This is an archive of past discussions about Wikipedia:Reliable sources. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
The answer is: It depends on which channel. Obviously, a random person uploading a video of their kids playing is not a reliable source. At the other end of the spectrum, a lot of television news shows put copies of their news on YouTube, and it would be silly to say that the news show is reliable if you watch it on TV but not reliable if you watch the same thing from the same news channel on YouTube. In between those two things, you have to use your best judgment. For example, if a musician makes a video saying why they wrote a particular song, or that they're 25 years old, then that's reliable as an WP:ABOUTSELF statement. But you wouldn't want to use a musician's video saying things about a political candidate or the price of eggs or something like that. WhatamIdoing (talk) 17:39, 8 May 2025 (UTC)
YouTube isn’t a source itself but a platform the actual source would be the uploader of any of the videos. Also one other thing to be careful of is the possibility of copyright violations since some people do upload content they don’t own the rights to.--65.92.245.71 (talk) 03:40, 9 May 2025 (UTC)
A source like this is definitely an RS because: (1) it is produced by the museum housing the ship in question (2) it is presented by the Director of Research of that museum (3) the presenter has edited, contributed to and written three books which are an RS for the relevant article (4) the presenter is a noted expert in their field, with numerous research papers which are cited by others. I don't think you need me to give examples at the other end of the spectrum. There might be some difficulty in assessing the value of videos in the "shades of grey" area in-between these extremes. ThoughtIdRetiredTIR12:28, 9 May 2025 (UTC)
The comments given at WP:RSPYT discourage use of YT for any use, partly because there should be an original text source that is easier/quicker to obtain encyclopedic background. YT sources may be WP:COPYVIO, and are mostly accompanied by margin ads, which are contrary to Wikipedia's no-ad policy. YT also requires higher bandwidth, creating access challenges for users with low-quality Internet service. Best to revert any YT source per RSPYT and find a WP:RS. Zefr (talk) 19:53, 4 July 2025 (UTC)
I suppose "policy" implies a technical definition not specifically included among those at WP:PG. From the Wikimedia FAQ:"Wikipedia is not funded through advertising", which I know you knew that I knew. Zefr (talk) 20:37, 4 July 2025 (UTC)
"Wikipedia is not funded through advertising" sounds nothing at all like "Reliable sources are not allowed to have margin ads", as you know very well. WhatamIdoing (talk) 20:41, 4 July 2025 (UTC)
If the margin of a YT source is populated with ads, or the video is interrupted by ads, this would signal to an editor that a better source should be found. Acceptance of a source with ads is acceptance of its advertising, then passed to the next Wikipedia user to wonder about source quality - an avoidable practice.
Is there a discussion or policy revision proposal specifically accepting sources with moderate-heavy margin or in-video advertising? Zefr (talk) 21:26, 4 July 2025 (UTC)
I think advertising is irrelevant to reliability. Remember that newspapers, which we cite extensively, typically have many advertisements. Advertisements are an aspect of the source's financial model, not an aspect of their reliability. Zerotalk00:48, 5 July 2025 (UTC)
As others have said… YouTube is a hosting service, not a source itself. Most of what is posted to YouTube is unreliable, but some of it is quite reliable. It depends on who posted it. Blueboar (talk) 01:05, 5 July 2025 (UTC)
If the margin of an ordinary daily newspaper is populated with ads – and pretty much all of them are – do you think we should stop using newspapers, because that's a signal to an editor that a better source should be found?
Advertising is not a signal of reliability. Some ad-heavy sites are generally reliable sources, and some ad-free sites are not. The New York Times is full of advertisements; the Scientology website has none. But it's the NYT that's reliable, not the scientologists' website.
AFAIK the only serious discussion about rejecting advertisements is about Wikipedia:External links, and the usual challenge there is to convince zealous editors that when WP:ELNO#EL5 rejects "objectionable amounts of advertising", it does not mean "any advertising at all is objectionable". Even that doesn't come up very often, though. WhatamIdoing (talk) 16:24, 5 July 2025 (UTC)
So I had a look at RSP on YouTube. It was marked as a "questionable source" in the early days. Template:Questionable sourcedisplays as "Generally unreliable". The initial version linked five previous discussions:
a video of a politician that was wanted for a direct quotation, because "MSM [mainstream media] censors this"
a video of guest lecture by an author, talking about the themes in his own work, at a major research university
a general question of whether clips of ordinary television news reports (CNN and Reuters were given as examples) are acceptable
a general question, to which editors say things like "very simple" because we accept "the channel of an official news organizations", "perfect examples of when YouTube can be used", etc.
a video of a speech at a convention, to which editors say things like ""YouTube videos" are not inherently reliable or unreliable as sources, any more than "books" or "TV programmes"", and it turns out that the OP wants YouTube as a whole to be declared unreliable because this one speech "presents commercial activities" (classic case of moving goalposts).
In other words, nothing that could support the designation as WP:GUNREL.
At the time, RSP was using a three-color model: "Good", "No consensus", and "Generally unreliable". I think that in the current model, a more accurate description would be "Additional considerations apply". Specifically, it doesn't matter if it's "YouTube"; it matters whether it's an official channel for an ordinary reliable source vs a self-published video.
and probably other pages, too. When a discussion could affect multiple pages, there's no such thing as "the" talk page. There are instead multiple talk pages for multiple pages. (Personally, if this proceeds to an RFC or other decision-making discussion, I think it should be on its own page, like Wikipedia:Verifiability/2012 RfC was.) WhatamIdoing (talk) 17:17, 15 September 2025 (UTC)
The top of this page says "This is the talk page for discussing improvements to the Reliable sources page." If there are three pages to change, then three RfCs are okay. That 2012 RfC seems to have been about changes to one page, WP:V, so I fail to see why it could be considered a precedent. Peter Gulutzan (talk) 17:44, 15 September 2025 (UTC)
Having discussions about the same thing on multiple pages is a violation of WP:MULTI and WP:TALKFORK. It's also not a logical thing to do according to common sense. What if the editors at one page decide "Yes" and the editors at the other page say "No"? Then we end up with WP:PGCONFLICT problems. It's much better to have a single centralized discussion. WhatamIdoing (talk) 20:48, 15 September 2025 (UTC)
Agree… one discussion is better than multiple discussions. It doesn’t much matter where the discussion is held - we can (and should) leave notifications on other related pages pointing to the discussion. Blueboar (talk) 21:40, 15 September 2025 (UTC)
Like with Youtube, if we know the image uploader is an official account of the entity in question, then that generally can be used as a primary source. If that can't be verified, then no, its not reliable. Masem (t) 15:46, 21 September 2025 (UTC)
Has anyone indicated that any given bachelor's thesis has "had significant scholarly influence"? (WP:THESIS says Masters dissertations and theses are considered reliable only if they can be shown to have had significant scholarly influence. No scholarly influence, no reliability.) WhatamIdoing (talk) 02:10, 25 September 2025 (UTC)
I can definitely say that MY bachelors thesis (on the colonial history of the town where I grew up) is unreliable and had no scholarly impact beyond the local high school history teacher showing it to his students as an example of how to create a bibliography. Blueboar (talk) 11:35, 25 September 2025 (UTC)
My bachelor's thesis basically positioned me like Cassandra with regard to the impact of war mobilization in the United States on the antiglobalization movement but, despite me being entirely correct, it had absolutely no scholarly impact in any regard and, frankly, I think I mostly got really lucky regarding the predictions I made. It would also not constitute a reliable source. I think, in general, we should not be using bachelors' theses. Simonm223 (talk) 11:39, 25 September 2025 (UTC)
Latest comment: 8 months ago3 comments2 people in discussion
The journal has been brought up in the talk page of the Dead internet theory, specifically the article The Dead Internet Theory: Investigating the Rise of AI-Generated Content and Bot Dominance in Cyberspace. It looks predatory, but I'm not 100%. Could someone please let me know if it passes sniff test. I would like a 3rd party opinion I can point to, as this topic gets a bit of attention and I don't want to be the only "bad guy" gate keeping the page. Thanks! GeogSage (⚔Chat?⚔) 20:53, 8 October 2025 (UTC)
“ a contemporary secondary news source can quickly become a historical primary source. Articles of recent current events must be periodically updated with new secondary sources.”
