Archive 35Archive 38Archive 39Archive 40Archive 41Archive 42Archive 45

An issue regarding documents available under the Freedom of Information Act (United States)

Can I draw attention to Wikipedia:Reliable sources/Noticeboard#A U.S. National Gang Intelligence Center document on 'Juggalos'. The discussion revolves around whether a document which may only be available via a FOIA request can be considered to meet WP:RS policy. It might benefit from further input - and may indeed require clarification within WP:RS policy itself. AndyTheGrump (talk) 20:53, 31 January 2013 (UTC)

Given the claims that this kind of document is being used in many articles, we may need to clarify this. WhatamIdoing (talk) 02:00, 2 February 2013 (UTC)

Discussion on Meta regarding WebCite service

Please come participate in this discussion: meta:WebCite. ···日本穣? · 投稿 · Talk to Nihonjoe · Join WP Japan! 19:01, 9 February 2013 (UTC)

Conspiracy theories

There probably needs to be some sort of exception for (for want of a better expression) conspiracy theories. You don't have to believe that Bush43 personally oversaw the 9/11 attacks to report that such beliefs are held by some and have some traction. And then reference some popular CT websites. That is not using the site as a RS for the theory itself, but for the fact that these twerps are out there. It seems like a reasonable distinction.Cross Reference (talk) 19:03, 9 February 2013 (UTC)

It's not usually necessary for the 'big' conspiracy theories like that, because scholarly sources and serious books cover the major ones. Try searching just for books or scholarly articles, and see what you get. WhatamIdoing (talk) 21:41, 9 February 2013 (UTC)
It is acceptable to support an attributed statement of someone's beliefs to the people who hold the beliefs. If you state: "According to some conspiracy theorists, the US government is suppressing proof that eating broccoli causes cancer", it would be appropriate to cite two or three conspiracy theory websites to support this statement.
However, that isn't the end of the discussion. We also have to ask whether the statement about the opinion of conspiracy theorists should be included in the first place (See: WP:UNDUE.) That depends on what the specific article is about. In a general article on Broccoli, I would say mentioning the rantings of conspiracy theorists would be UNDUE weight. In an article focused on Governmental censorship of health information (or some such) it might be more appropriate to include conspiracy theories. Blueboar (talk) 22:08, 9 February 2013 (UTC)
I agree with the general statements of Blueboar. It probably requires a judgment as to whether the CT has some importance beyond its relevance to some (in my view) disconnected individuals. So the lunatic fringe who believe the govt is forcing you to eat killer broccoli are mostly harmless idiots (IMHO) while 911 CTers are anything but. So citing a website devoted to their ingenious and ingenuous theories is not making the site a reliable source about the facts of the events of 911 but it is a RS that these cretins (IMHO) believe what they do.Cross Reference (talk) 01:12, 10 February 2013 (UTC)
We certainly can use those sources, but the 9/11 CTers' ideas have been so widely studied that it's not necessary. WhatamIdoing (talk) 01:45, 10 February 2013 (UTC)
Very true... essentially, a CT source would be a primary source for the claims of the CT... and while these can be used (with limitations), we always prefer secondary sources (and must use secondary sources for analysis and conclusions... see WP:PSTS). So, if secondary sources exist, we should use them instead of the primary CT sources. Blueboar (talk) 02:19, 10 February 2013 (UTC)

Citations on Wikipedia and discussion at meta:WebCite

There is a discussion at meta:WebCite regarding citations on Wikipedia that would be of interest to those that watchlist this page. For those who don't know, webcitation.org is used to archive newspaper articles and other reliable sources that disappear from the original websites. Wikipedia currently has 182,368 links to this archive site. Regards. 64.40.54.47 (talk) 11:45, 11 February 2013 (UTC)

Raising the bar for peer-reviewed academic sources

I just reverted this edit made in January, because it completely changed the meaning of the paragraph, which has been long-standing for several years now. The change raised the bar for PhD dissertations, defining some as as reliable for citations and some not reliable, based on how often the dissertation has been cited elsewhere.

Because completed PhD dissertations have been reviewed and vetted by a panel of experts in the field, they aren't much different from peer-reviewed academic papers.

Looking in the archives, I don't see a consensus for considering dissertations as citable only "if it is shown that they have entered mainstream academic discourse". Nor do I see consensus for another assertion in this change, that even scholarly publications must be used "with caution" depending on its presence in the citation index and the reputation of the author. Therefore I have reverted back to the long-standing version. ~Amatulić (talk) 14:43, 19 February 2013 (UTC)

Gaming Google Scholar Citations and Google Scholar Metrics

An interesting article on gaming Google Scholar Citations and Google Scholar Metrics: http://www.roughtype.com/?p=2841 The good news is that we don't depend on either and instead look for peer-reviewed papers. The lesson here is that you pretty much can't use "number of hits on Google" to prove anything -- as indeed we don't. --Guy Macon (talk) 09:49, 21 February 2013 (UTC)

Dailymail

Why is Dailymail used as a reliable source on Wikipedia? What if I were to use this as a source for the Global warming article? Th4n3r (talk) 22:53, 28 February 2013 (UTC)

See the /FAQ. Even The Daily Mail is reliable for some purposes. WhatamIdoing (talk) 02:57, 8 March 2013 (UTC)
I've always argued that the Mail is sometimes reliable, but I'm less sure of that now. Certainly it is never reliable for scientific fact. Itsmejudith (talk) 20:11, 19 March 2013 (UTC)

Columns

News sources often contain both factual content and opinion content. "News reporting" from well-established news outlets is generally considered to be reliable for statements of fact (though even the most reputable reporting sometimes contains errors). News reporting from less-established outlets is generally considered less reliable for statements of fact. Editorial commentary, analysis and opinion pieces, whether written by the editors of the publication (editorials) or outside authors (op-eds) are reliable primary sources for statements attributed to that editor or author, but are rarely reliable for statements of fact.

Some editors seem to assume that if an "article" in a generally reliable source (such as the New York Times, Washington Post, etc.) is not marked "opinion" then it isn't opinion. In addition to editorials and op-eds, most publiations have "columns", which are the work of a columnist (or reprinted from other publications), and, as far as I can tell, are not generally fact-checked, as they are clearly intended to be seen as the opinion of the columnist.

