list of most edited Wikipedia articles in early 2023.
Regarding the "core verification problem [...] whether a claim is backed up by the retrieved paragraphs [the researchers] found that there is a significant gap between LLMs (even GPT-4) and human performance [...]. Therefore, we conduct human evaluation via crowdsourcing, to classify each claim as supported, refuted, or [not having] enough information." (This observation may be of interest regarding efforts to use LLMs as a tools for Wikipedians to check the integrity of citations on Wikipedia. See also the "WiCE" paper below.)
In contrast, the evalution for "conversationality" is conducted "with simulated users using LLMs. LLMs are good at simulating users: they have the general familiarity with world knowledge and know how users behave socially. They are free to occasionally hallucinate, make mistakes, and repeat or even contradict themselves, as human users sometimes do."
In the paper's evaluation, WikiChat impressively outperforms the two comparison baselines in all three topic areas (even the well-known "head" topics). It may be worth noting though that the comparison did not include widely used chatbots such as ChatGPT or Bing AI. Instead, the authors chose to compare their chatbot with Atlas (describing it as based on a retrieval-augmented language model that is "state-of-the-art [...] on the KILT benchmark") and GPT-3.5 (while ChatGPT is or has been based on GPT-3.5 too, it involved extensive additional finetuning by humans).
Briefly
- Compiled by Tilman Bayer
Wikimedia Foundation launches experimental ChatGPT plugin for Wikipedia
As part of an effort "to understand how Wikimedia can become the essential infrastructure of free knowledge in a possible future state where AI transforms knowledge search", on July 13 the Wikimedia Foundation announced a new Wikipedia-based plugin for ChatGPT. (Such third-party plugins are currently available to all subscribers of ChatGPT Plus, OpenAI's paid variant of their chatbot; the Wikipedia plugin's code itself is available as open source.) The Foundation describes it as an experiment designed answer research questions such as "whether users of AI assistants like ChatGPT are interested in getting summaries of verifiable knowledge from Wikipedia".
The plugin works by first performing a Google site search on Wikipedia to find articles matching the user's query, and then passing the first few paragraphs of each article's text to ChatGPT, together with additional (hidden) instruction prompts on how the assistant should use them to generate an answer for the user (e.g. "In ALL responses, Assistant MUST always link to the Wikipedia articles used").
.png)
Wikimedia Foundation Research report
The Wikimedia Foundation's Research department has published its biannual activity report, covering the work of the department's 10 staff members as well as its contractors and formal collaborators during the first half of 2023.
New per-country pageview dataset
The Wikimedia Foundation announced the public release of "almost 8 years of pageview data, partitioned by country, project, and page", sanitized using differential privacy to protect sensitive information. See documentation
Wikimedia Research Showcase
See the page of the monthly Wikimedia Research Showcase for videos and slides of past presentations.
Other recent publications
Other recent publications that could not be covered in time for this issue include the items listed below. Contributions, whether reviewing or summarizing newly published research, are always welcome.
- Compiled by Tilman Bayer
Prompting ChatGPT to answer according to Wikipedia reduces hallucinations
.png)
From the abstract:
The authors tested various variations of such "grounding prompts" (e.g. "As an expert editor for Wikipedia, I am confident in the following answer." or "I found some results for that on Wikipedia. Here’s a direct quote:"). The best performing prompt was "Respond to this question using only information that can be attributed to Wikipedia".
"Citations as Queries: Source Attribution Using Language Models as Rerankers"
From the abstract:
"WiCE: Real-World Entailment for Claims in Wikipedia"
From the abstract:
_01.JPG)
The preprint gives the following examples of such an automatic decomposition performed by GPT-3 (using the prompt "Segment the following sentence into individual facts:" accompanied by several instructional examples):
"SWiPE: A Dataset for Document-Level Simplification of Wikipedia Pages"
From the abstract:
"Descartes: Generating Short Descriptions of Wikipedia Articles"
From the abstract:
"WikiDes: A Wikipedia-based dataset for generating short descriptions from paragraphs"
From the abstract:
From the introduction:
See also the "Descartes" paper (above).
"Can Language Models Identify Wikipedia Articles with Readability and Style Issues?"