Latest comment: 7 months ago13 comments7 people in discussion
What, precisely, is meant by this? Is it commanding us to purge articles on events that attracted lots of media interest at the time but failed to attract retrospective interest? Eldomtom2 (talk) 01:52, 10 November 2025 (UTC)
It’s not perfectly put.
A contemporary secondary source will become a historical primary source.
Articles of events must be updated with later secondary sources.
What time scale? I reckon at one year later, less than half the sources should date from the month of the event.
News sources for the most part are already primary sources at the time they are published. If they are simply recapping what's happend without any additional commentary, that's a primary source. Masem (t) 02:28, 10 November 2025 (UTC)
To put it more concretely:
When you are editing articles like Grooming gangs scandal, it would be ideal if an editor or two found a reputable book on the subject or a long analytical piece (scholarly or maybe a "serious" magazine) and replaced as many citations to newspaper sources with citations to a page in the book as was feasible.
Agreed, but this is really about reliability and not really about primary vs secondary. A later source which has the benefit of hindsight and full knowledge of the accumulated evidence will be more reliable than one written in the heat of the moment. As is often the case, the use of the primary/secondary distinction in the present wording serves mostly to obscure the meaning. Zerotalk12:13, 10 November 2025 (UTC)
I agree. The point about time-based primary to secondary transitions is off point. The paragraph conveys new, beyond the earlier “older sources may be inaccurate”. The third paragraph of WP:AGE MATTERS should be cut. SmokeyJoe (talk) 00:44, 11 November 2025 (UTC)
"Must" implies consequences if it isn't done. Are there? Is "too many secondary sources" an argument to use at a deletion discussion? If not, "must" should be changed to "should." Zaathras (talk) 14:37, 10 November 2025 (UTC)
Yes, this is what I'm getting at. If it was "Articles of recent current events should be periodically updated with new secondary sources where available", it would make perfect sense to me, but at the moment it seems to be "old primary sources = bad", which doesn't seem to be really something there's consensus on (see, for instance, WP:PRIMARYNOTBAD). —Preceding unsigned comment added by Eldomtom2 (talk • contribs) 16:20, 10 November 2025 (UTC)
I think this is "must" in the sense of "or else the citations will be lower quality than we'd really prefer". "Must" doesn't mean an editor will be punished in some way.
Old primary sources are usually bad, but it depends on exactly what we're using them for. An old primary source cited for its own contents (e.g., if quoting Shakespeare) is not bad. An old primary source cited for facts (e.g., The Daily Whatever wrote in 1860 that Abraham Lincoln had been elected as the 16th president of the United States) is kind of bad. WhatamIdoing (talk) 04:53, 13 November 2025 (UTC)
"Articles of recent current events must be periodically updated with new secondary sources" doesn't actually say anything at all about purging etc... Updating doesn't necessarily require replacing, it doesn't say that the sources must be updated it says that the article must be updated which it can be without removing anything. Horse Eye's Back (talk) 19:42, 10 November 2025 (UTC)
This language is new, it was only added in September of last year. There doesn't appear to be consensus for it anywhere on the talk page (Wikipedia talk:Reliable sources/Archive 72 covers the time period) for the addition. Therefore I have reverted the edit, anyone who wants to add this text back in will need to obtain an explicit consensus to do so on this talk page. Horse Eye's Back (talk) 16:53, 11 November 2025 (UTC)
Latest comment: 6 months ago10 comments4 people in discussion
We say: "Questionable sources are those...that are promotional in nature". However, reliable sources are not required to be neutral, unbiased, or objective.
I wonder if we have a shared understanding of what it means for a source to be "promotional in nature". For example, are the red carpet interviews before the Oscars "promotional in nature"? Is a positive book review "promotional in nature"? WhatamIdoing (talk) 23:34, 20 September 2025 (UTC)
There is no right answer to this… a lot depends on the “who, what and where” of the source. The key is to ask: is the interview/review independent of the thing being commented on. An interview with someone involved in a film is probably not independent… a gushing book review might be. Blueboar (talk) 15:36, 21 September 2025 (UTC)
So you see it as a question of WP:INDY? Why don't we just say "isn't independent of the subject matter", then? Editors have a shared understanding of what independent means. WhatamIdoing (talk) 15:53, 21 September 2025 (UTC)
I don't know that I've ever thought about that phrase, nor do I recall anyone bringing it up in any exchange I've had. (For that matter, I don't recall anyone ever referring to WP:QUESTIONABLE in a discussion.) It looks like "promotional in nature" was introduced in 2009. I'm inclined to think of it as sponsored content (this section wasn't introduced until 2018) + the content that gets excluded as "unduly self-serving" in BLPSELFPUB/ABOUTSELF/SELFSOURCE. So I don't see a strong reason to keep that phrase in light of those sections. The content of Oscar red carpet interviews largely falls under one or the other of those; at least, I think of someone promoting a film they're in as a form of sponsored content. A positive book review would be OK to use, as it doesn't fall in either of those; certainly practice is to pull from both positive and negative book reviews. As an aside, I find it odd that there's a section on questionable sources but not on generally unreliable sources. FactOrOpinion (talk) 19:31, 21 September 2025 (UTC)
I suspect that the absence of a "gunrel" section is because that concept was made up by WP:RSP and technically has no basis in policy. The community used to believe what the FAQ says:
Are there sources that are "always reliable" or sources that are "always unreliable"?
No. The reliability of a source is entirely dependent on the context of the situation, and the statement it is being used to support. Some sources are generally better than others, but reliability is always contextual.
– though the difference is largely in principle. In practice, there were always sources that were basically unusable (given that, e.g., so few articles actually need a statement like "A throw-away account on Reddit once said ____"). WhatamIdoing (talk) 19:59, 21 September 2025 (UTC)
Here's a new way of looking at this: Is a straight-up blatant advertisement a WP:QUESTIONABLE source? For example, if you open a dead-tree magazine, and there's an advertisement there saying that The Pure Soap Company makes Pure brand soap, which has assorted qualities that one expects soap to have, then is that WP:QUESTIONABLE on par with a content farm, extremist website, or a notorious supermarket tabloid? WhatamIdoing (talk) 05:03, 13 November 2025 (UTC)
It is for some things, it is not for other things. It is not on par with a content farm, extremist website, or a notorious supermarket tabloid. Last month I used this advertisement in ketchup chips to verify the claim "An earlier description of Smith's manufacturing a "tomato sauce flavour" crisps is given in an October 1968 advertisement in the Scottish newspaper the Montrose Review.". I did not feel it was questionable for such a claim, even though it contradicted what secondary GREL sources were saying. Rollinginhisgrave (talk|edits) 11:49, 15 November 2025 (UTC)
I agree with you, but I can't think of an example of anything that is more obviously "promotional in nature" than a blatant advertisement. I get an advertisement from the grocery store every week. There cannot be any purpose to that other than "promotion"; in fact, they call the various items listed in it "promotions".
For this guideline, you can see the historical development in these links: 2008 version, 2009a, 2009b.
Wikipedia:Verifiability#Questionable sources includes the word promotional ("websites and publications expressing views widely considered by other sources to be promotional, extremist, or that rely heavily on unsubstantiated gossip, rumor, or personal opinion"), but I think the key difference is in the prior sentence:
WP:V: "Questionable sources are those that have a poor reputation for checking the facts, lack meaningful editorial oversight, or have an apparent conflict of interest."