It may be going the other way as well, but I see a number of left-wing columnists' opinions being inserted as fact in articles about right-wing subjects. I'd like some guidance as to whether "columns" appearing a reliable source should generally be considered "reliable". — Arthur Rubin (talk) 00:58, 8 March 2013 (UTC)

Are you thinking of syndicated columnists like Maureen Dowd? WhatamIdoing (talk) 03:03, 8 March 2013 (UTC)
Quite possibly. She's not the one I had in mind, but she is probably a more clear example. — Arthur Rubin (talk) 04:59, 8 March 2013 (UTC)
I don't know what level of fact-checking they do. There is obviously editorial control at the syndicate, and also at each subscribed periodical, since the local editor could choose to omit a given column. I wouldn't expect normal fact-checking to happen on a syndicated humor column, and I'm not sure that it would happen on lifestyle columns (e.g., interior decorating). But for a political opinion column, which is likely to attract more attention from people with lawyers than advice on home decorating, it's possible that at least a minimal amount of fact-checking happens, at least to verify quotations. I wonder how we could find out what happens for an average syndicated column. WhatamIdoing (talk) 06:25, 8 March 2013 (UTC)

Another Wikipedia page as a source

Is in a Wikipedia article a reference to another Wikipedia page considered to be a reliable source?--Gyte75 (talk) 08:55, 20 March 2013 (UTC)

Normally, other Wikipedia articles are not considered reliable sources since there is no editorial control over them. However, not everything need be linked - see WP:BLUE. For example, in my view is it sufficient to write "The White House, home of the United States President ..." relying on WP:BLUE to justify the absence of a citation, but having a Wikilink to provide more information (having first checked that the article "White House" has adequate citations. Martinvl (talk) 09:31, 20 March 2013 (UTC)

Google research and identifying reliable sources

Why? This guideline is primarily focused on content, not article titles. WhatamIdoing (talk) 22:17, 18 February 2013 (UTC)
I'm confused by why you think the style guide's advice against showing off by using obscure foreign phrases in the body of an article is at all relevant. WhatamIdoing (talk) 19:28, 19 March 2013 (UTC)
WP:SET lists a number of valuable sources, and explains techniques for searching them in order to get valid results. Some of the {{Google}} series of templates make it easy to search multiple reliable sources—Encyclopedias like Britannica, magazines like the Economist, newspapers like the New York Times, broadcast sites like the BBC—simultaneously. {{Find sources}} is another useful oldie. LittleBen (talk) 18:12, 4 April 2013 (UTC)
So? This guideline is about whether a specific source is sufficiently strong to support the weight of a given bit of material in an article. What does how to use a search engine to determine what proportion of sources use which terms have to do with determining whether a specific source is sufficiently strong to support the weight of a given claim in an article? WhatamIdoing (talk) 22:40, 7 April 2013 (UTC)

URLs on category pages

Should External links to reliable sources or other websites be permitted on WP:Categorization pages? There is a discussion at Wikipedia talk:Categorization#Reference resources about a template that links to several other websites. If you have an opinion, please comment there. WhatamIdoing (talk) 21:46, 28 March 2013 (UTC)

Request for comment

Your attention is called to the above-named section, Use Mapping L.A. as reliable source?, which could use your expert input. GeorgeLouis (talk) 16:01, 7 April 2013 (UTC)

A source can be good enough to source a citation while not good enough for a statement of fact. There are also materials that can not be used as sources but are still valuable to the article. We have no policy or guideline for this. RS has us pretend the material is a source. Such distorted logic is not necessary.84.106.26.81 (talk) 07:18, 9 April 2013 (UTC)

Rather like people doing college research who use, but cannot cite Wikipedia. I think a note in the reference regarding the use made of the source is the way to handle this. User:Fred Bauder Talk 11:44, 9 April 2013 (UTC)
Not citing the sources used is plagiarism. User:Fred Bauder Talk 11:52, 9 April 2013 (UTC)
See Template:Notelist for creating footnotes that are not full citations. User:Fred Bauder Talk 13:24, 9 April 2013 (UTC)
The usual thing to do with "materials that can not be used as sources but are still valuable" is to put them in a ==Further reading== section, following the guideline on that at WP:FURTHER. WhatamIdoing (talk) 00:10, 10 April 2013 (UTC)
Ah! There is the guideline WP:FURTHER and even Wikipedia:Further reading. Thanks. Lets see how that ends first.
84.106.26.81 (talk) 04:32, 10 April 2013 (UTC)

Visual material

Some editors feel that the prohibition on using self-published sources also extends to visual material such as videoclips and photographs, despite these not being listed. Let's say for example that I have a photograph or a video of a distinctive fruit published as part of a blog. If I describe the fruit on the basis of such imagery (assuming that the ID of the plant is unequivocal and uncontested) then I find it difficult to see how that could be regarded as Original Research. As I commented above, there is no mention of visual material in the guideline, and perhaps intentionally so. Should the guideline be modified to clearly mention and permit the use of such material? Paul venter (talk) 07:42, 11 March 2013 (UTC)

TL;DR: "Visual material" needs expert commentary. —— I think an important issue here is that interpreting the visual material is best done by a reliable source, eg. a peer reviewed journal, etc. Let's use a concrete example. Calappa calappa contains (right now) this content: "The claws are large and flattened, also with a wavy line pattern, fitting snugly under the carapace and providing armoured protection for the crab's front." Then, there is a <ref> inline as if to support this content simply consisting of a link to this image on Flickr. The Flickr page does not contain any commentary or description at all, let alone by an established expert, explaining, for example, how the claws provide protection to the front of the animal. So while we non-experts may infer from the photograph that the claws are protecting the front of the crab, since that is what they appear to be doing, (and I don't doubt that they are) we don't really have any reliable source at this point to justify the claim. In that sense, we as Wikipedians concluding this functionality of the claw from looking at a photograph is original research. So I think it is clear that we'd rather have a journal article or something of that nature to justify the content, not a picture of the claws. But this raises another question—until such a proper reliable source is found, should the image link be included as a reference? It's a pretty good image, after all. I think no. Using the image as a reference says to the readers, I think, that this is a proper reliable source, so it is a bit misleading. It kind of makes us look unprofessional, too, since our sources that we use for all-important WP:V are just flickr pictures lacking expert commentary. Maybe it should be an external link. But WP:ELNO says we should only include external links that are to content we wouldn't on the article itself, and this doesn't qualify as that, either. It's a great picture. But it just doesn't fit in a Wikipedia article, since it isn't freely licensed. ErikHaugen (talk | contribs) 17:26, 11 March 2013 (UTC)
WP:ELNO forbids
  • 1 Any site that does not provide a unique resource beyond what the article would contain if it became a featured article.
  • 2 Any site that misleads the reader by use of factually inaccurate material or unverifiable research, except to a limited extent in articles about the viewpoints that the site is presenting.