From the abstract:
"Wikibio: a Semantic Resource for the Intersectional Analysis of Biographical Events"
From the abstract:
"Detecting Cross-Lingual Information Gaps in Wikipedia"
From the abstract:
From the paper:
"Wikidata: The Making Of"
From the abstract:
"Mining the History Sections of Wikipedia Articles on Science and Technology"
From the abstract:
References
New fringe theories to be introduced
On Wednesday, the English Wikipedia's Arbitration Committee took some sorely needed action on the long-standing subject of fringe theories and WP:PSEUDOSCIENCE, issues where tense disagreements and POV-pushing have been causing trouble for decades.
Drafting arbitrator Hubert Glockenspiel, in an interview with the Signpost, said that the Committee was introducing a set of brand-new fringe opinions, conspiracy theories, and pseudoscientific claims, which will be available for free to anyone interested in arguing on Wikipedia.
A bounty of topics
The new topics span a broad range of subjects, academic disciplines and national concerns. "We tried to get a little bit of everything", said Glockenspiel. "Because, after all, Wikipedia was meant to be the sum of all human arguments about politics. And we're committed to belonging, inclusion, and equity; we need to amplify diverse voices."
A full list (along with suggested arguments for and against each theory) is available at WP:NEWFRINGE, but here is a summary of each one:
- English is actually a dialect of Basque
- While it's true that virtually no grammar, syntax or etymology are shared between the two languages, it is obvious that they share a recent common origin; what else could account for the fact that the two languages use the same words for "Code of Conduct", "cookie", and "Wikipedia"?
- Oroville, California was founded by Alexander the Great in 317 BC
- "Oro" is a Greek prefix. How did a city in California (inhabited by Native Americans, then Spanish-speakers, then English-speakers) get a Greek name? The answer is obvious. The Macedonian Menace and his troops didn't stop at India, like we have been told: instead he and his armies kept going east through China, crossed the Pacific Ocean, and founded a city in the Golden State, getting the drop on other Europeans by several thousand years. Note that San Francisco, Los Angeles, Sacramento, et cetera, were indeed founded by the Spanish or the Americans: Oroville was the only Hellenic city in the state.
- Evidence of this historic achievement has been suppressed by the powers that be, since they are afraid people will realize how based Alexander really was, and return to his projects: uniting Macedon, Greece and Persia (thereby angering all three countries) and worshipping Zeus and Apis as the true progenitors of humanity (thereby angering the Church of the SubGenius).
- Lake Superior goes down to the center of the Earth
- Evidence can be found by trying to swim in it — even on the hottest day in the hottest month of the year, it will always be cold as hell. If you "wait for it to warm up in the afternoon", it will be even colder. How is that possible? It doesn't add up.
- The center of the Earth is cold instead of hot
- Why else would Lake Superior always be cold as hell?
- Hell is cold instead of hot
- Why else would we compare Lake Superior to it?
- Barack Obama never existed
- In reality, the United States simply didn't have a president from 2009 to 2017. The works attributed to "Barack Obama" were written by a variety of authors, orators and politicians; alleged videos of his public "appearances" were simply CGI. This one is fairly easy to figure out: he was allegedly from "Hawaii", an obviously fictional location (the United States somehow contains a tropical island with volcanoes on it?), and started his political career in "Chicago", another prima facie farcical city (a wacky noir setting filled with gangsters and tommy guns?)
- Joseph Stalin was a CIA plant
- Come on. They expect us to believe that a good old boy named Joey Ashville — from the sweet, sweet state of Georgia, no less — just happened to wander into the Russian Revolution halfway across the globe, somehow ended up in charge of the whole thing, and then coincidentally spent his entire career making Communism look terrible?
- Various artifacts on the Moon were planted by beings from outer space
- The native inhabitants of the Moon could never have developed such advanced technology — it had to have been put there by aliens. In fact, careful analysis of the so-called "Moon missions" reveals several entities bearing a distinctive resemblance to the animals of Earth.
- The media and banking system is controlled by the Church of the SubGenius
- This is the only explanation for the sheer scale of the suppression campaign regarding the reality of Hellenic Oroville.
- I don't care what anybody says: it's real to me.
