Wikipedia talk:Identifying reliable sources

From Wikipedia, the free encyclopedia
  (Redirected from Wikipedia talk:RS)
Jump to: navigation, search
Frequently Asked Questions (FAQ)
Where should I ask whether this source supports this statement in an article?
At Wikipedia:Reliable sources/Noticeboard. Don't forget to tell the editors the full name of the source and the exact sentence it is supposed to support.
Do sources have to be free, online and/or conveniently available to me?
No. Sources can be expensive, print-only, or available only in certain places. A source does not stop being reliable simply because you personally aren't able to obtain a copy. See Wikipedia:Reliable sources/cost. If you need help verifying that a source supports the material in the article, ask for help at Wikipedia:WikiProject Resource Exchange or a relevant WikiProject.
Do sources have to be in English?
No. Sources can be written in any language. However, if equally good sources in English exist, they will be more useful to our readers. If you need help verifying that a non-English source supports the material in the article, ask for help at Wikipedia:Translators available.
I personally know that this information is true. Isn't that good enough to include it?
No. Wikipedia includes only what is verifiable, not what someone believes is true. It must be possible to provide a bibliographic citation to a published reliable source that says this. Your personal knowledge or belief is not enough.
I personally know that this information is false. Isn't that good enough to remove it?
Your personal belief or knowledge that the information is false is not sufficient for removal of verifiable and well-sourced material.
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.
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.
What if the source is biased?
Depends on how it is used. In all cases, Wikipedia articles are required to be neutral, so Wikipedia cannot simply parrot a biased or non-neutral source's opinion as if it were fact. But, sometimes "non-neutral" sources are the best possible sources for supporting information (with due weight) about the different viewpoints held on a controversial subject.
Does every single sentence need to be followed by an inline citation?
No. Only four broad categories of material need to be supported by inline citations. Editors need not supply citations for perfectly obvious material. However, it must be possible to provide a bibliographic citation to a published reliable source for all material.
Are reliable sources required to name the author?
No. Many reliable sources, such as government and corporate websites, do not name their authors or say only that it was written by staff writers. Although many high-quality sources do name the author, this is not a requirement.
Are reliable sources required to provide a list of references?
No. Wikipedia editors should list any required sources in a references or notes section. However, the sources you are using to write the Wikipedia article do not need to provide a bibliography. Most reliable sources, such as newspaper and magazine articles, do not provide a bibliography.

Request for comment: Are "in popular culture" entries "self-sourcing" or do they require a reference under Wikipedia:Verifiability and Wikipedia:Identifying reliable sources?[edit]

I have started an RfC on whether "in popular culture" entries are "self-sourcing" or, conversely, require a reference under Wikipedia:Verifiability and Wikipedia:Identifying reliable sources.

The RfC is at Wikipedia talk:Verifiability, so discussion is centralized there. Comments are welcome. Neutralitytalk 23:56, 27 September 2015 (UTC)[edit]

Opinions are needed on the following matter: Wikipedia talk:WikiProject A WP:Permalink for it is here. This source has affected a lot of articles, and this discussion is important. Flyer22 (talk) 20:30, 7 October 2015 (UTC)

Dave mattocks Drummer[edit]

Hi Just want to point out in your article about Dave mattocks there"s no mention in the article that he Played and recorded with Bill Nelson"s Red noise? — Preceding unsigned comment added by (talk) 14:27, 20 October 2015 (UTC)

RfC announce: What does Wikipedia:Identifying reliable sources (medicine) cover?[edit]

There is a request for comments at [ Wikipedia talk:Identifying reliable sources (medicine)#What does MEDRS cover? ].

At issue is whether the lead paragraph OF WP:MEDRS should remain...

"Wikipedia's articles are not medical advice, but are a widely used source of health information. For this reason it is vital that any biomedical information is based on reliable, third-party, published secondary sources and that it accurately reflects current knowledge."

...or whether it should be changed to...

"Wikipedia's articles are not medical advice, but are a widely used source of health information. For this reason it is vital that any biomedical and health information is based on reliable, third-party, published secondary sources and that it accurately reflects current knowledge."

This has the potential to change the sourcing policy from WP:RS to WP:MEDRS on a large number of Wikipedia pages, so please help us to arrive at a consensus on this issue. --Guy Macon (talk) 06:09, 3 November 2015 (UTC)

Updated Bio for Denis Ryan[edit]

NEWSORG and science[edit]

Recently, news reports have come out declaring that a Nigerian mathematician has proved the Riemann hypothesis. For example, the BBC, The Telegraph, The Independent, CNN. It's blatantly clear that these stories are unsubstantiated nonsense. (At worst, this entire affair is an obvious hoax.) Aside from pointing to WP:SCHOLARSHIP and WP:REDFLAG, is there some bright line that we can point to regarding the unreliability of such otherwise well-regarded "news" outlets on scientific matters?

I've seen my fair share of news sources getting science just completely wrong (in some cases, just outright fabrication). In such cases, it's pretty clear to reasonably scientifically literate editors that this is the case. However, the trouble with news media is that it tends to bring a different crowd of editors to scientific pages: those that have had no exposure to the subject, apart from the news source that they read. In some cases, when a news outlet picks up a story, that story is immediately and uncritically copied by every other news source, without performing any additional fact-checking. So, we often get a situation where there is an overwhelming number of "reliable" sources, and editors without much scientific literacy lobbying for inclusion of content like this based on prevalence in those "reliable" sources. ⇔

For this reason, it seems like we need to firm up some guideline to clarify our collective position on using such sources. Editors without much scientific literacy tend not to be big on nuances, like: "The BBC is not reliable for scientific matters." Or, "That's a redflag claim. We need high-quality scholarly sources." I think some bright-line rule, to which we can refer such editors, would helpfully clarify our position and put an end to unproductive time-wasting. Sławomir
14:49, 18 November 2015 (UTC)