WP:RS: "Questionable sources are those with a poor reputation for checking the facts or with no editorial oversight."
Imo this is a symptom of discussions on whether we can trust something being oriented around sources rather than claims. And on reflection, an extremist website or supermarket tabloid saying "tomato ketchup flavour chips are being sold in this region" would probably usually be fine for verifiability if the authors are in the region. The source then is better understood as "a Glaswegian" than "killallminorities.co.uk".
I think losing promotional, gaining conflict of interest, and reframing around "a source may be questionable for a given statement if it has a poor reputation for checking the facts, lacks meaningful editorial oversight, or has an apparent conflict of interest." would better communicate what we are trying to say. Promotional in nature is too broad, and incidentally captures "promotional in effect" which includes positive book reviews. Rollinginhisgrave (talk|edits) 23:50, 15 November 2025 (UTC)
I've replaced the WP:RS sentence with the one from WP:V. That doesn't address the "promotional in nature" problem, but perhaps it's a step in the right direction. WhatamIdoing (talk) 03:19, 19 November 2025 (UTC)
RFC on restructuring RSP
Latest comment: 6 months ago2 comments2 people in discussion
The sale was mentioned yesterday on RSN, and it's come up in other discussion about the Telegraph. As with any change of ownership, it's definitely something to keep an eye on. -- LCU ActivelyDisinterested«@» °∆t°13:44, 22 November 2025 (UTC)
Keep an eye on it? sure… but right now we don’t have a crystal ball. A change in ownership might or might not indicate a shift in reliability. Time will tell. Keep in mind that it took YEARS of compiling examples before we finally (reluctantly) deprecated the Daily Mail. Even then, we carved out exceptions to that deprecation (sports reporting, for example). Blueboar (talk) 16:16, 22 November 2025 (UTC)
Yep, it comes up on RSN regularly - it's a watch and wait situation. Sometimes change in ownership causes a change in a publication, othertimes it changes nothing. -- LCU ActivelyDisinterested«@» °∆t°16:21, 22 November 2025 (UTC)
Ancient histories
Latest comment: 6 months ago5 comments4 people in discussion
Is there somewhere in policy where we address the use of ancient histories? In some articles (especially Chinese monarchs/nobility from the ancient world), I see zero contemporary coverage and only material from the 5th to the 11 century being cited. To my mind, this should not be allowed. Documents that old, even if secondary, require historical interpretation. One needs training as a historian to use that type of source. This should probably be addressed somewhere if it isn't already.4meter4 (talk) 22:04, 28 November 2025 (UTC)
(I come at this editing mostly ancient Greece and Rome articles.) I would agree that editors should not be using ancient histories for large portions of their narratives. This is in part what WP:PRIMARY says. But see also WP:GRLIT and WP:CLPRIM.
However, a lot of the content at least in classical studies is rather old and predates a lot of the library access which we have via WP:LIBRARY today. It wasn't as if the Cambridge Ancient History (or more analogously the Cambridge History of China didn't exist); it was that Wikipedia editors circa 2008 couldn't easily get their hands on it. Thus, I think a lot of these problems are largely holdovers – both in terms of content and, since people are still sometimes doing original research with primary source citations, in terms of practice – from a then-absence of readily available scholarly sources to consult. Ifly6 (talk) 23:22, 28 November 2025 (UTC)
Old sources are prone to being labelled as “primary sources”, but source typing always depends on what information is being used, and how.
Ancient secondary sources are unusual. Do you have an example. Common challenges are the accuracy of the translation, and the reliability of assumed information, such as context, and what the writer believed the readers already knew. Often, something is written in stone because it is not true and someone powerful wants to make it true by assertion. SmokeyJoe (talk) 07:13, 29 November 2025 (UTC)
Are we talking about reliability, or notability? Notability is rarely decided on questions of reliability, of the source information, but more depends on the reputation of the publisher. SmokeyJoe (talk) 07:23, 29 November 2025 (UTC)
Question about reliable sources?
Latest comment: 6 months ago4 comments4 people in discussion
Is the emails from "Contact Us" on many websites a reliable source?
Is personal communication from an expert a reliable source?
No. It is not good enough for you to talk to an expert in person or by telephone, or to have a written letter, e-mail message, or text message from a source. Reliable sources must be published.
The answer is no, an e-mail message sent to you is never a reliable source.
Can opinion polls and surveys even be considered Reliable sources anymore?
Latest comment: 5 months ago5 comments4 people in discussion
I’ve been wondering whether opinion polls and surveys should still be considered reliable sources, given the recent trends in survey research methodology. In North America, for example, response rates for random-sample surveys have fallen dramatically — often hovering around 5%, which is a historic low. This raises questions about how representative these samples really are.
Organizations like Gallup and the Pew Research Center earned much of their credibility from accurately predicting election outcomes in the past, but since the mid-2010s, both have largely moved away from “horse-race” election polling. Gallup stopped after disappointing accuracy in the 2010 and 2012 cycles, and Pew followed by 2016, citing a desire to focus on methodological improvement. Ironically, issue polling (which they now emphasize) doesn’t have the same clear benchmarks for accuracy that elections do, there’s no definitive “result” to compare with.
I know that groups like Pew compare their surveys to high-quality government surveys for benchmarking, but those benchmarks don’t exactly cover every topic that private or nonprofit pollsters study, and they focus mainly on demographics rather than attitudes, beliefs or opinions. Moreover, even major government surveys have faced declining response rates of their own.
Another related concern is the rise of online opt-in panels (notably YouGov), which rely on volunteer participants and have been criticized (often by pollsters that still advocate randomized sampling) for accuracy and representativeness.
Given all these developments, should the guidelines on Reliable sources consider qualifying or clarifying how (and when) public opinion polls and surveys can still be treated as reliable sources? Belson 303 (talk) 14:58, 11 November 2025 (UTC)
I think the most important principle is that every source is reliable for something – but that often, it's is reliable for nothing more than "This document contains the following words:" (or, in the case of a survey, "One survey claimed...").
Seconding WhatamIdoing on this one, definitely more of a due weight question. I would also note that Gallup and Pew earned their credibility for their non-political polling... Theres never actually been a period when you could describe political polling as "accurate" in the RS sense of a meaning beyond the poll or survey itself unless there is actual analysis at that level. Horse Eye's Back (talk) 23:30, 12 November 2025 (UTC)
I disagree, Gallup and Pew earned their credibility because their election polls largely matched the election results within the margin of error and their credibility on issue polling was largely derived from their success in election prediction. I suggest you also lookup the Literary Digest fiasco of 1936, which catapulted Gallup's election poll credibility. Belson 303 (talk) 03:24, 9 December 2025 (UTC)
As a general rule I feel that polls should usually be cited via secondary sources anyway. Polls are virtually always primary sources; even high-quality ones like Gallup and Pew are still primary. I wrote an essay about this, Wikipedia:Polls are primary sources. As primary sources, we have to be careful not to use them in a way that performs interpretation or analysis, or which uses them to make an argument; the question of "what does this poll mean, exactly?" underlines the problems with using them as primary sources. And especially if someone is pulling one question out of a larger poll that asked many questions - often we see this with those big Gallup or Pew polls, where they'll ask 100 questions and someone will pull one out and slap it into an article in a way that implies a clear conclusion. That's often going to be WP:OR. --Aquillion (talk) 21:44, 14 January 2026 (UTC)
WP:RS/AC - limit to "academic"?