Reactions
While many of the new fringe theories have already been associated with one of the two American political parties, others remain undecided. The Hellenic Oroville theory, in particular, is currently the subject of ardent debate as to what political affiliation its supporters have: some have said that it's an obvious left-wing dogwhistle and critique of American imperialism, whereas some argue that it's an obvious right-wing dogwhistle and fantasy of Macedonian imperialism. There is also a secondary, less-important argument about whether it is correct or not.
One thing's for certain, though: we will have a bunch of AN/I threads about it.
Long-time tendentious editor (and WMF-banned troll) Snowpisser said, through a spokesman sockpuppet, that he welcomed the challenge of the new theories. "I can't wait to start a big clusterfuck over these. Nobody even knows what side they're supposed to be on yet! I will probably be able to catch a few dozen people off guard, and get them to freak out and get themselves banned."
Meanwhile, controversial administrator DarkAngelBlademaster666 said in a talk page comment that she was looking forward to figuring out what the right opinion was to have on them, and then immediately INVOLVED-blocking everyone who she disagreed with. "It's perfect, because none of my existing topic bans apply to this stuff yet. By the time they're expanded to cover these, I will have already gotten to fire off like thirty indefs".
- Note: My friends Hazzard and Skutz came up with two of these (the moon one and the Stalin one, respectively).
If you're reading this, you're probably on a desktop
- Editor's note: I don't know how in tarnation this never ended up getting published. When I rewrote some templates to better categorize Signpost drafts, it revealed that this had been languishing in the doldrums for quite some time. Oopsies woopsies!!! Anyway, it is a fine piece, and here it is. —J
- *Desktop views and mobile views as per toolforge pageviews and massviews analysis
The Signpost in 2020, our sixteenth year of publication, contained twelve issues and 161 articles, compared to the 155 articles of 2019. This article reports data on articles, contributors, pageviews, and comments from 2020 and compares them to data from previous issues.
The 161 articles of 2020, created by 88 Wikipedia users, received a total of 330,911 pageviews. Adding in views from the first page and the single-page edition, the total pageviews reached 354,786. This is a decrease of 
110,574 views from last year. This is also a five year low. The twelve issues have seen comments totaling 122041 words. This is a decrease from last year by 2,735 words. Despite the noticeable fall in views, the amount of discussion has remained relatively stable.
_Article_pageviews_(one_week_vs_all-time).png)
_Article_pageviews_(26_Jan_2020_to_20_Feb_2021)_logarithmic_scale_via_massviews.png)
A sharp fall in views is seen in the view count from the third and fourth issues. This general fall remained visible until the end of the year. Further 2021 signaled a five year low in total view count.
Orange highlight represents pageviews of the top two articles while blue represents pageviews for the remaining.
Contributors and comments
- How many users have contributed to The Signpost in 2020?
- 10 Signpost contributors in 2020 account for around 52% of the total byline mentions, with the remaining 77+ users accounting for the remaining byline mentions. However this does not include the additional 831 editors if you count articles such as the essay Wikipedia:An article about yourself isn't necessarily a good thing or the humour collaboration Cherchez une femme from the French Wikipedia.
- The contributions of the top 10 are similar to last year.
- What about the state of comments and discussion in The Signpost in 2020?
Articles with the most discussion:
- Most words:
- Op-Ed: Re-righting Wikipedia: 7509 words
- Op-Ed: Anti-vandalism with masked IPs: the steps forward: 7235 words
- Most participants:
- From the editor: Reaching six million articles is great, but we need a moratorium: 46 unique editors
- News and notes: Jimmy Wales "shouldn't be kicked out before he's ready": 31 unique editors
Signpost desktop versus mobile views
85% of Signpost views are from desktop devices: For the period 27 January 2020 to 18 February 2021 and considering all issues of volume 16, 85% of Signpost views came from desktop. The remaining 15% views came from mobile views (mainly mobile web views and relatively negligible mobile app views). For comparison Wikipedia articles get around 35% to 60% of views from desktop as per the sample I have taken. It is also interesting to note that out of a sample of ~62M edits on the English Wikipedia taken in a study by Ainali, 10% of the edits were from mobile devices. Do Signpost creators use mobile or desktop to create their articles?
Outliers
Q. What do Signpost outliers have in common?
Ans. Here are a few observations related to outliers:
- Outliers reach an internal audience wider than the usual Signpost readers through discussion on multiple fora such as RfC or MfD.
- Outliers reach an external audience which in turn can drive external conversation.