Taking the BBC's article as an example, the key word that actually justifies them is "claims". Similarly, the CNN link is not outright sayng its solved, just that this professor claims to have solved it. So us using those sources and just stating that there's a claim it was solved is fine, doesn't break RS in any way. I will agree that there are likely some newspapers that might have taken it at face value but the highest quality sources here clearly are not stating 100% the problem is solved, only that the professor and this one conference think it is, so unless there's something else missing here, these seem to capture the situation appropriately. --MASEM (t) 16:59, 18 November 2015 (UTC)
I don't think this is a serious enough piece of crankery to even mention in our article despite the newspaper buzz. I think you are very wrong to suggest doing so. Even the part you state, that he "claims to have solved the hypothesis", is not reliable. What this professor appears to have actually done is put together a collection of crank attempts at the Riemann hypothesis by other people, on a preprint site. But you can't get out of this sourcing issue by pointing to weasel-worded articles and saying that the weasel wording saves their reliability. Take a look at this story linked from Talk:Riemann hypothesis. It's in a mainstream newspaper. Its headline says ""Professor becomes a millionaire". This is false — the Clay math prize has definitely not been awarded. And yet, it's in a newspaper. That doesn't mean it can be included in our article. It means our standards of reliable sourcing for this sort of story are bad. —David Eppstein (talk) 17:19, 18 November 2015 (UTC)
Two points:
  • First, keep in mind that we have outright rejected any claims of sourcing based solely on the headline of a news article. That's nearly always written by a copyeditor and not the author, and they're going to pick up something to entice readership and potentially be wrong. There's no reliability to that. We go by the article body.
  • Second, when looking at the range of sources, with sources like the BBC and CNN at the top end, it's clear that the higher quality sources are clearly aware this is only a claim made by the professor and not the actual proof. I'm sure if I poke around I can find a smaller paper that outright says in the prose the problem is solved, he got the million. That means there's conflicting information, and there we have to use the more reliable sources, which is clear that the problem is only claimed to be solved, and not an absolute. Our article should reflect those sources.
  • Now, whether this should be included, I don't know. I thought I have remembered a few cases of people having thrown proofs at the Millennium Prize Problems that have gotten mainstream attention and later proven wrong; for that, I would think that would cover the notable failures (whether well-intended but flawed solutions, or straight up hoaxes) as part of the demonstration of the importance of providing a solution to these proofs offers. I don't see that on these articles, but maybe it was decided not to. If so, then yes, we don't need to add this at all. --MASEM (t) 17:30, 18 November 2015 (UTC)
You didn't actually read the article I linked to, did you? You just picked up on the word "headline" in my comment. Because the first actual text sentence of the article, "A professor has received a prize of $1 million...", is flat-out false. —David Eppstein (talk) 18:18, 18 November 2015 (UTC)
A lot of today's online news journalism seems to have been written by interns on zero hours contracts. An article about the Riemann hypothesis should be written by someone who knows the score and the long history of people claiming to have found a proof for it. There has been a failure of basic journalistic standards in a lot of the reporting of this latest claimed proof, with some sources getting the facts wrong, such as the Irish Independent story. In its current form, the claimed proof by Dr Opeyemi Enoch is no more notable than the many other similar claims floating around on and "publish anything you like" websites. This type of website is not much more reliable than a Facebook page, as been pointed out to various claimants at Talk:Riemann hypothesis‎ in the past. Some journalists should go to boot camp and be licked into shape before writing anything about academic research, as they are likely to go for whatever makes the most eye-catching headline, as we have seen here.--♦IanMacM♦ (talk to me) 18:23, 18 November 2015 (UTC)
I did allude to the fact that I suspect (and rechecking the Independent, confirmed) that less-reputable news sources would say he got the prize and the problem was solved. That's why there's the strength of journalistic integrity to consider. The BBC article nor the CNN article says he got it, just that he's claimed to have solved it, this conference claims its valid proof, but the prize-awarding institute has to review it. The BBC + CNN got it right. My point is that if we were to include this, we'd obviously be wanting to cite the BBC and CNN articles, over the Independent, and have to recognize that new editors are going to use less-reputable papers like the Independent to make the claim the problem was solved. --MASEM (t) 20:31, 18 November 2015 (UTC)
@Masem: Since these sources no longer appear to be engaging in fact-checking, they are no longer reliable. Any random crank can "claim" anything. Part of the responsibility of real science journalism is to interview independent experts. That didn't happen here, and even a moment of looking into the matter reveals that the whole affair is an obvious hoax. (The paper in question is just a plagiarized copy of an old preprint; it doesn't even have Enoch's name on it.) Journalists don't get to hide behind the weasel-word of "claim" any more than Wikipedia editors do. The sources are saying something that is clearly nonsense and obviously false. Sławomir
18:46, 18 November 2015 (UTC)
What's false about [1]? If that story is false, we should be able to point to a story that says so. If it is undue (or too soon) that's something else. Alanscottwalker (talk) 19:27, 18 November 2015 (UTC)
Saying that the claim by Dr Opeyemi Enoch is an obvious hoax might run into problems with WP:LIBEL. People often believe that they have succeeded in squaring the circle or whatever and won't be told otherwise, but they are not acting in bad faith. The modern trend for self-publishing academic papers online offers endless possibilities for doing this sort of thing, and journalists need to be aware of this phenomenon. Some basic fact checking and asking an independent expert would have avoided most if not all of the mistakes that have been made by the mainstream media here. The CNN article is one of the better ones because it has avoided mistakes that other sources have made.--♦IanMacM♦ (talk to me) 19:30, 18 November 2015 (UTC)
Indeed. It's just a matter of following our sourcing guidelines - this probably falls in the "extraordinary claim" category, but it's always best practice to check multiple refs for multiple angles (extraordinary or not). Just look at the links above, we have already disproved the Irish Independent story (or cast such serious doubt on it that it will not be used) by looking to multiple reliable sources - so the guidelines are fine. (RS has never meant infallible.) And yes, I hope no one on Wikipedia says that this guy intends a hoax (if that claim is unsourced to RS), if they do -- it should be rev deleted per BLP. -- Alanscottwalker (talk) 19:39, 18 November 2015 (UTC)
In agreement with Alanscottwalker: the CNN and BBC seem to be 100% truth: the professor claims he solved it. They aren't judging or confirming as journalists, they went to speak with appropriate organizations that would be experts and learned that it will take time to confirm, so they still with this being a "claim". The only "sin" I could see here is that this is jumping the gun on the reporting, as it will take a few years to confirm (assuming the proof is not shot down sooner), but that's far from making either unreliable. --MASEM (t) 20:35, 18 November 2015 (UTC)
When you say "the proof", what do you mean? What proof? There is no proof. You are being far too generous to the reporters. Their sin is not quick trigger fingers, it is a lack of any pretense of fact-checking. That's why they should no longer be considered reliable for this sort of story. —David Eppstein (talk) 20:52, 18 November 2015 (UTC)
There is a document that anyone can see. It is the proof spoken of in these BBC and CNN articles. Whether that proof is valid and actually demonstrates the solution, I don't know, nor do I expect reputable journalists to know. Hence why they contacted the group that hands out the Millennium Prize awards and a conference to see if the proof was legit. That's fact checking on that the proof exists, but in no way validates the proof as the solution. --MASEM (t) 21:10, 18 November 2015 (UTC)
"There is a document that anyone can see." This is silly. Go look at the document. Sławomir
21:31, 18 November 2015 (UTC)
There is no such published proof, the BBC does not say there is, the Telegraph says there is no proof, yet, and CNN says publication has not occurred yet. Alanscottwalker (talk) 23:25, 18 November 2015 (UTC)
This is a claimed "proof", and it is "published" by someone claiming to be Enoch. Sławomir
23:50, 18 November 2015 (UTC)
Not RS, so not usable. "Someone claiming to be" and it makes some "claim", should tell you right there it cannot be useful to us. Alanscottwalker (talk) 00:06, 19 November 2015 (UTC)