Latest comment: 4 months ago58 comments8 people in discussion
Anyone else think WP:RS/AC should be broadened to encompass not just academic or scientific consensus, but also non-academic/non-scientific consensus (e.g., statements asserting consensus outside science or academia, e.g. the consensus of food critics, sportswriters, journalists, pundits, etc. etc.)? The principle seems sound, I don't see a reason to limit it to just academic/scientific fields. Levivich (talk) 18:17, 13 January 2026 (UTC)
That would make me very nervous. I certainly would be wary of asserting political consensus in wikivoice. We all are subject to living in political bubbles, where “everyone” (ie all those we have contact with) agree on a political stance, not realizing that dissenting opinions might be larger than we think. We can mistake a consensus within our bubble for a consensus on a broader scale. Our current NPOV policy asks us to step outside our bubbles, and examine other viewpoints (and present them in accordance to how prevalent they actually are). Blueboar (talk) 18:37, 13 January 2026 (UTC)
I agree with you, which is why I think the prohibition in RS/AC should be expanded to also apply to, e.g. "political consensus" and other types of consensus. Right now, RS/AC doesn't apply to "political consensus" or other types. So editors are prohibited from writing in mainspace that there is academic consensus without an RS explicitly stating so, but editors are currently not prohibited from writing that there is political consensus without an RS explicitly stating so. I think they should be so prohibited. Levivich (talk) 18:57, 13 January 2026 (UTC)
How often does this come up? A search for "political consensus" turns up a couple hundred articles, but they're mostly not saying "There is a political consensus that ____". It's mostly used in historical contexts and could often be replaced by words like "the general belief among politicians" or "didn't have the votes to ____" or "Party leaders should pay attention to what the rank-and-file politicians say". WhatamIdoing (talk) 21:26, 13 January 2026 (UTC)
That example is a footnote saying "There is substantial consensus across multiple reliable sources that the SUV drove forward and to the right" (followed by a long list of sources), after a sentence saying in WP:WIKIVOICE that "The SUV began to drive forward and to the right". This feels like a bit of wikijargon unnecessarily creeping into an article rather than a real attempt at stating "there is a consensus". WhatamIdoing (talk) 20:59, 14 January 2026 (UTC)
I wouldn't expected most claims of political consensus to use the phrase "political consensus", so I don't know how often such a consensus is asserted by articles. However, I do think that it shouldn't be asserted without a source that asserts it. That would be a clear NOR violation, so it is already against policy but further clarification never hurts. Zerotalk01:03, 14 January 2026 (UTC)
I'm having trouble thinking of an example that wouldn't be covered by academic sources (e.g., there was obviously "a consensus" to implement the Marshall Plan, because it got implemented, but we can find plenty of history books about that) or isn't trivial (e.g., there was "a consensus" among the attendees that last night's meeting of the Smallville city council was the most interesting they'd ever seen). WhatamIdoing (talk) 21:35, 14 January 2026 (UTC)
It's not obvious there was a consensus to implement the Marshall Plan (or any other political act/opinion), since most politics happens at the expense of subaltern peoples. DiaperSommelier (talk) 00:03, 16 January 2026 (UTC)
A consensus doesn't mean that everyone agrees. Politicians and governments can have a consensus among themselves even if disadvantaged people don't join in the discussion. WhatamIdoing (talk) 00:46, 16 January 2026 (UTC)
"A process of decision-making that seeks widespread agreement among group members" – In the case of a political decision like the Marshall Plan, the "group members" do not include Subaltern (postcolonialism) people. It's still a consensus among the group making the decision, though.
"General agreement among the members of a given group or community, each member of which exercises some discretion in decision-making and follow-up action" – In the case of a political decision like the Marshall Plan, "the members of a given group" does not include subaltern people. It's still a consensus among the group making the agreement, though.
I'm going to have to disagree on both of those. A group that excludes people from a decision-making process automatically precludes consensus. DiaperSommelier (talk) 15:07, 16 January 2026 (UTC)
Not according to the definitions at Wiktionary, which explicitly include the underlined words.
You picked Wiktionary. Would you like to try another dictionary? Maybe you can find one that says it's only possible to have a consensus if everyone in the world agrees to it, because anything less than everyone in the world would be excluding someone. WhatamIdoing (talk) 21:06, 16 January 2026 (UTC)
The purpose of WP:RS/AC is to say that we can't claim consensus unless we have a source specifically stating that it exists; it warns people against using OR to claim a consensus by eg. just gathering a bunch of scientists who agree, none of whom actually state that there's a consensus across the field as a whole. Is your argument that we can't state that eg. there's a consensus among film critics or the American public even if we have high-quality sourcing explicitly stating it? I think that when that bar is met we obviously have to reflect the sources - a high-quality source saying eg. "the leadership of the colonies eventually reached a consensus that they should reject British rule" or "there is a broad consensus among major tech companies not to do XYZ" or "there is a broad consensus among courts that this law means XYZ" is obviously something we reflect in the article voice. --Aquillion (talk) 21:18, 14 January 2026 (UTC)
I wonder if we're paying too much attention to the word "consensus". Would we ban "the leadership of the colonies eventually reached a consensus that they should reject British rule" but not "the leadership of the colonies eventually agreed that they should reject British rule"? Those mean the same thing. WhatamIdoing (talk) 21:36, 14 January 2026 (UTC)
Well, either way, the underlying logic of WP:RS/AC applies to both - if we want to say "group X reached a consensus" or "group X agreed", we'd want a source saying as much. The real point of RS/AC is that you can't just cite a bunch of individual people to make the argument that there was a consensus or agreement yourself. --Aquillion (talk) 00:44, 15 January 2026 (UTC)
I want to agree with you, but I also wonder: Isn't it kind of sky-is-blue obvious that the American Revolution wouldn't have happened/been successful if the leadership of the colonies didn't agree to reject British rule? But it might also be a little strange to pretend that the Second Continental Congress (they of the United States Declaration of Independence) didn't reach a consensus about the desirability of British rule unless we can find a reliable source that uses that exact word.
I do agree with your point that you shouldn't usually cherry-pick a couple of individuals and pretend that they are The Consensus™. (They might be; they might not be.) We tend to accept this in WP:PARITY situations but not elsewhere. For example: Faith healing says that "Virtually all scientists and philosophers dismiss faith healing as pseudoscience", citing a book that says virtually all scientists and philosophers dismiss topics like faith healing as "either pseudosciences or at least lack the epistemic warrant to be taken seriously". We wouldn't, outside of a FRINGE topic, accept a source saying "either A or B" as a reason to say "A" alone, but the rules get stretched in this area. WhatamIdoing (talk) 02:06, 15 January 2026 (UTC)
I mean, for the specific example of the American Revolution, I suspect it wouldn't be hard to find sources that say that the colonies or leadership or the like agreed on this if we needed to state it. But I still think it's useful to actually find those sources, especially in more controversial areas, because even if the sources definitely exist, the precise context, framing, and wording they use might be relevant to how we cover it (ie. who exactly agreed, when, how, why; who dissented, etc.) --Aquillion (talk) 02:27, 15 January 2026 (UTC)
The colonies example is not representative of the problem because there were a fixed set of them that is not in question. Saying that they all agreed to something is quite different from claiming a consensus among a poorly defined set such as scientists or historians. Zerotalk06:04, 17 January 2026 (UTC)
And… quite a few colonial “leaders” were Tories and sided with the British (or at least attempted to stay neutral). Blueboar (talk) 00:30, 18 January 2026 (UTC)
Both these statements are true but neither detracts from my point that a consensus among a small well defined set is quite different from a consensus among a large poorly defined set. That should be borne in mind as examples are chosen for the policy page. Zerotalk02:06, 18 January 2026 (UTC)
Yes, and I think that's true even if we can find a source somewhere that asserts that "All scientists agree" – a statement too vague to be meaningful, even if what they're agreeing is that water is wet (because sometimes it's not). WhatamIdoing (talk) 02:13, 18 January 2026 (UTC)
MOS:FILMCRITICS does this for movies (The overall critical reception to a film should be supported by attributions to reliable sources that summarize reviews; do not synthesize individual reviews.), although it is somewhat a farce since following WP:BALASP the reader should be able to understand overall critical reception but we need to do the gamesmanship of pretending they don't. Rollinginhisgrave (talk|edits) 01:11, 14 January 2026 (UTC)
I think that the intended meaning behind FILMCRITICS is something like "We've had a problem with editors who cherry-pick the individual reviews they agree with and write a summary based on that, and leave out or downplay all the reviews they disagree with. To avoid that problem and end up with neutral content, we recommend using reliable sources that summarize multiple reviews". WhatamIdoing (talk) 21:02, 14 January 2026 (UTC)
I had the same reading as you, although I don't see a meaningful difference between with the rule and without it.