- Outliers can be highly controversial resulting in point one or two.
- Examples: Humour: Pesky Pronouns
The following input from Smallbones helped explain some of the statistical attributes of the Signpost data. The distribution of any data related to Wikipedia pageviews is "normally highly non-normal", i.e. it's usually highly skewed to the right and non-Gaussian. The skew to the right - some unusually big numbers or "outliers" - makes some usual statistical tools - like average (mean) - difficult to interpret.
The cause of the skew might be: some rather very good or very bad journalism! Internal controversy - which was really good (or bad) for us to have brought up. We touched a nerve. Posting links on other platforms, which can be good - an indication of quality journalism and interest by the outside world, or bad - somebody is trying to manipulate our pageviews or trying to make some type of point. The concept of outliers was expanded to this report after it was suggested by Smallbones in 2019.
Bingeing on Signpost
When conjuring up this article, I noticed that it inadvertently involved bingeing on the Signpost. I wonder if anyone else has done so. I have my reason, what was yours?
- Part data: Signpost Statistics 2020; via toolforge pageviews massviews
- Notes
Scrollin', scrollin', scrollin', keep those readers scrollin', got to keep on scrollin', Rawhide!
_2.jpg)
Well, here we are! Yet another featured content. This one's meant to come out - and of late that's actually meant something for once – on the 17th, which is World Day for International Justice. We'll find out in ... two issues, I believe (we work about a fortnight behind so we actually have time before publication to write things, but that's just past the cutoff for next issue), whether or not that resulted in any featured content of note.
It's also World Emoji Day, but 🖕 to that. Anyway! I probably make these rather chatty, but, I don't know. I have about 600 readers, according to the stats, maybe more given that doesn't include those of you that read the single-page version of the Signpost (and if you do read the single-page version, and something I did didn't get changed in editing, I am so sorry). It feels like we used to get a lot more readers, but, well, it also feels like Wikipedia as a whole has gotten a lot less chatty than it used to be. I'd like to go back to writing the multiple articles I was doing on here for a while, but, well, one does what one has the mental capacity for, and I'm just dealing with too much other stuff to be able to do much more. As it is, probably going to have to finish the April Fools' retrospective next year, which is kind of fine.
Getting back on topic, this was a fantastic fortnight (well, fiftnight, I guess?) for featured pictures, and okay-ish for lists and articles. It marks the return of featured topics for the first time since "Animals in lagomorpha" in October last year. Or ... probably does? The trouble with featured topics is that good topics automatically become featured topics when over half their articles become featured, and sometimes that doesn't get announced properly. Template:Announcements/New featured content says that it's the first since "Animals in lagomorpha", though, and I don't have evidence to say otherwise.
Hope everyone's doing well!
Ike & Tina Turner by Rob Mieremet for Anefo, restored by Adam Cuerden, yet another featured picture. Spoiler: We have a lot this month.
Featured articles
Ten featured articles were promoted this period.
- Merenre Nemtyemsaf I, nominated by Iry-Hor
- In a black basalt sacophagus, for centuries he lies,
- Until removed in the 19th century by a pair of German guys.
- (Or is it really him? Could it be an imposter?
- Did some other person sneak their corpse into the roster?)
- Anyway, they took him to show a friend, and, in a massive gaffe,
- He was heavy so they went and broke the mummy right in half.
- They didn't even take both parts: One's been lost since World War Two!
- What an amazingly daft and stupid thing for archaeologists to do!
- Henry II of England, nominated by Unlimitedlead
- He reconquered Wales, lands in France he increased,
- But will no-one rid him of that turbulent priest?
- Battle of New Carthage, nominated by Gog the Mild
- For what else is our beloved featured article status for
- But to promote each and ev'ry battle of each and ev'ry Punic War?
- Seriously, though, great work, Gog! You're doing WP:MILHIST proud!
- Frances Cleveland, nominated by Thebiguglyalien
- She was First Lady of the United States, but after Grover's death she did a ton,
- Like promoting education, women's suffrage and being prepared for World War One.
- Tunic (video game), nominated by PresN
- When I was a child, when I got a game in a store,
- I read the manual on the ride home to get myself ready for more.
- This game's manual, like back then, has quite a lot to say,
- But you have to unlock it to read it, a page at a time as you play.