and even though it doesn't say so explicitly, the gist seems to be about textual content. Also I agree that only incontestable conclusions should be drawn from visual material, whether video or still. Paul venter (talk) 09:22, 12 March 2013 (UTC)

ELNO doesn't "forbid" anything. ELNO is about what "one should generally avoid providing external links to" (emphasis added). WP:ELNEVER is about what's forbidden.
Additionally, the External links guideline has absolutely nothing to do with reliable sources that support article content, as it says in at least six separate places. WhatamIdoing (talk) 21:02, 13 March 2013 (UTC)
There are actually slightly different rules for including images. See WP:PERTINENCE and WP:OI.
What editors are not allowed to do is to look at a picture and draw conclusions from the picture, e.g., I can see in this picture that the crab has claws (that much is likely permissible), and therefore the crab uses the claws to eat/defend itself/engage in interpretive dance/whatever (definitely not permissible). WhatamIdoing (talk) 01:18, 12 March 2013 (UTC)
Agreed that speculating about function based on a still is OR, but if the visual material is a video clip, TV documentary or film clearly showing behaviour which is unambiguous in its interpretation, then I think that should be allowable. Paul venter (talk) 09:04, 12 March 2013 (UTC)
There isn't anything hugely qualitatively different between a still or a video. An expert can learn or determine things from stills, too. We still rely on expert interpretation where we can. See WP:SECONDARY. ErikHaugen (talk | contribs) 16:27, 13 March 2013 (UTC)
'Function' is about movement. One cannot tell, for example, from a still that a horse's teeth are used for eating, but a video would provide clear proof of this. Paul venter (talk) 07:25, 14 March 2013 (UTC)
PERTINENCE and OI aren't really relevant here, we're talking about links to eg. an image on Flickr, used as references. This is not a question about what images are ok to include in the article. ErikHaugen (talk | contribs) 05:53, 12 March 2013 (UTC)
Whether the link is used as a reference or external link should depend only on relevance and verifiable factual content. Paul venter (talk) 12:23, 12 March 2013 (UTC)
I'm not sure how to parse this. Are you saying whether to use it at all should depend on those things, or are you saying that picking between EL or ref should depend on those things? ErikHaugen (talk | contribs) 16:24, 13 March 2013 (UTC)
Right - I should have been clearer in saying that if a source of visual material is to be used it shouldn't matter whether it's used as an EL or ref as long as the material is relevant AND verifiable. Paul venter (talk) 07:17, 14 March 2013 (UTC)
If you have a reliable source that obviously shows a particular type of crab with claws (still or video, drawing or photography), you can use that source to make a claim that this crab has claws. If you have a source that obviously shows a crab using those claws to eat something, then you can use that source to make a claim that that the crab uses its claws to eat.
This isn't that hard, guys. Maybe you should put the sticks down for a minute and think about what we want. We want accurate, verifiable information in articles, right? Does anyone here honestly believe that crabs don't use their claws for eating? So if we all already know this, then why are we getting so worried about whether the cited source is text or video? WhatamIdoing (talk) 21:51, 13 March 2013 (UTC)
The Reliable sources/Noticeboard should be used when in doubt about the advisability to linking to a particular external image. Bus stop (talk) 22:09, 13 March 2013 (UTC)
The point here is that these were not reliable sources. A flickr image, youtube videos, a gallery on a website with user uploaded content - not taken by or accompanied by expert commentary. Beyond that, in this particular instance, there are original observations made in Calappa calappa that are attributed only to the gallery or image. For example, the gallery contains an image of the underside of the animal and the sentence it's a source for in the article refers to the appearance of this and likeness to horseshoe crabs. Visual material wouldn't be a terrible reference - it just depends on the author and publisher and what information we cite from it, avoiding WP:SYNTH. A BBC film narrated by David Attenborough is an OK reliable source, just as the many NOVA PBS videos are fine reliable sources. But a Youtube video, flickr image, or gallery of images with an unknown author where the information divined from it is an original observation or synthesis seems to be unacceptable when there are likely so many other great reliable sources out there for the exact same information.
FYI, another example of a current discussion along this same topic is the article sandfall where it was brought up at AfD: Wikipedia:Articles for deletion/Sandfall, where it was noted there weren't any reliable sources to verify the article's topic when it was brought to AfD. Cheers, Rkitko (talk) 22:29, 13 March 2013 (UTC)

Let me take this a step further. Can I assume that information taken from Google Street View is reliable (assuming of course that I can reasonably expect that the object to which that information relates will be there for the foreseeable future and that no interpretation of the object is needed - for example stating that there is a speed restriction sign at location X)? Martinvl (talk) 23:00, 13 March 2013 (UTC)

I think there are cases in which there is little likelihood of visual misrepresentation therefore I think this should be decided on a case-by-case basis with recourse to the Reliable sources/Noticeboard as appropriate. Bus stop (talk) 23:18, 13 March 2013 (UTC)

Let's get back to the original question. The OP specifically asked about the omission of any mention of visual media from WP:USERGENERATED, taking the omission as a sign that it was permitted because it wasn't specifically mentioned. Should we add a phrase also cautioning against the use of user generated unreliable visual media as references? Rkitko (talk) 00:50, 14 March 2013 (UTC)