The POV pusher before: I include reviews supporting my POV. The reader will see a section of overwhelmingly positive/negative reviews and infer this is representative of critical opinion (WP:SYNTH's "state or imply something"). At the start of the section and in the lead, I write "The film was positively/negatively received."
The POV pusher now: The exact same, but now the summary reads "the film received multiple positive/negative reviews", implying the same, but permitted under The Rules™. Rollinginhisgrave (talk|edits) 22:00, 14 January 2026 (UTC)
I was actually thinking about this exact case. It's more complicated than it sounds and I don't think we can write a hard-and-fast rule for it. The problem is that the lead needs to summarize the article. If the article contains a bunch of individual people[/scholars/reviewers/etc] saying XYZ, the only real way the lead can summarize that without listing them all out (which is impractical) is to basically summarize that to "a bunch of people said XYZ." This can sometimes lead to a false impression, but provided the reviews actually make up enough of the body to justify summarization in the lead, I would argue that that's a problem with the body, not with summarizing it in the lead - ie. if the reviews are not representative, or are being given undue weight or something, that needs to be addressed anyway, then the lead can be more representative too. Basically, we do have an obligation to summarize the actual things in the article; if that summary is accurate but ends up implicitly biased by giving undue weight to a particular position, then that's a problem with the weight of the entire article. (Of course, sometimes things that aren't really significant enough for the lead get promoted there - but in a situation where the reception makes up a huge chunk of the body, it's not unreasonable to say that we need to summarize it somehow; we have to be careful with the wording to imply things that go beyond what the individual sources say, but we can't just ignore them in the lead, either.) --Aquillion (talk) 02:36, 15 January 2026 (UTC)
This may be naive, but I wonder if can draw a distinction between the views of reviewers and a group like scientists, and if so whether such a distinction would be helpful.
For the former, by definition all the relevant views in the group are published. If we proportionately represent all significant published views, a summary of these would not be a misleading summary of the views of reviewers, black swan events notwithstanding. The lead of Lemurs of Madagascar (book) would be acceptable.
For the latter, only some of the views in the group are published. If we summarised only the views of scientists that have been published and tried to pass that off as the views of scientists, it would be misleading. Rollinginhisgrave (talk|edits) 03:30, 15 January 2026 (UTC)
Consider this sentence again: "Virtually all scientists and philosophers dismiss faith healing as pseudoscience".
This sentence indicates that "Virtually all scientists and philosophers" have spent time thinking about the subject. This seems doubtful to me. Summarizing only the views of published scientists might not give us an accurate view of all scientists, but the views of relevant scientists should be the ones that we are trying to summarize. Biologists and physicists are scientists. But biologists should be ignored if you're trying to identify a consensus about high-temperature superconductors, and physicists should be ignored if you're trying to identify a consensus about the number of insect species. WhatamIdoing (talk) 01:22, 16 January 2026 (UTC)
I agree of course with "the views of relevant scientists should be the ones that we are trying to summarize", but I'm not sure where it leaves us in terms of SYNTH. Rollinginhisgrave (talk|edits) 01:45, 16 January 2026 (UTC)
===Academicconsensus==={{Shortcut|WP:RS/AC}}Astatementthatallormostscientistsorscholarsholdacertainviewrequiresreliablesourcingthatdirectlysaysthatallormostscientistsorscholarsholdthatview. Otherwise,individualopinionsshouldbeidentifiedasthoseofparticular,namedsources. Editorsshouldavoidoriginalresearchespeciallywithregardtomakingblanketstatementsbasedon[[WP:SYNTH|novelsynthesesofdisparatematerial]]. Statedsimply,anystatementinWikipediathatacademicconsensus exists on a topic must be sourced rather than being based on the opinion or assessment of editors. [[Review article]]s, especially those printed in academic review journals that survey the literature, can help clarify academic consensus.
+
===Consensus==={{Shortcut|WP:RS/C|WP:RS/AC}}Astatementthatallormostscientists,scholars,oranyothergroupholdacertainviewrequiresreliablesourcingthatdirectlysaysthatallormostofthegroupholdthatview. Otherwise,individualopinionsshouldbeidentifiedasthoseofparticular,namedsources. Editorsshouldavoidoriginalresearchespeciallywithregardtomakingblanketstatementsbasedon[[WP:SYNTH|novelsynthesesofdisparatematerial]]. Statedsimply,anystatementinWikipediathatconsensus exists on a topic must be sourced rather than being based on the opinion or assessment of editors. [[Review article]]s, especially those printed in academic review journals that survey the literature, can help clarify academic consensus.
Not sure. The general issue is saying members of a group hold views without sources directly attesting to it. This guideline addresses a subset of these cases, when "all or most" members of a group hold a view. This seems weird here: why have a guideline saying "you need reliable sources directly stating there is a consensus among pundits that" but not one saying "you need reliable sources directly stating that the majority of pundits hold the view that"? A consensus among scientists seems particularly meaningful in a way it is not in other groups. Rollinginhisgrave (talk|edits) 02:41, 16 January 2026 (UTC)
I would support a further expansion that applied the "must-have-a-source-directly-stating-it rule" not just to consensus but also the majority of or other similar statements ("the minority", "half", or whatever). Would you support that?
Also, can you expand on your last sentence: why is consensus among scientists/scholars meaningful in a way that consensus among other groups (e.g., film critics, journalists, sportswriters, etc.) isn't? Is it because "scientific consensus" is a 'thing' (like a recognized concept in the world) but, say, "journalism consensus" isn't, or are there other reasons? Because my response to that is that the consensus of critics, or sportswriters, or Olympic judges, are 'things' in the same way as scientific consensus. Would you support an expansion that applied not to all "groups" but just to some specified subset (specification TBD)? Levivich (talk) 02:59, 16 January 2026 (UTC)
Hmm: "You need a reliable source saying that adherents of <belief> hold <belief> before you can say that they believe <that belief>."
I dunno about this. What if the group is defined by adherence to a view? Is there a world in which supporters of Donald Trump don't hold the view that Trump is worthy of their support? WhatamIdoing (talk) 07:05, 16 January 2026 (UTC)
Well… there ARE “hold your nose and vote the party line” Republicans, who don’t think Trump is “worthy” of support, but support him anyway. Blueboar (talk) 15:22, 16 January 2026 (UTC)
Is your point that if we write "there is a consensus of Trump supporters" we should not need a source saying "there is a consensus of Trump supporters", or we should need a source saying that? Because I don't understand the discussion about different kinds of groups. Is there anybody here who thinks we should be able to write "there is a consensus of [group]" without a source directly stating "there is a consensus of [group]", for any group, and if so, for what group(s) and why? Levivich (talk) 22:49, 16 January 2026 (UTC)
I don't think we should treat the phrase "there is a consensus that ____" as special.
Therefore, "There is a consensus among Trump supporters that Trump is good" is the same as "Trump supporters support Trump".