I'm not arguing to use the source for Wikipedia though. I'm arguing that there are some other sources that are bad sources. See the difference? (I'd ask that if you really want to pursue the minutiae, that we can conduct it privately on my user talk.) I agree completely that it would be inappropriate to use this in an article, but it is a clear indication that something is fishy with the story run by CNN/BBC. It is not the only indication (e.g., you can find some blogs, including the aforementioned guy that collects wrong proofs). But, generally speaking, when WP:FRINGE things make headlines like this, you won't see actual WP:RS jumping at the opportunity to prove the headlines wrong. Mostly it just doesn't even make it on the radar, unless someone has a blog whose mission it is to do just that (e.g., Peter Woit). We even have a humorous essay that lampoons precisely this situation: WP:CHEESE. This is a very real and relevant problem for us, though, precisely for the reason that you've correctly identified. Even if a story doesn't even pass the most cursory sniff test, it might still pass the letter of WP:RS (even if it fails the ephemeral "common sense" test MASEM alluded to, like the Nobel teen). That is a real problem, and this case shows precisely how. Sławomir
01:20, 19 November 2015 (UTC)
It's clearly not a problem. No one has even been successful in getting this story into the article. Why? Because if you actually know how to read and use sources, and how to operate per this guideline, you would know that all the story is, is someone has made a currently unpublished, unproven claim. That they have made the unpublished, unproven claim is sourced to RS, but just because something is in RS does not mean it goes into articles, like this has not gone into the article. Your other weak case on the noble prize fluff, is just morning program fluff, so hardly any problem. -- Alanscottwalker (talk) 01:55, 19 November 2015 (UTC)
The article is edit-protected, and edits are swiftly reverted by experienced editors who understand both the mathematics, and the guidelines. That's "why" it hasn't gotten in. However, this is not an isolated incident of some editors attempting to misuse sources like that, which suggests that the guidelines should be clearer. The "morning program fluff", that is a problem. Indeed, reams of proverbial ink were spilt defending that source and other similar sources at Wikipedia:Articles for deletion/Jacob Barnett (2nd nomination) and the subsequent Wikipedia:Deletion review/Jacob Barnett. The sources were defended as absolutely reliable at Talk:Jacob Barnett. If you don't think it's a problem, you're wrong. There are many editors who think that sources like this are reliable, because they have the name "BBC" attached to them. You see very clearly that this is a mistake, but I assure you that not all editors do. The guideline needs to be made clearer on this point. As you aptly put "This is not at all a problem for anyone who understands sourceing, which apparently you do not". Surely the task of a guideline is to make me (or other editors) understand sourcing. Where does it say that "morning fluff" is not allowed? Sławomir
02:38, 19 November 2015 (UTC)
Jacob Barnett is apparently some prodigy or "whiz kid" that has been noted by multiple sources. So what? That's not a problem of reliable sourcing. It's therefore not a problem for this guideline. (You may wish to take up your concerns over at WP:GNG). Alanscottwalker (talk) 02:48, 19 November 2015 (UTC)
No, no. I'm curious about the sourcing. And I can only share my experiences with other editors regarding sources like the "morning program fluff", which you feel is not a problem. But very recently, many kilobytes of debate and hours of editor time were spent at Talk:Jacob Barnett discussing this very point. And it keeps coming up again and again. It's a long story. But the short version is: yes it is a problem. So I'll ask again earnestly, where does it say BBC Breakfast (aka "morning fluff") is not a reliable source? Many editors believe that, because it has "BBC" in it, it is reliable period. Even after pointing at WP:REDFLAG and WP:NEWSORG, these beliefs persist. And these aren't necessarily new editors either. What's the nuance that they're missing, that you seem to see very clearly? Sławomir
03:05, 19 November 2015 (UTC)
What you are missing is that just because something is considered RS does not mean you use it for any particular sentence in an article. RS is not a state of being, it is only determined in the context. Alanscottwalker (talk) 03:24, 19 November 2015 (UTC)
True enough. But saying "context matters" does not answer the question either. The fact is these sources have been a problem. You have at various points agreed that sources like BBC Breakfast are fluff. But nowhere in the guideline is it really spelled out when such sources are acceptable and when they aren't. So, yes, a source can be a reliable source for one thing and not another, and so context does matter. But where in the guideline is it actually spelled out that the CNN source is not a reliable source establishing weight of Enoch's solution for our article? Or that the BBC Breakfast is not a reliable source on Barnett's future Nobel prospects? Maybe there's no easy answer to questions like those. Mostly, we get misconceptions of the kind that "context doesn't matter". That, because it's written by the BBC or CNN or whomever, it is 100% reliable (see MASEM's post below). Sławomir
08:04, 19 November 2015 (UTC)
This guideline gives guidance on what is by consensus considered RS in general, it does not determine the issue of whether something is WP:DUE, nor whether a sentence construction is WP:OR, nor whether something offends WP:BLP, nor meets WP:GNG, nor issues covered by other policies and guidelines. It cannot say that the sentence "Person claims to have solved this problem", belongs anywhere, in particular, in the encyclopedia. That sentence on its face is not science - it is either reasonably certain that they have made that claim, or there is not sufficient evidence that they have made that claim. Similarly, the sentence "BBC morning program calls person a future prize winner" is not science - either the BBC program has or they have not. Alanscottwalker (talk) 10:51, 19 November 2015 (UTC)
So you're saying that news items are reliable as primary sources, then, which I agree with, not as secondary sources. Perhaps the guideline should be adjusted to reflect that. This is already what MEDRS says about news concerning medicine. Sławomir
11:24, 19 November 2015 (UTC)
Have you read the first paragraph of WP:NEWSORG? As for whether some source is primary or secondary, it depends on text and context, see generally WP:NOR. It's not a neat dichotomy. CNN, for example, used multiple primary sources, and that CNN article is generally a secondary source for the history of the award. Alanscottwalker (talk) 12:09, 19 November 2015 (UTC)
──────────────────────────────────────────────────────────────────────────────────────────────────── Yes, I read the first paragraph of WP:NEWSORG. In the past, there has been confusion over what "fact" means there, probably because people don't really understand the difference between primary and secondary sources. A source (like the BBC or CNN, recently discussed), while reliable as a primary source of the fact that Enoch claims to have a solution, is not a reliable secondary source on which to base encyclopedic content. It seems like this is a grave omission from the guideline, since in assessing the worthiness of a news source, this is perhaps one of the most important considerations. Yes, we cover that in detail in policy, but there is no rule against linking to policies when those can helpfully clarify a guideline. Sławomir
12:16, 19 November 2015 (UTC)
In agreement with Alanscottwalker: the CNN and BBC seem to be 100% truth: the professor claims he solved it. They aren't judging or confirming as journalists, they went to speak with appropriate organizations that would be experts and learned that it will take time to confirm, so they still with this being a "claim". The only "sin" I could see here is that this is jumping the gun on the reporting, as it will take a few years to confirm (assuming the proof is not shot down sooner), but that's far from making either unreliable. --MASEM (t) 20:35, 18 November 2015 (UTC)
Bullshit. The BBC states, flat out, that Enoch has "been shown to have solved [the Riemann hypothesis]". It gives the misleading impression that it's already been decided that Enoch will receive the million dollars. So half of the interview is about when and how he'll get the money. The CNN piece does not interview any experts on the Riemann hypothesis. Instead, they cite one of the conference organizer: 'Nina Ringo, a member of the conference committee, said in an email statement: "I consider his results to be very important and confirm his discovery."' That's the only "expert" commentary they have. Then there is a bunch of text about when and if Enoch will collect the money. It's fine if they don't want to confirm or deny it as journalists, but if they want to be journalists, the expectation is that they will do journalism. Obviously, a journalism is not necessarily knowledgeable about the Riemann hypothesis. So, presumably it's a best practice to do research and interview people who do know something. That's where the reliability of science reporting comes from. It's called journalism. If someone at the BBC or CNN doesn't want to do research on a topic, that's fine too. But they don't get to be considered reliable sources. Sławomir
21:31, 18 November 2015 (UTC)