It is already understood that "user generated unreliable visual media" would not be permissible. We find at WP:OI that we are "encouraged to upload … [our] own images, releasing them under the GFDL, CC-BY-SA, or other free licenses", and that "Original images created by a Wikipedian are not considered original research, so long as they do not illustrate or introduce unpublished ideas or arguments, the core reason behind the NOR policy." I think that in some instances an image linked-to would be no different from an image uploaded as concerns its potential to "illustrate or introduce unpublished ideas or arguments". Bus stop (talk) 01:35, 14 March 2013 (UTC)
Quite - I think that if one uploads an image clearly showing a crab species with claws, then one shouldn't be surprised if the casual reader goes away with the impression that a certain crab species has claws, no matter what the text may say. Bottomline - the information imparted by visual material may be a rich source of original research without having to go to the trouble of spelling it out. Paul venter (talk) 14:10, 18 March 2013 (UTC)
I think you've hit on one of the key arguments against allowing such interpretations. You can't upload an image of a crab species. You can only upload an image of an individual member of that species. If you uploaded an image of a snow crab with only one claw, could we say that snow crabs have a single claw? We leave it to experts to identify and describe the characteristics of a species; we don't engage in such original research ourselves. Pburka (talk) 00:00, 19 March 2013 (UTC)
I don't think that "reliability" is the real problem here. You don't need a source at all to claim that a crab uses its claws when eating. Your average 12 year old already knows this. So who cares whether this video is "reliable"? No policy requires any source at all (unless, of course, some user was WP:POINTy or silly enough to officially challenge the statement.
Martinvl, you could use a Google Street View image to support a statement that the speed limit sign (or any sign, saying whatever) was posted at that location. I'm not sure that it would always be acceptable to claim that the speed limit sign is automatically accurate or up to date. WhatamIdoing (talk) 03:47, 19 March 2013 (UTC)
I respectfully disagree. There are many things that your average 12 year old knows which are plain wrong. When I was 12 I knew that scientists can't explain how bees fly. If something is so well known, it ought be to be easy enough to find a reference in a reliable source. Pburka (talk) 01:06, 20 March 2013 (UTC)
If the material is uncontested—and I do seriously hope that you have not challenged material that can be found in children's books—then no citation is required at all, and thus whether this or that citation would be reliable is unimportant. WhatamIdoing (talk) 22:03, 20 March 2013 (UTC)
If I may, the issue is not as simple as you're making it out to be. The OP specifically asked about using visual material as references to describe aspects of the subject -- behavior in a video or detail of a plant's fruit. At best it would appear that you're saying those references are unnecessary since things like that could be verified by visual material that we can trust don't need references. Ok, so let's remove the youtube links, flickr image link, and gallery link from Calappa calappa as I had done before but was reverted. More broadly, though, visual material has its limitations. A description of a plant's fruit written by an expert botanist has the advantage of being written by someone who 1) has seen many, many specimens and not just a single individual so that the final description is based on variety within the species and 2) has been able to examine these specimens in three-dimensions instead of a two-dimensional photo. I think Pburka hit on the right argument here - visual material is far inferior because it represents a single individual, in most cases, and to take data from an image or video and present it as fact about the species at large could be a gross misrepresentation. The inferences and extrapolations bit (WP:OR or WP:SYNTH) from user-generated visual material is another issue, but of course that is prohibited on Wikipedia. --Rkitko (talk) 22:08, 24 March 2013 (UTC)
Agreed that visual material which is not representative and is abnormal or unusual in its depiction should not be used - unless the abnormality itself is being described. One does not need an expert to analyse every image, some of which fall well within everyday experience (such as the colour of the sky or the number of yolks in a double-yolk egg) - at the other end of the range images may well need expert commentary (such as whether the radicle growth of a pine seedling is abnormal in shape, or whether the way in which a left ventricle contracts reflects a healthy heart). Common sense should prevail and unless challenged, interpretations should not automatically be termed original research. Paul venter (talk) 19:40, 25 March 2013 (UTC)
Rkitko, I think that you have confused "reliable" with "best". We only require that sources meet the absolute, rock-bottom, minimum level for reliability. We do not require that the best source be used. If you want to use the best possible sources, then more power to you—but that doesn't mean that you get to challenge material that you know is accurate (e.g., crabs use their claws when they're eating) or blank someone else's barely-good-enough sources. WhatamIdoing (talk) 00:30, 28 March 2013 (UTC)
Images aren't "barely-good-enough". They're simply not good enough. Using a photograph, especially a self-published one, is no different than referring to a blog where someone describes having seen crabs eat with their claws. They're fundamentally unreliable sources. We have no way of verifying that the image is representative or unmanipulated, nor can we verify the interpretation of the image. We cannot use them as references. Pburka (talk) 02:44, 28 March 2013 (UTC)

WhatamIdoing: I certainly have not; I fear that you have have misinterpreted what I've suggested above. How, exactly, can a user-generated image or video be considered a reliable source when it has not been evaluated, edited by, or published by an expert in that field? That has always been the standard for user-generated material. That is the example we've been working with here at Calappa calappa - unknown authors of user-generated material. Earlier in this discussion it was suggested that the authors of these materials could have equally likely uploaded it to Commons. True, but would it ever be appropriate to use material uploaded to Commons as a reference? I have never seen this done and indeed I would not think these could be reliable sources, especially in the case of taking images of individuals of a species and applying that description to an entire species. At Calappa calappa, we have a flickr image that is serving as the source for the statement that the species has claws with a wavy lined pattern. Is that true of all individuals? Is that true of females and males in the species (sexual dimorphism)? Perhaps there's a polymorphism in the species where some populations do not have the wavy line pattern. Is the wavy line pattern only present in adults or is it also seen in juveniles? This is not just a case of the image being a poor source -- it is a case of the image being wholly unreliable because you are entirely unable to extrapolate any information beyond the level of the individual in the visual material. Further, I am also challenging Paul's use of sources on that page for original research. For example, he used a gallery of images, one of which included a photo of the underside of the animal, to describe it this way: "covering the eight ambulatory legs in a design similar to that of horseshoe crabs." The comparison to horseshoe crabs is a novel one as far as I know and appears to be Paul's original thought based on this image gallery. It, along with these rotten excuses for sources, should go. This is why relying on user-generated visual material as a reference is a dubious practice - there is no valuable information in the content that can be applied to the subject of the article. User-generated visual material illustrates an article (ours, from Commons), but I don't think it can be a reference for it. --Rkitko (talk) 02:46, 28 March 2013 (UTC)