Not sure if your comment "I don't think we should treat the phrase "there is a consensus that ____" as special." is intended to cover the current guideline, but a consensus of SMEs will generally have a large bearing on what we can put in wikivoice. Rollinginhisgrave (talk|edits) 08:39, 17 January 2026 (UTC)
I don't think we want this for "Trump supporters support Trump" statements, but I do think we want this for "All scholars agree" statements. WhatamIdoing (talk) 05:05, 17 January 2026 (UTC)
WAID, do you really think we will write "Trump supporters support Trump" in an article? Can we stick to realistic examples? "There is a consensus among Trump supporters that the 2020 election was stolen" = should require a source directly stating that or no, in your opinion? Levivich (talk) 05:22, 17 January 2026 (UTC)
I would expect "There is a consensus among Trump supporters that the 2020 election was stolen" to have the same kind of source as a statement like "Trump supporters believe that the 2020 election was stolen". WhatamIdoing (talk) 05:51, 17 January 2026 (UTC)
I'm honestly confused because you seem to want to discuss the issue, but you are not answering any of my questions to you. Levivich (talk) 06:21, 17 January 2026 (UTC)
Here are the answers to your questions:
do you really think we will write "Trump supporters support Trump" in an article? – Yes, I really think that statements like this appear in articles. Specifically, they often appear in the very first sentence of articles.
Can we stick to realistic examples? – I believe we are, but I sometimes find that "unrealistic" examples illustrate the principles more clearly than realistic ones. For one thing, if you have an example about (e.g.,) dragons that live on Mars, nobody can sidestep the question by saying that they found a heretofore unknown source that makes the question moot, or that there is some obscure detail that makes the example fail. Obviously, at AFD or in a specific dispute about an article, finding that source is the best thing you can do. But in a policy-type discussion, having unavoidable constraints is helpful. If you think that editors will be tempted to avoid the real question by invoking the Spell of Status Quo, then you can set your "unrealistic" example to specify that the entire article is only ten minutes old and the edit war over the specific point started before the article was 30 seconds old.
"There is a consensus among Trump supporters that the 2020 election was stolen"=should require a source directly stating that or no – I believe that all plausible forms of this require an inline citation, but that the phrasing "There is a consensus" does not require any additional or special sourcing.
In particular, for that last item, it is important that the source cover the stated group (Trump supporters) and not a small subset (e.g., one individual, one organization, Twitter users, etc.). For example, if we have three sources that describe Alice, Bob, and Chris' individual beliefs, then:
Y Alice, Bob, and Chris believe _____.[Alice][Bob][Chris]
Y Some experts such as Alice, Bob, and Chris believe _____.[Alice][Bob][Chris]
N Experts believe _____.[Alice][Bob][Chris]
N There is a consensus that _____.[Alice][Bob][Chris]
Thank you:-) What do you think about: A statement in Wikivoice attributing a viewpoint to a group (e.g., "consensus amongst scientists that ...", "a majority of film critics agree ...") must be supported by a reliable source directly attributing that viewpoint to that group.? Levivich (talk) 02:58, 18 January 2026 (UTC)
What about approaching it from the opposite side? That would sound something like this:
Do not attribute a viewpoint to a group (e.g., "consensus amongst scientists that...", "a majority of film critics agree...") if your sources are merely examples of individual members of the group expressing that viewpoint.
I still think this would cause problems for WP:PARITY sources (where one blog post from one mainstream science-oriented person is taken to be representative of all right-thinking rational scientists), but it might be clearer about the key point, which is that "Alice Expert said ____" is not the same as "All experts say ____", even though everyone knows that Alice is an expert in expertise. WhatamIdoing (talk) 20:31, 18 January 2026 (UTC)
I think it's better the other way: tell people when they can say it, rather than one instance in which they can't say it, because there are many more instances in which they can't say it, than in which they can say it. Levivich (talk) 01:27, 19 January 2026 (UTC)
"You can't say this unless you have exactly this source" is the most restrictive rule, and it doesn't appear to be compatible with community practice around PARITY. WhatamIdoing (talk) 02:23, 19 January 2026 (UTC)
I wonder why the community treats "There is a consensus among scientists that ____" as different from "Scientists believe that ____" (because IMO we do treat these differently). WhatamIdoing (talk) 20:25, 18 January 2026 (UTC)
MOS:WEASEL explains: The examples above are not automatically weasel words. They may legitimately be used in the lead section of an article or in a topic sentence of a paragraph when the article body or the rest of the paragraph can supply attribution. We can say "scientists have criticized XYZ for..." or the like in the lead without naming any individuals, provided we have the attributions in the body. I'd be strongly opposed to changing that - it's important because we do need to actually summarize prominent views in the lead, when they are sufficiently WP:DUE to reasonably have a large presence in the body; and requiring a massive list of individual names in the lead would be clunky and would discourage actual summaries. "You can't summarize this in the lead at all without a secondary source performing the summary" is wrong, we just have to word it carefully to avoid implying that the scientists we summarize are a specific proportion of all scientists. --Aquillion (talk) 17:11, 19 January 2026 (UTC)
Latest comment: 4 months ago2 comments2 people in discussion
Just something I feel should be added, but I feel like there should be a section on weither or not first hand accounts should be considered reliable sources. An example would be Misturi Yoshida's Requiem for Battleship Yamato which is his account of her last sortie in 1945. Panzer VI13:58, 21 January 2026 (UTC)
Latest comment: 4 months ago12 comments7 people in discussion
A disagreement at WP:DRN is hinging on debate as to the meaning of WP:EXCEPTIONAL. An editor has provided an example of a claim they believe falls under this: "if someone says, "In my opinion, the Earth is flat", saying they hold that opinion would still be exceptional." This is an analogy with someone saying "this video game is the best". Any input here is appreciated. Rollinginhisgrave (talk|edits) 14:16, 23 January 2026 (UTC)
Specifically, I was responding to a claim that WP:EXCEPTIONAL applies only to statements of fact, and not statements of opinion. Nothing in the guideline states this (it just says "Any exceptional claim", and includes "Surprising or apparently important claims"), so a clarification is needed. Phediuk (talk) 14:23, 23 January 2026 (UTC)
But "the world is flat" is not an opinion, even if someone calls it an opinion. It's a counterfactual. Opinions are a kind of non-T/F claims. FactOrOpinion (talk) 15:23, 23 January 2026 (UTC)
I should really like to note for the record that you did in fact not [respond] to a claim that WP:EXCEPTIONAL applies only to statements of fact, and not statements of opinion. The comment you responded to said that "X is the best" is not a statement of fact but a statement of opinion (but would be exceptional if it were a statement of fact), that "X is considered the best" is a WP:YESPOV statement of opinion, and that "X is considered the best by Y" is standard WP:INTEXT attribution which does not require multiple sources. TompaDompa (talk) 15:32, 23 January 2026 (UTC)
Attribution doesn't mean that other relevant policies can be ignored. "In my opinion Bob is a murderer", is still as problematic as "Bob a murderer". Putting "the Earth is flat" in Earth would still be a problem whether it was attributed or not. Who is so important that their claim of the Earth's flatness was appropriate for that article? If the claim is exceptional then the person making it should be as well, exceptional claims require exceptional sources. In the case of attributed claims the source is the person making the statement. -- LCU ActivelyDisinterested«@» °∆t°14:46, 23 January 2026 (UTC)
Thanks AD. As I understand it, the relevant claim is "X video game is considered the best by multiple magazines". Maybe @Phediuk can affirm / deny that and rephrase. Do you view such a case as requiring exceptional sourcing? Rollinginhisgrave (talk|edits) 14:52, 23 January 2026 (UTC)
"X is the best video game" would be an exceptional claims, "Y says that X is the best video game" is no different. If this was the opinion of a video game grandee, Miyamoto / Kojima / Newell / Carmack / etc, then I could see the argument for its inclusion. But if it's some random games journalist then no, it's putting undue weight on that journalist opinion. Also opinions need to be put in context, only including one opinion when other having differing opinions is a WP:NPOV issue.