- Al-Adid, nominated by Cplakidas (a.k.a. Constantine)
- Who kept Egypt Shia despite Saladin ?
- Al-Adidn't. His quest he failéd in.
- "My Man" (Tamar Braxton song), nominated by Aoba47
- A blues song that covers, as it goes on its course,
- The trauma arising from Tamar's parents' divorce.
- Neptune (Alexander McQueen collection), nominated by Premeditated Chaos (a.k.a. PMC)
- Ancient Greece meets the eighties, and then they combine
- In a dazzling display: McQueen's fashion design.
- Edward Dando, nominated by SchroCat
- In the early eighteen hundreds, he ate food, though without cash
- By inventing something along the lines of the modern "dine-and-dash".
- 1919–20 Gillingham F.C. season, nominated by ChrisTheDude
- Throughout the First World War, the team couldn't exist
- So they had to form a new one, and kept swapping who's on the list.
Featured pictures
Twenty (!!!) featured pictures were promoted this period, including the ones at the top and the bottom of this article, and shoved in between other sections. Because there's twenty of them.
Kowloon Walled City by Ian Lambot
Thomas Mundy Peterson by William R. Tobias, restored by Adam Cuerden
Theodore von Kármán inspecting two of the models used in wind tunnels. U.S. Air Force photographer, restored by Chris Woodrich
_male_Bia%C5%82owieza.jpg)
_Finca_El_Pilar.jpg)
.jpg)
_female_underside_Dobruja.jpg)
_male_Burren.jpg)
.jpg)
_underside_Phuket.jpg)
-20230407-RM-124525.jpg)
.jpg)
Sol de Mañana by Kallerna. Another featured picture.
Featured lists
Ten featured lists were promoted this period.
- List of diprotodonts, nominated by PresN
- An order of Australian marsupials for you to view,
- Like a wallaby, a bettong, or a rat-kangaroo.
- List of awards and nominations received by Modern Family, nominated by RunningTiger123
- A mockumentary sitcom, which all the critics say
- Is great at portraying relationships, one of which is gay.
- List of Best Selling Rhythm & Blues Singles number ones of 1968, nominated by ChrisTheDude; List of Music Bank Chart winners (2018), nominated by EN-Jungwon and Jal11497; List of Best Selling Soul Singles number ones of 1969, nominated by ChrisTheDude; and List of Billboard Latin Pop Airplay number ones of 2000, nominated by Magiciandude (a.k.a. Erick)
- Sometimes lists might start to cluster
- Which any poet's mind will fluster.
- It's always hard to know what to say
- About lists of what was hot on each specific day.
- List of international goals scored by Kelly Smith, nominated by Idiosincrático
- At football, she proved very clever:
- Steph Houghton declared her "England's best-ever"
- And Houghton probably was quite right,
- Unless you prefer Ellen White.
- List of malvid families, nominated by Dank
- There's cabbage and broccoli, but I'd ask you don't knock all yet:
- For lo! There's mangos and maples and cashews and chocolate!
- List of World Heritage Sites in the United States and List of World Heritage Sites in Pakistan, nominated by Tone
- Over the years, our good Tone has unfurled
- Lists on World Heritage from throughout the world.
19th-century Chinese scroll depicting Al-Hajj (part of the Qu'ran), unknown creator. This is one of our newest and tallest featured pictures. Templates like {{tall image}} exist for a reason, but I'm ignoring those reasons because I find terrible layout choices funny. And, yes, it's a bit distorted. We're pretty sure that's because the paper distorted as it aged.The Idol becomes the Master
- This traffic report is adapted from the Top 25 Report, prepared with commentary by Igordebraga, Milowent, Ollieisanerd and Capsulecap.
I'm tryna put you in the worst mood, ah (June 25 to July 1)
Rebel rebel, party party, sex sex sex, and don't forget the violence (July 2 to 8)
Exclusions
- These lists exclude the Wikipedia main page, non-article pages (such as redlinks), and anomalous entries (such as DDoS attacks or likely automated views). Since mobile view data became available to the Report in October 2014, we exclude articles that have almost no mobile views (5–6% or less) or almost all mobile views (94–95% or more) because they are very likely to be automated views based on our experience and research of the issue. Please feel free to discuss any removal on the Top 25 Report talk page if you wish.