I don't think there is a qualitative difference between using an image for illustration and using it as a reference. As I've mentioned before in this discussion the moment an image accompanies an article the reader draws his own conclusions. Writing "covering the eight ambulatory legs in a design similar to that of horseshoe crabs" suggests nothing more than that the one resembles the other, a resemblance that is obvious to anyone with normal eyesight, and should fall under WP:BLUE. To suggest that ANY inference made from an image is original research is going to ludicrous extremes and I'm sure was not the intention of the framers of OR. Let's apply a bit more common sense here instead of hysterically wanting to label ideas with labels they don't deserve. Paul venter (talk) 20:56, 29 March 2013 (UTC)
WP:OI states "Original images created by a Wikipedian are not considered original research, so long as they do not illustrate or introduce unpublished ideas or arguments". Does this mean that one may publish as many images as one likes on WP showing wavy lined patterns on the claws of Calappa calappa as long as the wavy lines are not mentioned in the text? Paul venter (talk) 21:08, 29 March 2013 (UTC)
You missed the point. There is a deep and important difference between drawing inferences from an individual photo for the entire species and uploading a photo to Wikipedia, using it to illustrate an article, and casually describing the individual photo in the caption. The former requires extrapolation that you, the editor, cannot possibly do unless you are making certain assumptions, e.g. "the individual in this photo is representative of all members of this species." The latter requires no assumption and no original research -- it's just an image that illustrates what the text (supported by actual reliable sources) describes. Further, when you go on to make the leap to say something about what you, the editor, thinks the subject of the image looks like, this requires your own knowledge. You may think it's WP:BLUE but does everyone know what the underside of a horseshoe crab looks like? And what, exactly, would an expert in this field say about this sentence if they were happen upon it? Would it be misleading to infer from this photo that the underside of these two species is arranged similarly? Perhaps there are obvious differences. Regardless, this comparison was your original observation, and Wikipedia is not a publisher of original thought. User-generated visual media serves to illustrate articles, not as a references or source of information for it. Rkitko (talk) 00:41, 30 March 2013 (UTC)
@Rkitko You write "You missed the point. There is a deep and important difference between drawing inferences from an individual photo for the entire species and uploading a photo to Wikipedia, using it to illustrate an article, and casually describing the individual photo in the caption. The former requires extrapolation that you, the editor, cannot possibly do unless you are making certain assumptions, e.g. "the individual in this photo is representative of all members of this species." The latter requires no assumption and no original research -- it's just an image that illustrates what the text (supported by actual reliable sources) describes." Well, no....WP:OI does not recognise any difference and states quite clearly "Image captions are subject to this policy no less than statements in the body of the article." Paul venter (talk) 08:02, 15 April 2013 (UTC)
The photograph illustrating Drosera trinervia has been taken from this site. The photographer's credentials are unknown and therefor his ID of the plant is suspect. His note (in Czech) on the page states "This plant has not yet been processed, the data listed here are far from complete". Is this ID original research by the photographer or by Rkitko, or both? and is the photograph representative of the species as a whole or does it illustrate an aberrant form? We have no authoritative support that it is what it says. Paul venter (talk) 14:14, 12 April 2013 (UTC)

This discussion reminds me of a joke (parable?):

An engineer, a physicist, and a mathematician were on a train heading north, and had just crossed the border into Scotland.
•The engineer looked out of the window and said "Look! Scottish sheep are black!"
•The physicist said, "No, no. Some Scottish sheep are black."
•The mathematician looked irritated. "There is at least one field, containing at least one sheep, of which at least one side is black."

Putting this into the Wikipedia context vis-a-vis visual material, we need first to ascertain whether or not the original editor was being honest - did the three passengers in question really see a black sheep in the field just after crossing the border into Scotland? If we are satisfied that they were being honest in what they saw (a 2 year-old can identify a black sheep and a 12 year-old can reliably identify a black sheep), then we can quote the mathematician's view on the basis of the visual image, but we need an expert in order to quote the physicist's or the engineer's views. Martinvl (talk) 13:00, 15 April 2013 (UTC)

Biased or opinionated sources

WP:RS#Biased or opinionated sources

I find this section confusing. If a source is reliable for facts then there should be no need for in-text attribution. If it is not reliable for facts, then it is not a reliable source. The examples provided are also confusing. The policy already says that opinion pieces, even in reliable newspapers, are not reliable sources for facts. Does this section add anything? Can anyone provide an example where this section would apply? TFD (talk) 17:19, 11 April 2013 (UTC)

It gives us an official "guideline" to point at when <insert a nice word for POV pushers here> demand that we ban all use of (pick all that apply: Fox News, The New York Times, anything published in Israel, anything not published in Israel, anything that supports the use of psychiatric medications, etc.).
Furthermore, a source can be "unreliable for facts" and still be a reliable and useful source. See, e.g., anything sourced to a work of fiction. Twilight is "unreliable for facts" (e.g., about whether vampires exist), but it's a perfectly useable primary source for what that novel's plot is. WhatamIdoing (talk) 20:00, 11 April 2013 (UTC)
Four Deuces, consider the following two sentences and how they are referenced...
  • Jews caused Germany to lose World War I <cite: Adolf Hitler, Mien Kampf>
  • Adolf Hitler believed that Jews caused Germany to lose World War I <cite: Mien Kampf>
I hope you would agree that Mein Kampf is not a reliable source for the first sentence. Mein Kampf is, however, a reliable source for the second sentence (despite the fact that it is extremely biased and opinionated). Hope that example clarifies the situation. Blueboar (talk) 01:36, 12 April 2013 (UTC)
I think I agree in general with Blueboar, but the example is poor. Mein Kampf is a primary source for Hitler's views and Nazi ideology. There are shedloads of secondary sources that comment and interpret. Itsmejudith (talk) 17:50, 20 April 2013 (UTC)
RS already says that sources written by subjects are rs for the views of the subjects. Contrary to what WhatamIdoing says, the only time I find editors quoting the section is to say that a source is unreliable because it is biased. For example, they may claim that all the books that are negative on Hitler are written by anti-nazis, therefore unreliable. TFD (talk) 19:17, 20 April 2013 (UTC)

Bogus open access scholarly journals

We need to include a section on phoney scientific journals, see Criteria for Determining Predatory Open-Access Publishers (2nd edition), this New York Times article, this "Sham journals scam authors: Con artists are stealing the identities of real journals to cheat scientists out of publishing fees." by Declan Butler 27 March 2013 in Nature. Obviously this information may be added to appropriate articles also. User:Fred Bauder Talk 11:24, 8 April 2013 (UTC)