"X video game is considered the best by multiple magazines" would require a source that says that, it's meant to be directly verifiable. It's not an exceptional claim though, the claim is that multiple magazines have said that X game is the best not that X game is the best. -- LCU ActivelyDisinterested«@» °∆t°17:01, 23 January 2026 (UTC)
I don't think that "This video game is considered the best by multiple magazines" is exceptional.
The concept of WP:EXCEPTIONAL comes from Extraordinary claims require extraordinary evidence, which mostly starts with things in the range of "I used the natural sciences to prove the existence of a supernatural being" and then gets stretched just a little to include things like "Wonderpam cures every kind of cancer" and "I did cold fusion in my kitchen".
By comparison, there are thousands of magazines in the world, and the fact that "multiple" of them (which could be as few as two) have expressed an opinion on about the "best" video game is ordinary and expected.
I think there is a big difference between "X is the best" and "Y said that X is the best". If there is a source that says something like "Several magazines named this as the best video game of 2025" (i.e., so the claim is verifiable), then the question isn't whether it's extraordinary, but instead whether it's WP:DUE. WhatamIdoing (talk) 23:42, 23 January 2026 (UTC)
It feels like the question at hand is not whether something is exceptional, but whether it is due weight. Why is the particular opinion worth highlighting? Is the opinion itself commented on in another source? CMD (talk) 04:45, 24 January 2026 (UTC)
Source needed for Annapurna fatality rate
Latest comment: 4 months ago3 comments3 people in discussion
I am raising this for sourcing discussion only.
The article includes a statement regarding the fatality rate of climbers on Annapurna that is currently marked with a citation needed tag. I have not found a high-quality secondary source cited directly for this specific figure.
I should note that I have a conflict of interest, as I am associated with the website linked below, so I am not proposing its direct addition as a reference.
That source contains a summary of historical ascent and fatality figures for Annapurna compiled from publicly available mountaineering records. It may be useful for verifying basic, non-controversial factual data if editors feel such information is required, though a higher-quality secondary source (such as an academic publication, alpine journal, or authoritative mountaineering database) would be preferable.
Latest comment: 4 months ago5 comments5 people in discussion
If a reliable source republishes and cites material from an explicitly unreliable publisher, should we (a) assume due diligence on the part of the former and allow its use, or should we (b) assume the former otherwise trusts the original source more than (we do/it should) and not use the republished info? — Fourthords|=Λ=|20:28, 31 January 2026 (UTC)
Well care is certainly worthwhile! But yes if a reliable source republishes something as fact then the first assumption should be that they have checked it. NadVolum (talk) 20:35, 31 January 2026 (UTC)
The only certain answer is "it depends". Even a stopped clock is correct twice a day, and the world's worst publishers sometimes publish good information. So if Fake News publishes something that accidentally happens to be good, and The Daily Trustworthy re-published it, there is a reasonable chance that it's good. But if AllNewsAggregator re-publishes it – well, they re-publish a lot of stuff, and there's no reason to think that this was specially selected or checked at all.
One thing I'd suggest is that you look at is the amount of re-publishing that the second publication does, and the types of things they re-publish. There was at least one US newspaper that reprinted huge numbers of press releases. This was actually handy for us (if you want to see whether a supposed news article was effectively written by the PR agency, then you could look up the original), but you had to be careful to make sure that anything on their site was from their actual newspaper or from a reputable wire service, and not an automated post of a press release. If your second publication does a lot of re-publishing, I'd not assume that the second publication spent time checking it. WhatamIdoing (talk) 20:38, 31 January 2026 (UTC)
It will depend on how they republish and the exact wording they use. If they're just republishing the exact same report in the way a news aggregator would then the reliability remains with the original source. If they're repeating the report that was made by another source in their own reporting then the language they use is important. News organisations are very aware of the language they use and use it very specifically "X is Y" and "The Chronicle says X is Y" are very different statements. One is reporting that X is Y, the other is saying that if this isn't true the Chronicle is to blame. The important part is that the language of the second type of statement only supports "The Chronicle claim that X is Y" not that "X is Y". -- LCU ActivelyDisinterested«@» °∆t°22:20, 31 January 2026 (UTC)
An "explicitly unreliable" publisher is a rare bird, although some RfCs have declared a publisher "generally unreliable" for certain facts. "Generally" means not specifically or "most of the time". If anothersource.com has endorsed generally.com by quoting it, then generally.com's article is reliable and is the real source. So WP:RS/QUOTE says the cite should be of generally.com if possible, though it's probably best to cite anothersource.com too. Peter Gulutzan (talk) 01:30, 3 February 2026 (UTC)
Latest comment: 4 months ago21 comments8 people in discussion
At present, WP:HQRS (High Quality Reliable Sources) redirects to WP:RS with no specific direction as to the distinction. Following a recent discussion at WT:FAC, there is a proposal to upgrade WP:HIGHQUALITY (part of an essay) to a guideline as part of WP:RS. This would be in addition to the currently existing guidelines.
The intent of this guideline would be to consolidate the existing guidance existing in various places about source quality. It would support assessment processes where source quality is measured above the level of WP:RS -- most frequently and obviously WP:FAC, but it may also be of use to processes like A-class review as operated by certain WikiProjects. It would also be of value to editors looking to improve their articles to a higher and more academic standard, whether or not they end up putting them through formal certification processes.
The text as currently written is transcluded below. In the interests of workshopping the proposal, it might be useful to discuss (inter alia) the following:
Should there be guidance on what constitutes a "high-quality" reliable source with the status of a guideline?
If so, what should that guidance consist of?
Does the current framing at WP:HIGHQUALITY constitute a reasonable starting point, and what changes (if any) should be made to it before making it a guideline?
At present, WP:HQRS (High Quality Reliable Sources) redirects to WP:RS with no specific direction as to the distinction is not quite true. A few days ago, I went ahead and changed the HQRS shortcut to point to WP:HIGHQUALITY. That solved the immediate problem; now when you click WP:HQRS you at least get to something explaining the difference. I'm fine with elevating that to a guideline and/or moving that text to be in WP:RS, but I don't think it's essential. RoySmith(talk)13:21, 7 February 2026 (UTC)
In addition to the usual reliability requirement, the text of Featured Articles (FAs) must be "verifiable against high-quality reliable sources". Note that while WP:FACR requires high-quality reliable sources, WP:GACR only requires reliable sources. This means that sources which were acceptable in a Good Article will not always pass the higher bar enforced at the FA level. Reviewers with some expertise in the subject of the article will more easily be able to determine whether the sources used meet the required quality standard. The general questions on which all reviewers should try to satisfy themselves are:
Do the sources represent the best available for this particular subject?
Is the source that supports each point the most appropriate for that point?
Are the main sources reasonably up-to-date, and therefore likely to represent the most recent scholarship? Older sources, particularly contemporaneous primary sources, are often appropriate, but the nominator may need to explain why they've been chosen.
In the case of anything contentious, are primary sources being used in accordance with the secondary literature?
Do the sources appear collectively to provide a comprehensive account of the subject, or is there over-reliance on a particular source or group of like-minded sources? Reviewers should be aware that even the highest-quality sources can be used selectively in a way that affects the neutrality of the article.
Making these judgements takes time, and raising them will sometimes invoke the ire of nominators, but if reviewers have any doubts about sources' quality, individually or collectively, they should pursue the matter.
Discussion re Proposal to add WP:HIGHQUALITY to WP:RS
My participation in the last discussion was on clarifying what HQRS was intended to mean, so I'll raise my two concerns (really one) here:
Some facts that we may really want to include (e.g. when a company was founded) may only be included in a poor or marginal source, such as the company's about us page. Are there any circumstances where following this guideline we are excluding these facts that would generally have acceptable sourcing on a non-FA candidate?