So even if the BBC didn't contact the Millennium Prize committee, CNN did and confirmed that there's no confirmation the proof has been shown valid. There's nothing broken here about RS relating to science, its basically being smart that on controversial statements (in this case, if the proof has been validated which has different answers depending on source), we avoid stating anything as fact. I will note this has little to do with any actual science or math understanding of the hypothesis, but a lack of understanding of how the Millenium Prize system works that is rearing its head. --MASEM (t) 21:42, 18 November 2015 (UTC)
Ok, let's set aside the Millennium Prize. CNN confirmed the validity of Enoch's proof by asking Nina Ringo. Do we accept that her opinion is made with the appropriate independence and scientific authority? Should CNN have bothered asking an expert specifically on the Riemann hypothesis, like Peter Sarnak, Pierre Deligne, or Michel Lapidus? Or is this misleading by omission? Sławomir
22:00, 18 November 2015 (UTC)
Nowhere in CNN's article, in the words of author Thomas Page, did they validate that the proof was right. Page spoke to Ringo, Ringo's response is quoted. Page spoke to CMI, CMI's reply is quoted, including the process that the proof needs to go through before being validated. Page makes no attempt to affirm the proof's validity. That's good reporting, and nothing like the Independent that jumps on Ringo's statement as conclusive. --MASEM (t) 22:05, 18 November 2015 (UTC)
They interviewed an "expert" who confirmed the validity. But they didn't bother to ask a real expert on the Riemann hypothesis. It's lying by omission. Sławomir
22:13, 18 November 2015 (UTC)
Absolutely not. They quotes a person who I have no idea of their expertise on the hypothesis who claims it was solved. That is not CNN saying it was solved. They are not lying. Further, they turned to the group that administers the prize and were told that just the act of publication was not sufficient for saying it was solved; that's trying to get the whole story and a sign of good journalism. Contrast that to the Independent story that took Ringo's statement as word of god, which is lying and/or poor reporting. --MASEM (t) 22:41, 18 November 2015 (UTC)
I'm sorry MASEM, but the hypothesis that the BBC and/or CNN piece is "good journalism" is just a non-starter, and I don't think even you believe that. Good journalism should at least try to get the story right. In this case, CNN is presenting their source (Ringo) as an authority. There are hundreds of authorities that CNN could have consulted with, many of whom would be very willing to supply their commentary. But the best expert CNN came up with was this one.
"They quotes a person who I have no idea of their expertise" - seriously? If Ringo is not an authority on the Riemann hypothesis, then why is she being interviewed? Why is there a story here at all? There are people that think the CIA sends secret messages to their fillings. But we don't see viral news stories saying "Mr. Quackenbush of Hackensack, New Jersey, claims to have discovered secret radio transmissions from a CIA satellite to his dental fillings. According to Dr. Pembroke, at the University of New Hampshire, 'The CIA definitely sends messages to fillings.' The CIA issued a prepared response that 'Thank you for your interest in the CIA. Per executive order, we cannot confirm or deny any clandestine programs at this time.'" Just because you wrap nonsensical bad journalism in the phrases "claim" and "according to so-and-so" does not magically make it good journalism. Sławomir
23:31, 18 November 2015 (UTC) also uses wording like "claim" and "according to so-and-so" to place contentious statements as claims (see WP:YESPOV). That is good practice to avoid asserting a statement that one cannot otherwise verify as the truth. CNN went the extra mile to provide a counter-point to Ringo's statement in that the proof is not considered conclusive until certain steps happen. So a reader from CNN's article will walk away known that a proof has been presented but the world doesn't know yet if it passes muster, only that one conference head thinks its looks legit. There is nothing wrong with that approach. If BBC and CNN were reporting like the Independent, your concerns are fully valid, but they simply don't exist here. --MASEM (t) 23:54, 18 November 2015 (UTC)
I'm astounded at your apparent contention that using words like "claim" and "according to so-and-so" is a necessary and sufficient condition for good practice, in journalism or encyclopedia writing. One cannot write a sentence without a noun. But not all sentences noun with grammatical are. Sławomir
If we are talking about a situation where the "right" answer is not known - whether it be if the mathematical proof here is legit or a more controversial social issue - using wording like "John Q Smith claims X" or "According to John Q Smith, X" for WP articles is perfectly good writing as to avoid stating a potentially incorrect statement as fact, and thus if it is shown wrong, no damage is done to Wikipedia; we are just reporting it like it is. CNN did it here too, a good sign of an RS. --MASEM (t) 15:42, 19 November 2015 (UTC)
Ok, thanks. I think I can more clearly see the divide here. You're thinking of news items as reliable primary sources. In that case, I would agree, CNN is a reliable source that so-and-so has such-and-such view. Whether so-and-so's views on a subject should be represented on Wikipedia is an issue of policy (WP:NPOV, WP:NOR). Do we agree on this? I disagree that using this terminology is a sufficient condition for using the source as a secondary source. For example, although the CNN quotes Ringo, Wikipedia should treat Ringo's view as (essentially) the view of a random person on the street, and thus carries practically no weight. The mere fact that CNN interviewed an individual doesn't implicitly stamp that person with more authority and weight than if they had (for instance) just blogged about it.
Now, a news source can be a secondary source, if many eyewitnesses were interviewed and the reporter comes to conclusions in the editorial voice of the newspaper. A lot of wire news has this character. But we've run into problems in the past with sources that tend to blur the lines more (a good example is The Independent) and present facts without sources. It's tempting to use those as secondary sources. How does this guideline help us determine that this is not acceptable? Is there a way to make it clear why one class of sources or one way of using sources is acceptable and the other isn't? Sławomir
16:18, 19 November 2015 (UTC)
I think that's what I'm thinking. CNN's article, specifically, is clearly reliable as a source (I'd say it's secondary, but it doesn't matter on this point) for the fact that Enoch has published a proposed proof for the hypothesis, but CNN's article, and all the others provided as links, is absolutely not a reliable source to assert the proof is valid, even with the quote from Ringo, who I wouldn't necessary say is just a random person (they represent this conference, so they have to have some scientific knowledge), but not necessary an expert to judge fully if the proof is valid. We ( know that nearly all scientific advancement goes through the peer review process, this will too before plus more before the prize is award, and only a few articles (like the CNN) get this fact correct. If we assume that every news source said the proof was valid based on Ringo's statement, even though we know CMI hadn't made its judgement yet, we have the ability to correct that. And to that end, yes, we do have to be careful when newspapers make scientific conclusions without lack of established evidence. The way I would take this without having to modify any policy or guideline is to follow what's at WP:YESPOV - if a contentious statement has been put forth in sources, primary or secondary, (here "The proof is valid" asserted by Enoch and Ringo, questioned by CMI), we have the ability to write it as a claim made by a person or work, in this case, Enoch's offering and Ringo's statement. We do no harm as we're not asserting something true that hasn't been proven out, and maintain WP:V as we are still reporting what the sources say. In the larger generalized case of a mainstream newspaper making a novel statement of scientific fact without presenting or indicating the origin of that fact, we know nearly all such facts are only accepts after peer-review, and the lack of such would make that statement contentious. (Most good RS newspapers will always link to the published paper or indicate the upcoming paper when a new scientific theory is presented). --MASEM (t) 16:31, 19 November 2015 (UTC)
But this view seems based on the presumption that mention of the claimed proof is suitable for inclusion in Wikipedia. Rather it is pretty clearly not suitable. Lot's of people claim to have proven the Riemann hypothesis, and we don't cover all of them. How do you convince me of that? What guideline or policy (beyond the one under discussion) rules it out as suitable? WP:NOR? WP:WEIGHT? WP:REDFLAG? Sławomir
17:02, 19 November 2015 (UTC)
A combination of WP:WEIGHT/WP:UNDUE , WP:CONTEXTMATTERS, and WP:SCIRS/WP:SCICON. If there are lots of proposed proofs that these problems get and this one only just happened to get an amount of media coverage, to cover it and not the others is clearly undue. To also add WP:NOT#NEWS and WP:RECENTISM - just because a large number of mainstream sources presented this, doesn't mean we have to include it until we understand it better. --MASEM (t) 17:08, 19 November 2015 (UTC)