And "Predatory publishers are corrupting open access" an editorial in Nature. User:Fred Bauder Talk 11:29, 8 April 2013 (UTC)
Is this really any different from any other source? "But it's a journal" is no more convincing than "But it's a newspaper". Reputation matters. Type of publication is not a guarantee of suitable reputation. WhatamIdoing (talk) 23:07, 8 April 2013 (UTC)
Actually, a list is being published of "journals" with particularly bad reputations. I suggest a brief note and a link to Criteria for Determining Predatory Open-Access Publishers (2nd edition) User:Fred Bauder Talk 11:29, 9 April 2013 (UTC)
Agree to some kind of link/note. Don't we also compile a list of self-publishing "publishers"? Alanscottwalker (talk) 11:53, 9 April 2013 (UTC)
Can we agree that "journals" that aren't even notable enough to be rated by SOA there are also not reliable? I witnessed a lengthy debate involving Haibun Today: A Haibun and Tanka Prose Journal[1] going on on English Wikipedia a while ago, and that and a lot of other online "literary journals" seem to just publish anything. We should probably specify that these are not acceptable as reliable sources. Konjakupoet (talk) 15:11, 21 April 2013 (UTC)

In discussions on what constitutes a reliable source, the point is often made that while a publication may be categorised as 'reliable', portions of it may not be. Conversely, an 'unreliable' publication may contain undeniable truths. The bottom line is that the information world is full of grey areas and that it is extremely misleading to label a source as wholly reliable or unreliable. Paul venter (talk) 14:15, 30 April 2013 (UTC)

Overview clarify

we only publish the opinions of reliable authors

Could we clarify what a "reliable author" is defined as here? Is there another section we can link to? Is this synonymous or different from "reliable source"? I am not sure if there is a situational context to 'reliable'. Reliability is very subjective so we need to clarify what it means to Wikipedia for people to understand this policy.

reliable publication process

What distinguishes a publication process as reliable?

authors who are regarded as authoritative

Regarded by whom? When does regard become "regarded" by Wikipedia standards?

These qualifications should be demonstrable.

What is Wikipedia's standards for demonstrability?

audio, video, and multimedia materials that have been recorded then broadcast, distributed, or archived by a reputable party may also meet the necessary criteria to be considered reliable sources

What is a 'reputable party' by Wikipedia definition? Reputable to whom? Ranze (talk) 03:41, 15 April 2013 (UTC)

The second sentence of WP:RS#Overview would normally read "reliable sources", except that it's a little silly to say that a document, rather than its creator(s), has an opinion.
As for the rest, if you honestly can't glork it from context and an application of common sense, then it seems to me highly unlikely that any explanation will suffice. WhatamIdoing (talk) 22:58, 15 April 2013 (UTC)

Scrambled item embedded in talk page header

IMO the "I personally know that this information is true (or false). Isn't that good enough to include it (or remove it)?" scrambles two two completely different topics (inclusion and removal) together and implies some incorrect advice. It mistakenly implies that there is a verifiability requirement restriction against removal of material. I plan to semi-bold / BRD edit it. Sincerely, North8000 (talk) 11:48, 8 April 2013 (UTC)

Did it. Narrowed it to inclusion, i.e. what is in wp:ver and wp:nor. North8000 (talk) 12:05, 8 April 2013 (UTC)
Hmm... not so sure about that. I agree that we should separate the questions, but we should answer the question: "I personally know that this information is false. Isn't that good enough to remove it?" I would answer that question with: "Not necessarily"
It really depends on whether the material in question is supported by a reliable source or not. For example, a creationist could not go to the Evolution article, say "I personally know all this is false (my Pastor said so)" and gut the article.
I think this needs more discussion. Blueboar (talk) 16:49, 8 April 2013 (UTC)
A perfect answer would take pages to write out, but for now, I've started a parallel question that says only that personal knowledge or beliefs are not enough. WhatamIdoing (talk) 23:11, 8 April 2013 (UTC)
I remember an instance where there was a source, not the best, but a source, that said that someone was dead. As he lives near where I do and I sometimes saw him and talked to him some I challenged that. I got some static over that, but as I had followed his "career" such as it was, for years I was not inclined to just let it go. He had a listed number in the local phone book but people still wanted to argue. User:Fred Bauder Talk 11:41, 9 April 2013 (UTC)
@ WhatamIdoing I think that that is a straw man version of the actual situation, and is creating something that is not in any policy. In any debated item, it's never going to happen that one person can force something to come out based on their personal belief. What you wrote is more likely to get misused in a realistic situation to say that falseness isn't a valid consideration / can't be taken into consideration in a conversation about potentially taking the material out.
On a secondary note, if I'm the only editor on some obscure article, and I see something that I think is obviously wrong, you are saying that policy forbids me from taking it out? Where is this policy that forces material to stay in in that case?
IMHO the real answer takes not even one word. It is to "do no harm" here, and not to try to get into that, and leave it to the other Wikipedia practices and mechanisms to determine what happens on those cases. Sincerely, North8000 (talk) 11:51, 9 April 2013 (UTC)
Re: "if I'm the only editor on some obscure article, and I see something that I think is obviously wrong, you are saying that policy forbids me from taking it out?" It depends on whether there is a citation to a source or not. We actually have three scenarios to deal with:
  1. The information is completely unsourced - WP:BURDEN explicitly allows you to simply remove it.
  2. The information is sourced, but you think the source is less than reliable - WP:BURDEN allows you to challenge the material (or the source). If your challenge is not responded to after a realistic amount of time has passed, you may remove it.
  3. The information is sourced to a reliable source - Per WP:NPOV, you may not remove it. However, you can edit it, to better indicate that there are opposing view points. Blueboar (talk) 13:42, 9 April 2013 (UTC)
I'll add a rare #4: The information is sourced to a reliable source, but that source unfortunately contains an error, and you can demonstrate it with other reliable sources. Then per WP:DUE you are free to remove the error.
What we don't want is to have people gutting articles because "I just know that it's wrong". For example: I've got someone over at Alternative cancer treatment that "just knows" based on her personal experience that a water fast is shrinking her cancer. Shall we permit her to remove sourced statements that dietary changes don't kill cancer cells, because she has personal beliefs to the contrary? WhatamIdoing (talk) 00:08, 10 April 2013 (UTC)
Respectfully and friendly-ly (is that a word?), I think that you are both making up policies that don't exist, and that doing that unilaterally via things embedded in the talk page header is not correct. I think that using Whatamidoing's example is a useful way to discuss this at this time. My answer is that the editors at the article discuss that proposed removal, and with such an incredibly weak argument, the person trying to make the removal would "lose" that debate. And if that doesn't resolve it, use the other Wikipedia methods (RFC etc) to resolve it, rather than misleading people into thinking that there is a policy against that removal. Sincerely, North8000 (talk) 09:58, 10 April 2013 (UTC)
Are there any objections to me removing that? And is so, let's discuss. Sincerely, North8000 (talk) 15:50, 11 April 2013 (UTC)
Removing what? Blueboar (talk) 15:55, 11 April 2013 (UTC)
This item from the FAQ's "I personally know that this information is false. Isn't that good enough to remove it? No. Your personal knowledge or belief is not enough." While well-intentioned it is having the effect of using a straw man example to imply a policy that does not exist. North8000 (talk) 16:53, 11 April 2013 (UTC)