If information is not in the best sources, but it is in the non-best sources and the two are not contradicting, can we include information from the non-best sources? HQRS makes claims not just on reliability but also DUE weight, is this what we want? Literature on social phenomenon may be dominated by critical perspectives, and literature on chemical processes dominated by aspects of interest to industry.
The broader concern is this being deployed to exclude information that is sufficiently reliably sourced, that readers want and we should want in an encyclopaedia. Rollinginhisgrave (talk|edits) 11:44, 7 February 2026 (UTC)
Yes, I think that excluding ordinary information would be the net effect, and I therefore think that's not a good idea.
I think this should be turned into an essay, and that anything we say should (a) be in WP:BESTSOURCES and (b) should be written in a way that is applicable to all articles, not just FA and GA candidates. The "over-reliance" point in particular is a NPOV situation. WhatamIdoing (talk) 19:02, 7 February 2026 (UTC)
Ideally, editors will be able to find and use the best available sources for the particular subject. The best sources will vary by field (e.g., a reference work for technology or a mass-market book for a celebrity), but they should be reasonably up-to-date and collectively provide a comprehensive account of the subject, without over-reliance on a particular source or on a group of like-minded sources. Editors should aspire to having most of the article cited to secondary sources that are independent of the subject..
I think that pushing this sort of material into WP:RS is WP:CREEPy wishful thinking. Wishful thinking both because so many of our articles use sources that are marginally or not reliable that polishing our language on what perfection would look like is paying attention to the wrong end of the curve, but also because the sort of major-publisher biography or comprehensive reference book that this language asks for only rarely exists for our subjects, unless we want to heavily scale back our ambitions how comprehensive our coverage should be and cover only first-world mainstream-media topics. These sorts of sources are good to use when we can, but I think this belongs in an essay, not in a guideline that could easily become used as a cudgel to eliminate specialized, technical, or non-first-world topics. —David Eppstein (talk) 01:43, 8 February 2026 (UTC)
By defining "high-quality" sources as "the best available sources for the subject", the proposed guideline would help encourage the creation of better articles. It would prompt editors and reviewers to show that they have carried out careful source research and have considered source quality, rather than relying only on what is easiest to access. This would also strengthen our projects in comparison to AI-generated articles, which often depend heavily on material that happens to be available online, and therefore often overlook recently published work from leading academic institutions. Finally, the assumption that only first-world media can produce the "best available" sources seems questionable. Source quality should be judged in context and in relation to the subject, not assumed from geographic origin. Borsoka (talk) 02:46, 8 February 2026 (UTC)
Yes, but if you look at the preceding comments, we're talking about putting this text into WP:BESTSOURCES, which is (a) in WP:NPOV instead of WP:RS and (b) a section that we've talked about re-writing for a couple of years now, because it provides almost no information on what the best sources are. WhatamIdoing (talk) 04:16, 8 February 2026 (UTC)
In re "By defining "high-quality" sources as "the best available sources for the subject"":
One thing I have considered for this guideline and/or WP:V is introducing the concept of a "good enough" source. Yes, if you've got plenty of high-quality sources, then use them! But also don't revert a source that meets the minimum reliability standards just because it's not The Best™ Possible Source. WhatamIdoing (talk) 04:18, 8 February 2026 (UTC)
Yes -- my worry about "what if there aren't any great sources?" would be the other way to that raised by Rollinginhisgrave above -- I've worked on a few biographical articles where there are often-told stories about someone that don't make it into the scholarly literature, or where key details (such as place of birth) are found in popular/non-scholarly sources but not repeated by the academic presses. One possible explanation for this is that the evidence behind those stories/facts isn't great -- so we probably shouldn't repeat "facts"/accounts that are only stated specifically in marginally reliable sources, when plenty of better sources which might be expected to report them don't. Instead, we shouldn't pass a judgement at all. UndercoverClassicistT·C10:04, 8 February 2026 (UTC)
If you're writing about a subject with lots of good sources (e.g., World War II), then you should use multiple sources, and basically all of your sources should be high-quality sources (e.g., scholarly books).
In such a scenario, it would be possible but not desirable to write an entire article from a biased set of high-quality sources. For example: all scholarly books, but they're all British. Or all scholarly books, but they all view the war through an economic lens. Or all scholarly books, but they're all about technological advances.
Instead of half a dozen excellent all books about Nazis, we instead want one excellent book about money, one excellent book about Britain, one excellent book about Japan, one excellent book about Italy, one excellent book about politics, one excellent book about anti-Semitism, one excellent book about science and engineering, one excellent book about armies, navies, and air forces, etc. WhatamIdoing (talk) 22:46, 8 February 2026 (UTC)
When I see the idea that the best sources are those "collectively providing a comprehensive account of the subject", I think of how more marginal / primary sources are deployed to fill gaps in the secondary/tertiary literature. In that context, would these be considered best sources? Not sure if this is just doing word meaning debate or substantive. Rollinginhisgrave (talk|edits) 04:25, 9 February 2026 (UTC)
I don't know if any editors will interpret it that way. But I'd start by saying that "the best sources are those" is the first problem. Collectively providing a comprehensive account isn't a definition of "the best sources". It's how you will know whether your collection of sources is incomplete. WhatamIdoing (talk) 06:10, 9 February 2026 (UTC)
I think one question is what "high quality" means. I would say that a company's About Us page is a high quality source for certain aspects (e.g who the staff is, their stated mission). It would not be a good source for certain other things. OK, I correct myself - "high quality" is relative to the to-be sourced fact, it's not an intrinsic property of a source. Donald Trump's posts are HQRS for his own statements and claims, for instance, barring questions about authenticity. Jo-Jo Eumerus (talk) 11:28, 8 February 2026 (UTC)
We have sometimes used the word authoritative to describe that kind of source. I think this is a statement is about the intrinsic property of the source. It's more like being "generally reliable" than being reliable in context for the particular statement it supports. WhatamIdoing (talk) 22:49, 8 February 2026 (UTC)
There's no one size fits all. For example, if a national newspaper raises concerns that a prestigious.edu's professor's website said perpetual motion machine is possible; citing the archive of that website alongside the newspaper article is a very reliable source for the fact that a prestigious.edu's professor's website indeed made such a statement. Graywalls (talk) 06:39, 9 February 2026 (UTC)
Yes, but I think the idea of "high-quality sources" (as well as WP:BESTSOURCES) is addressing the editing process at a different level. It's not about what the best source for a given claim is; it's about what you'd tell someone if they're trying to write a whole article. If Perpetual motion were a red link, you wouldn't start by looking for a webpage at prestigious.edu, or even for newspaper articles. You'd start by looking for a history of the idea, or a physics textbook that addresses its impossibility. Those would be your high-quality sources, and the newspaper tutting about how Prof. Speculator said something on the internet would be quite far down the list – suitable for giving an example of the idea never quite dying off, but not really a high-quality source for the article. WhatamIdoing (talk) 07:00, 9 February 2026 (UTC)
There is currently a discussion at Wikipedia talk:Reliable sources/Perennial sources regarding re-categorising self-published platforms, along with amending the sections in this guideline Definition of a source and User-generated content. Please note this is not an RfC, this a drafting process for an RfC. The thread is Drafting the RFC question about platforms. CNC (talk) 10:33, 21 August 2025 (UTC)
You are invited to join the discussion at Wikipedia talk:Citing sources § WMF seeking feedback on Reference Check. Sdkb-WMF talk 18:09, 25 September 2025 (UTC)
Not done: please provide reliable sources that support the change you want to be made. Happy Editing -- IAmChaos 22:37, 29 November 2025 (UTC)
Alice, Bob, and Chris believe _____.[Alice][Bob][Chris]