I've reverted a BOLD good-faith edit by Slawekb to create a presumption, if not a near-prohibition, against general news sources being used for scientific "matters" based upon this discussion. That edit goes far beyond anything which has been discussed here and, with a principle as broad as that edit would create, really needs specific, focused discussion here (and probably a RFC published at VPP) before being adopted here. Regards, TransporterMan (TALK) 20:48, 18 November 2015 (UTC)

I didn't really think it would be all that controversial to say that we should demand science news sources to have a reputation for fact checking and accuracy. That's already part of RS and NEWSORG. All I've done is add "science" to it. It's clear from this episode that some news sources that may have a good reputation for fact checking in some areas (e.g., world news) may run columns that do not have as solid a reputation (e.g., science news). We already advise caution in citing opinion content. Sławomir
21:20, 18 November 2015 (UTC)
We already demand that news sources have a reputation for fact checking and accuracy, as you say, so if that was your only point, then the addition would have been objectionable as rule creep. But that wasn't all that it said. It said, "News sources are not generally reliable for scientific matters, unless the column has a reputation for scientific fact-checking." (Emphasis added.) That's quite different and far more restrictive than having a general reputation for fact checking and accuracy (especially, but not only, when the breadth of the word "matters" is taken into consideration). It seems to me that the second bullet point of NEWSORG already deals with this sufficiently. Regards, TransporterMan (TALK) 21:37, 18 November 2015 (UTC)
If it were just this one issue, they rulecreep would apply. But there is a pattern of disregard for scientific fact-checking in otherwise respectable media outlets like the BBC. For example, here the BBC puts a Nobel Prize in a teenager's future, as does this article by Medical Daily ("now a top candidate spot for the Nobel Prize"). This is a problem with viral news stories that appear to get picked up by one source, and then reprinted everywhere else. Facts appear to take a hiatus in a lot of science reporting. So I think special emphasis is called for. Just because it's the "BBC", and isn't "opinion content", doesn't automatically mean that it is subject to the same degree of scrutiny as world or political news. Context really matters, and I think (I hope) we all agree about that. The danger is that viral news stories also tend to bring swarms of editors, and we shouldn't rewrite articles based on these stories. So, it's good to have a bullet that specifically addresses this issue. (Anyone remember Wikipedia:Articles for deletion/Vinay Deolalikar?) Sławomir
21:53, 18 November 2015 (UTC)
But here again, it's less about deep understanding about science or math, and more about what I'd consider as human interest reporting; they are adding a spin that brings humanity and personality to a story but that would have no place in a proper discussion on the scientific merits, if only to draw readership. It's not bad reporting in terms of the science, but exaggerating claims to make the story seem more important as a human interest element; eg the viral news stories. The Nobel Prize links are a good example, and while I don't know how many people attempt proofs of these Millenium Prize problems, I suspect this story got interest because the professor is from a rather impoverished area of the world. We need to clearly recognizes puffed-up claims that are associated with these science stories but I don't think this is necessary an issue with how RSes report on scientific details. It doesn't just affect articles in science, but any type of modern journalism where works are fighting for pageviews. --MASEM (t) 22:00, 18 November 2015 (UTC)
Agreed. But surely such human interest stories should almost never be used in an encyclopedia. Perhaps we at least could agree on that? Sławomir
22:13, 18 November 2015 (UTC)
I don't think we can ignore these as it becomes more prevalent in sources, but we can be using enough common sense to know when stories are trying to tickle the human element and discount those claims (like the Nobel prize teenager above) while keeping other more factual claims that are not so puffed-up. Its unfortunately how WP has to deal with new media which I've seen rear up across the board over the last few years years. --MASEM (t) 22:44, 18 November 2015 (UTC)