Sure, let's use that example. She says that she personally knows that it works. Do you agree that her personal belief alone would be enough justification to remove a verifiable statement that diets do not cure cancer? WhatamIdoing (talk) 20:02, 11 April 2013 (UTC)

Yes, it's a good example. But you're going to have to hang in there while I split logical hairs. First let me answer what I think should happen at that article. The editors there should discuss it and be free to decide. And if I were in the discussion, my thoughts would go like this. On one side we are getting the opinion of one anonymous editor, saying something that conflicts with reliable sources, and an editor, who even if they are doing their best to be accurate, in reality has no clue has no basis (other than a wild guess on a cause-effect relationship). The weight of their argument for removal of the material is approximately ZERO. They will certainly lose the "argument". It would be a inconceivable travesty if the material got removed.
But that is not the question. The question is: should we put something in a major guideline (faq's on wp:rs) which mistakenly implies that a policy or guideline (rather than editor discussions) dictates the results of that discussion. My answer to that would be "no". Sincerely, North8000 (talk) 20:36, 11 April 2013 (UTC)
So the answer actually is no, but you object to us telling people that the answer is no? We should just keep the answer a secret? WhatamIdoing (talk) 23:10, 11 April 2013 (UTC)
But WAID, the answer to the question isn't a flat "No"... the answer is more nuanced: a) if the "false" statement is sourced, then the answer is "No, your personal knowledge or belief is not be enough to remove it - you need to raise your concern on the talk page and gain a consensus before you remove."... However, b) if the "false" statement is completely unsourced, then the answer is "Yes, your personal knowledge or belief is enough remove it". Blueboar (talk) 00:58, 12 April 2013 (UTC)
First I think that we should acknowledge that we are coming at this from two different angles. I think that you two (WhatamIdoing and Blueboar) are coming at this as say what should happen in a situation such as the one described. And we should note that all three of us are in agreement on what should happen in that particular situation. But where we disagree is on putting a statement in the header which falsely implies that there is a policy restriction against such a removal.
Let me give you a real example to explore this further. There is an article about a (now deceased) famous person where (other than a few tweaks in one-time visits) I have been the only editor for years. I visited this person in their cabin at location "A". When they died, they moved their cabin to location "B",and I visited their cabin at location "B". Then they moved their cabin to location "C" and I visited their cabin at location "C". There was an old statement (cited to an off line source which I don't have access to) (by a long-gone editor) which said that there was one move of the cabin, which was from location "A" to location "C". I deleted that statement. There is NO Wikipedia policy that prevented me from doing that. (as an irrelevant sidebar (because we are talking about the removal),in full disclosure, I added the location "B", I admit that the addition violated policy, and later a major biography was published I bought and which described the move from "A" to "B" to "C" and I was able to source my previously added statement) North8000 (talk) 14:53, 12 April 2013 (UTC)
Blueboar, do you mean "sourced" or "verifiable" in your answer? I have already specified that the statement is verifiable. WhatamIdoing (talk) 22:44, 12 April 2013 (UTC)
I mean "sourced" (ie a citation has been provided). Per BURDEN, if there is no source provided, the information can be challenged (and removal is a legitimate form of challenge). Now... since you stipulate that the statement in question is in fact verifiable, then it probably should be returned... with a citation. At which point the question changes... it becomes "I still think this information is false, can I re-remove it?" And the answer to that question is now: "a) Since the "false" statement is sourced, the answer is 'No, your personal knowledge or belief is not enough to remove it - you need to raise your concern on the talk page and gain a consensus before you remove it a second time'." Blueboar (talk) 14:24, 16 April 2013 (UTC)

I plan to take it out. If anyone objects, that's cool, we can discuss further. North8000 (talk) 12:30, 16 April 2013 (UTC)