Newspapers can not afford to pay fact checkers any more - at least that is the usual excuse given for the reliance on "press releases" even by the most recognized journals now. Any suggestions how to deal with this rapidly increasing unreliability of "reliable sources"? Collect (talk) 23:05, 18 November 2015 (UTC)

As a small followup, I have seen evidence that the BBC did change their initial story, which ran as stating , without question, Enoch proved it and won the prize. The article had been changed since. --MASEM (t) 05:57, 23 November 2015 (UTC)


Details of this particular case aside, I'd like to summarize three basic points that I hope we can more or less agree on.

1. Journalistic standards have eroded with the rise of new media. One can no longer regard the "BBC" as a reliable source for all news. Context matters. For example, BBC Science might be more reliable for scientific content than BBC Breakfast. Collect has pointed out that "Newspapers can not afford to pay fact checkers any more", and that research shows a growing reliance on press releases. (This seems to be true across the board. Even BBC Science relies increasingly on press releases, per Collect's observation.)

2. I think there is (probably) agreement that there are viral human interest "news" stories, often relating to scientific matters, that tend to get picked up by a lot of sources. These sources may not be the most reliable sources on scholarly matters. However, these are also the same sources that are likely to attract a lot of prospective editors to articles. Many such editors might be new editors, or editors who typically edit in some other content area. (See the recent history of Riemann hypothesis, Talk:Jacob Barnett, or the inordinately long Wikipedia:Articles for deletion/Vinay Deolalikar). Probably there are other examples, but I am not very active as an editor these days.

3. Wikipedia editors aren't expected to be experts in the subject matter. Although User:Masem advises "common sense", Talk:Riemann hypothesis has four proposals over the last few days to add these human interest pieces as stories, and a number of editors have even attempted to add those sources to the article Riemann hypothesis. It's clear that "common sense", while a useful guiding principle for those with some experience in the subject matter, may not be so readily available to all editors. Another example is Talk:Jacob Barnett and the archives therein, which is full of claims that the sheer number of sources must be reliable. Common sense (apparently) be damned! (And invocations of "common sense" are dismissed by some parties as "original research".) Then there is the aforementioned case of Wikipedia:Articles for deletion/Vinay Deolalikar, which was a contentious battle.

We already do have a guideline in place for just such a situation, namely WP:MEDRS:

  • The popular press is generally not a reliable source for scientific and medical information in articles.


  • Newspapers and magazines may also publish articles about scientific results before those results have been published in a peer-reviewed journal or reproduced by other experimenters. Such articles may be based uncritically on a press release, which can be a biased source


  • For Wikipedia's purposes, articles in the popular press are generally considered independent, primary sources.
  • A news article should therefore not be used as a sole source for a medical fact or figure. Editors are encouraged to seek out the scholarly research behind the news story. One possibility is to cite a higher-quality source along with a more-accessible popular source, for example, with the |laysummary= parameter of {{cite journal}}.

The last item seems to be equally true with "medical" replaced by "scientific", in light of recent examples.

The wording in NEWSORG is horribly unclear: "For information about academic topics, scholarly sources and high-quality non-scholarly sources are generally better than news reports.... News reports may be acceptable depending on the context." What is the acceptable context? At least for scientific topics, MEDRS already gives the context, namely if that content can be independently supported by reliable scholarly research. I think this is a reasonable standard for most scientific topics, and NEWSORG should be adjusted to reflect this standard. Sławomir
01:59, 19 November 2015 (UTC)

Before you assert that "At least for scientific topics, MEDRS already gives the context", you might want to read the very first paragraph of MEDRS, which defines what MEDRS covers. Hint, It is WP:MEDRS, not WP:SCIRS. SCIRS is for scientific claims, MEDRS is for biomedical claims. --Guy Macon (talk) 03:44, 19 November 2015 (UTC)
Thanks, I wasn't aware of that guideline. Is there a reason it isn't linked more prominently in this one? Sławomir
04:04, 19 November 2015 (UTC)
Given your silence on this point, I'm assuming you agree that WP:SCIRS should be linked somewhere from this guideline. Sławomir
11:15, 20 November 2015 (UTC)
Note: I mentioned the WP:MEDRS dispute at that guideline's talk page. Flyer22 Reborn (talk) 02:03, 19 November 2015 (UTC)
Wikipedia:Administrators' noticeboard/Incidents#User:Sławomir Biały keeps changing Wikipedia:Identifying reliable sources without consensus --Guy Macon (talk) 02:28, 19 November 2015 (UTC)
  • Context is important. You cannot simply lift a statement like "The popular press is generally not a reliable source for scientific and medical information in articles" from WP:MEDRS, where it is in a paragraph that contains phrases such as "most medical news articles", "news articles too often convey wrong or misleading information about health care." and "they tend to overemphasize the certainty of any result, for instance, presenting a new and experimental treatment as 'the cure' for a disease or an every-day substance as 'the cause' of a disease" and plop it down in WP:RS, where it will likely be interpreted as applying to any scientific information. There is no consensus for doing that. Again, do not edit the page without first obtaining affirmative confirmation that there is a consensus for your changes. --Guy Macon (talk) 03:35, 19 November 2015 (UTC)

News items as primary sources[edit]

Per Alanscottwalker's analysis, I think the issue hinges on whether news items constitute reliable primary or secondary sources. WP:PRIMARYNEWS already indicates that news is typically a primary source. This is especially true of news concerning scientific and medical subjects (the principle is already discussed at WP:MEDRS). I think the relevant bullet point of the guideline should be clarified on this point. Something like the following:

  • For Wikipedia's purposes, articles in the popular press are generally considered independent, primary sources. A news article should therefore not be used as a sole source for a scientific fact or figure. Editors are encouraged to seek out the scholarly research behind the news story. One possibility is to cite a higher-quality source along with a more-accessible popular source, for example, with the |laysummary= parameter of {{cite journal}}.

Regardless of the specific form, I think it is essential that we link to WP:PRIMARYNEWS, since this is directly relevant to assessing the source. Sławomir
11:44, 19 November 2015 (UTC)

NOTNEWS is an essay, not a guideline or policy, and as such is merely the opinion of those who have contributed to it, not the community as a whole. And even if it wasn't, it doesn't say what you say it does, at least as not as broadly as you say it in your proposed text. Frankly, even when properly analyzed I think that many would disagree with the black-and-white way that it analyzes news sources. Regards, TransporterMan (TALK) 15:44, 19 November 2015 (UTC)
Ok, fair enough. I'm guessing that you also take issue with how this essay is referenced at WP:MEDRS. But anyway, is a link to WP:PSTS appropriate? As discussed above, it looks like often there is a real issue in distinguishing between whether a given news source is a primary or secondary source. It looks to me like others are saying that human interest stories are primary sources, that so-and-so said such-and-such, but are seldom reliable as secondary sources. How do you propose we incorporate that into the guideline? Sławomir
15:50, 19 November 2015 (UTC)
The NEWSORG guideline as currently written appears to create a false dichotomy between opinion content, which is reliable as a primary source, and news items which are reliable as "factual" sources (which the reader is to infer is a "secondary" source). As discussion here shows, there is a large grey area. News sources may be primary or secondary sources, depending on the context. For example, the CNN piece can be used as a primary source for Enoch and Ringo's opinions, or as a secondary source on the policies of the Clay Mathematics Institute. Do I understand correctly? If so, a sentence should be added to the second paragraph of the newsorg guideline clarifying that news items can be primary or secondary sources, depending on context. A link to WP:PSTS should be provided, and possibly also a link to WP:CONTEXTMATTERS. Sławomir
12:48, 20 November 2015 (UTC)
An initial proposal, that at least conveys there is an error in the dichotomy currently presented: "News sources may be primary or secondary sources, depending on how they are used." To be inserted at the beginning of the second paragraph of NEWSORG. Sławomir
13:50, 20 November 2015 (UTC)
A refined proposal. In the paragraph on academic topics in WP:NEWSORG, replace "News reports may be acceptable depending on the context." with the much clearer "News articles are not usually written by a specialist in the subject. So while news is often reliable as a primary source in the sciences, it should be only be used as a secondary source on scientific and academic topics with caution." Sławomir
01:45, 26 November 2015 (UTC)

Press releases[edit]

I'm a science journalist. I read original medical journal articles, and sometimes read press releases after I'm finished as a final check. All the press releases I see based on a peer-reviewed journal article are issued simultaneously by the journal and the author's institution. They're reviewed by the author, the journal and the institution's public relations office, who are all responsible for the release's accuracy or lack of it. This process itself has been the subject of many peer-reviewed journal articles, eg doi:10.1001/jamainternmed.2015.2147 and I recommend you read the substantial peer-reviewed literature before discussing it.

The news media vary greatly in their reliability, quality, accuracy, and methods.

At the top end, many peer-reviewed journals published by professional societies have news sections, such as the Science magazine News & Comments, or the BMJ news stories, which in my experience are as reliable (and sometimes more reliable than) the peer-reviewed journals. I believe most of Science's reporters have PhDs. For example, Science hired Gina Kolata, who has a PhD in mathematics, to cover mathematics, but she moved on to the New York Times. (Interestingly, she said it turned out to be particularly difficult to explain mathematics to non-specialists.) The reporters in Science know or can find out who the top experts are on any story, and they have editors who review their stories. Other publications, such as the New Yorker, have extensive fact-checking.

So if you're going to argue that news organizations are unreliable because they uncritically reprint inaccurate press releases and abstracts from peer-reviewed journals, you should also argue that peer-reviewed journals are unreliable because they print those inaccurate press releases and abstracts in the first place.

At the bottom end, some web sites simply scrape press releases. There are all different gradations in between. And in a given organization, there are gradations in journalists. There's a difference between the science editor and the fashion editor at the BBC.

Ivan Oransky, the former editor of Reuters Health and now editor of Medpage Today and journalism professor at NYU, famously said that for a reporter to write a story from a press release and/or abstract, without reading the original article, was "journalistic malpractice." (Like a lot of journalists, Ivan Oransky also has an MD.)

I had a job in which the editor told us that we should just write news stories from the abstracts, and not bother reading the original article. We refused.

This should be distinguished from press releases based on a meeting presentation of research that hasn't yet been published in a peer-reviewed journal. These releases can also be reported either verbatim by bottom-scraping web sites and patient forums or by journalists who often have been covering the subject for years and understand the context, singnificance and tentativeness of a meeting report. For that matter, they can write a story based on the abstracts in the abstract book (which is often, technically, a special section of a peer-reviewed journal and arguably a WP:MEDRS). Reuters, Medpage Today, and similar publications do have a level of fact-checking for these stories. (Here's a quick way to identify a reliable meeting article: Do they get a comment on the meeting report from an independent expert?) They also have editors in the home office, some with MDs or PhDs, who have been following these issues for years, and can identify a claim that is out of the ordinary.

So my bottom-line message is: You like high-quality peer-reviewed journals? So do I. There's lots of research in high-quality peer-reviewed journals on the process of writing science news. We should depend on published research rather than our own WP:OR observations and personal opinions of the newspapers and TV. --Nbauman (talk) 17:11, 19 November 2015 (UTC)

Thanks for this view. It's indeed a conundrum, that the current guideline doesn't really go far enough to distinguish between "good" science reporting, and "bad" science reporting. I had rather ham-handedly attempted such an edit. Naturally, that edit was reverted, but there's clearly a huge spectrum between "science journalism" on the one end, and churnalism on the other. Currently, the guideline only distinguishes between news content, and opinion content. Should the guideline be made more clear that there is an enormous difference between responsible scientific journalism, and the kind of churnalism detailed in the previous sections? Sławomir
17:51, 19 November 2015 (UTC)
In medical news, the best way I have to evaluate news sources and stories is the checklists used at, whose mission is to evaluate and improve medical journalism. They don't have a list per se; they just tell you how they evaluate news. It does require some judgment to apply, but it's well-designed like any other medical checklist. --Nbauman (talk) 19:42, 19 November 2015 (UTC)
For medical news, there's already WP:MEDRS, which you probably feel is too restrictive. Scientific news more broadly is covered at WP:SCIRS. That probably should at least be linked from this guideline somewhere. Sławomir
21:11, 19 November 2015 (UTC)

A related question[edit]

How does one tell if a "news item" comes from a reliable news service? For example, the BBC interview mentioned previously is part of Newsday (radio series), not (as far as I can tell) BBC News, which is mentioned prominently as a reliable source of "news". So, I guess my question is, does NEWSORG intend news agencies, or does it cover other types of media such as the program just mentioned? If not, what part of RS does? Sławomir
11:45, 20 November 2015 (UTC)⇒

Well, as an interview it is generally a primary source, but can it be considered as an RS, sure, and that it is carried by the BBC confirms that the BBC finds it reasonably reliable too - ie., it is not a mock interview. Alanscottwalker (talk) 11:52, 20 November 2015 (UTC)
But it's not a news source under NEWSORG, because it's not published by BBC News or any other news agency. Do I understand correctly? I see I may be misreading the guideline. If that's the case, how does one decide what is news, if "published by a news agency" is not the correct criterion? Given the amount of churnalism masquerading as news these days, I think it's essential to get clarity on that point. Sławomir
12:49, 20 November 2015 (UTC)
You do misunderstand: the guideline applies to news outlets. Alanscottwalker (talk) 13:41, 20 November 2015 (UTC)
What does that mean? How do we determine what is "news" and what is not "news"? Do human interest stories, churnalism, and other pseudonews count as "news"? Sławomir
13:42, 20 November 2015 (UTC)
Common sense. It's always difficult to do anything without it. Alanscottwalker (talk) 13:45, 20 November 2015 (UTC)
Ok, but I find that what I consider news may differ from what others do. For example, at Talk:Jacob Barnett, the program BBC Breakfast was presented as "news" (in fact, as part of BBC News). But I wouldn't consider this a news source. Is it news? How do we decide? Sławomir
13:49, 20 November 2015 (UTC)
Discuss. Guidelines are not meant to prevent talking to other Users. If you think it needed and discussion did not do it, see, WP:DR. Alanscottwalker (talk) 13:51, 20 November 2015 (UTC)
I'm fine with that. But is it clear that there is scope in the guideline for disagreement about whether "BBC Breakfast" is "news"? The current dichotomy presented could be interpreted as indicating that either it's "news" or "opinion". Can it be neither? Sławomir
14:31, 20 November 2015 (UTC)
Neither? It does not really matter. It can be news or not and that won't make it 'not a reliable source', in and of itself. If any particular usage is so perplexing, that is why we have WP:RSN. Alanscottwalker (talk) 18:16, 20 November 2015 (UTC)
It basically comes down to context: what article you want to use it in, what information wants to be included, etc. Take the piece above about the proof. I would certainly consider that a news source for Enoch if we had an article on him, but not for the hypothesis. --MASEM (t) 18:26, 20 November 2015 (UTC)

BBC and other sources[edit]

Alas, BBC (even BBC Science) does not fact check material any more - they run with press releases just like everyone else. [3] gives some of the bad news about the print media relying on press releases. It is just too expensive to have expert fact checkers around on staff in these lean days for the media. Collect (talk) 22:04, 18 November 2015 (UTC)

Is an otherwise-reliable source that uses unreliable methods still reliable?[edit]

For example, if a newspaper says "95% of the respondents to our online poll are in favour of ____", it's falling foul of the self-selection bias. The results cannot be trusted in any way. Is the statement still reliable enough to include in an article? Similarly, if a newspaper says "Young children who sleep with the light on are much more likely to develop myopia in later life, therefore, sleeping with the light on causes myopia", which breaks the rule that correlation does not imply causation, is the statement still reliable enough to be included in an article? Banedon (talk) 01:02, 19 November 2015 (UTC)

China Daily - "On China's portal websites like sina, 85 percent of those polled showed support for the couple." Banedon (talk) 00:39, 20 November 2015 (UTC)
I believe that each specific is context specific: "The appropriateness of any source depends on the context." I don't think it's usually a question of a source being either 100% reliable or unreliable. I think your example potentially could be included in an article, but it would need to be put into context. You would specifically need to identify that the number came from an online poll on China Daily's website, and link to the article reporting that number, not the poll itself. But I imagine this would need to be determined on a case by case basis. FuriouslySerene (talk) 14:21, 20 November 2015 (UTC)
A problem with that is that many dubious statements suddenly become "reliable". For example, statements such as "I tried this diet for two months and I lost weight!" would be perfectly acceptable as long as it is stated as such, and a page could potentially contain a sentence such as "100% of the respondents to an online poll on this diet's efficacy said they lost weight within two months". Such a statement may be true, but its impact would range from meaningless (for those readers who are aware that studies with self-selected samples are pointless) to dangerous (for those who aren't). I'm hesitant about allowing it at all. Banedon (talk) 05:49, 23 November 2015 (UTC)
A key step here is to make sure the chain of sourcing is explained in the article. So taking initial example, if this is reported by a normal RS, it would probably be best to say "An online poll taken by X show that most people think Y is true", instead of just saying "Y is true". We're not changing what the RS says, but we're not simply waving awy the use a questionable data gathering method. But as noted it is a case by case basis. --MASEM (t) 05:55, 23 November 2015 (UTC)
Again a problem with that is that it gives the online poll legitimacy. The online poll can be accurate, but it can also be completely wrong. A couple of articles on this from cursory Googling: [4] and, more sympathetic, [5]. The first source cites a great example of why online polls (and similar polls with self-selected samples) aren't reliable: it vastly overestimated the percentage of people who thought Bill Clinton should resign over his relationship with Monica Lewinsky. But from what you said, it could be used to legitimize the statement "According to an online poll, a majority of Americans think Bill Clinton should resign over his relationship with Monica Lewinsky". Like I said, I'm hesitant about allowing it at all, just like fringe Nasa-never-landed-on-the-moon theories don't show up on the moon landing page. Banedon (talk) 06:08, 23 November 2015 (UTC)