I object to taking it out completely... The question ("I personally know that this information is false. Isn't that good enough to remove it?") is an important one that needs to be answered... the hard part is reaching a consensus on how to answer it. The reality is that it can not be answered by a simple "Yes" or "No" ... like so much in Wikipedia, the actual answer is "Sometimes Yes, but at other times No". Blueboar (talk) 13:16, 16 April 2013 (UTC)
Well, to take a close look at that statement, we need to dissect it. And then it comes down to it being ambiguous in two ways:
  • Being in the FAQ's on the talk page of a guideline page, are we implying that the answer is policy/guideline, or that it is just advice?
  • By saying no, such is not "good enough to remove it" do we mean:
  1. Such is not always enough to remove it?
  2. Such is insufficient to win a debate / override others to remove it?
  3. Such is never sufficient reason to remove it?
Sincerely, North8000 (talk) 13:53, 16 April 2013 (UTC)
Re: implying that the answer is policy/guideline, or that it is just advice?... neither. It is simply the answer to a Frequently Asked Question (put on the talk page so we don't have to repeat ourselves by answering the same question over and over again). Our answer should be based on policy/guidance... but it isn't policy or guidance in itself. It's similar to the answers we give to questions at RSN. Blueboar (talk) 14:24, 16 April 2013 (UTC)
But, Blueboar, (assuming that you meant policy/guideline not policy/guidance) this is not based on policies or guidelines. If you feel otherwise, could you point me to the applicable item? Sincerely, North8000 (talk) 16:56, 16 April 2013 (UTC)
I agree... the simple answer "No" isn't backed by any policy or guideline (our policies and guidelines are more nuanced than that). Which is why I keep saying that we need to amend the answer (so that it is based on policy statements and guideline guidance).
Look, this should be simple... We have a frequently asked question: I personally know that this information is false. Isn't that good enough to remove it?
We need to answer this frequently asked question. So... based on what our policies and guidelines say... what do you think the answer to that question should be? Blueboar (talk) 17:12, 16 April 2013 (UTC)
I believe the answer is accurate—your personal knowledge or belief is not enough to remove any verifiable statement. Personal knowledge is insufficient. It is, essentially, irrelevant: If I know—really, truly, deeply, undoubtedly know—that X is ___, Wikipedia just doesn't care. The only thing that matters is what the sources say, not what individuals believe.
An expansion might be better than the bare statement that belief is insufficient. It might be better to say "No. Your personal knowledge or belief is not enough. You must be able to demonstrate that reliable sources have published your personal knowledge, or that no reliable sources have published the material that you believe to be erroneous." WhatamIdoing (talk) 01:13, 17 April 2013 (UTC)
OK... but what if I not only personally know the information is erroneous, but also suspect that it is unverifiable? Can I remove it in that situation? Blueboar (talk) 01:36, 17 April 2013 (UTC)
Assuming that by "remove it", we mean "remove it and have a reasonable expectation of it staying removed", not just "remove it temporarily to indicate a CHALLENGE that will probably result in the material being restored along with a citation to a reliable source, assuming any knowledgeable person notices the removal":
  • If it is uncited and a reasonable preliminary search produces no evidence to support it, then you may remove it.
  • If it is uncited and a reasonable preliminary search produces some evidence to support it, then you may not remove it.
  • If it is cited (to a passably reliable source that actually says what it's purported to say), then you may not remove it.
Of course, if by "remove" you mean "formally CHALLENGE the material", then you may remove anything uncited whenever you want, up until someone decides you're being disruptive and POINTy. You just shouldn't expect most of the challenged material to remain removed, since most uncited material. WhatamIdoing (talk) 03:09, 17 April 2013 (UTC)
OK... but that means that the answer to the original question ("I know it's false, can I remove it?") isn't a simple "No, you can not" (nor is it a simple "Yes, you can")... it means that the answer is a much more nuanced: "It depends". (It depends on whether the information is reasonably verifiable or not). Blueboar (talk) 13:58, 17 April 2013 (UTC)
The FAQ doesn't say "No, you cannot". It says "No, your personal knowledge or belief is not enough." That's "not enough", meaning that it depends on other factors, including verifiability and neutrality. WhatamIdoing (talk) 15:57, 18 April 2013 (UTC)
Ok... I see what you are saying... in which case we still need to amend the answer... to explain why it is "not enough"... we need to tell the reader (briefly) what some of the other factors that have to be considered actually are. Blueboar (talk) 12:09, 19 April 2013 (UTC)
I agree that a brief expansion would be helpful. Perhaps a link to WP:VNT or a similar essay would work. WhatamIdoing (talk) 17:16, 19 April 2013 (UTC)

I think that the fact that we are trying to answer an ambiguous question (what do you mean by "good enough to remove it") is preventing moving forward. So it must be answered which of these meanings of the ambiguous "good enough to remove it" are we trying to answer:

  1. Such is weak or insufficient to win a debate / override others to remove it?
  2. Such is never sufficient reason to remove it?

Sincerely, North8000 (talk) 12:39, 19 April 2013 (UTC)

I think that the inevitable answer will be that policy / guidelines do not dictate the answer, and so other Wikipedia processes (discussions, RFC's etc.) are free to decide whether or not to remove it under the zillions of possibilities. North8000 (talk) 12:54, 19 April 2013 (UTC)
This isn't addressed to groups of editors in discussions or RFCs. It's addressed to lone newbies: "I personally know", not "a bunch of people have looked into the sources and determined". I think you are trying to answer a completely different question. WhatamIdoing (talk) 17:16, 19 April 2013 (UTC)
Again, it's clear that we are viewing this from different angles. You two are approaching it as (excellent) good advice and general theme which is applicable 99% of the time. I am approaching it from a "logical dissection"/ structural analysis type view and the impact of such on some bigger picture items outside of those cases. I might try a tweak to resolve this. North8000 (talk) 20:58, 20 April 2013 (UTC)
I did a tweak. IMHO it would be better to leave it out (for above reasons) but this is a possible compromise. Sincerely, North8000 (talk) 19:05, 24 April 2013 (UTC)
Was subsequently modified by WhatAmIDoing, and then tweaked by me to include a point from the question which it answered. IMHO it would still be better to leave it out. North8000 (talk) 11:17, 26 April 2013 (UTC)
North has opened a question at WT:Talk page guidelines about whether FAQs should be permitted. WhatamIdoing (talk) 17:44, 9 May 2013 (UTC)

Don't use the best source, and doesn't have to be appropriate to the claims made

WhatamIdoing, can you please link to the discussion before this change to WP:RS? In ictu oculi (talk) 15:09, 27 April 2013 (UTC)

I changed this:
"Each source must be...the best such source for that context."
to this:
"Each source must be...an appropriate source for that content."
How does adding the word appropriate turn into "doesn't have to be appropriate to the claims made"? How does this become "Don't use the best source"?
There is no prohibition on using the best possible source. However, we don't actually require the best possible source. We require (NB the difference between "prefer" and "require") a source that is at least minimally reliable for the content. We do not, for example, say tell people that they have screwed up because they sourced the sentence "Abraham Lincoln was the president of the USA during the American Civil War" to a pop history book or a government-run history website, when "the best such source" would be a major history text. We don't require "the best such source". We do require appropriate sources. WhatamIdoing (talk) 20:28, 27 April 2013 (UTC)
WhatamIdoing,
Thanks for reply, but can you please link to any discussion first so we can see what was already discussed before going further. In ictu oculi (talk) 18:28, 28 April 2013 (UTC)


Nice personal attack. It's not just IIO: it's all good-faith Wikipedians. You are trying to force your sources on every article on Wikipedia, allowing Britannica and the New York Times to determine how we write our articles on all subjects, regardless of whether those subjects get any significant coverage in your sources. Konjakupoet (talk) 10:28, 28 April 2013 (UTC)
Additionally, the guidelines should not specify that we are looking for "adequate" sources. If an article is "adequate", we can still replace barely adequate sources with better ones. The proposed change to the guideline would effectively ban such replacements. By the way, LBW: how did you find this discussion? It wouldn't be that you are "stalking" In ictu oculi? Konjakupoet (talk) 10:36, 28 April 2013 (UTC)
BenW may not have said it well, but he actually raises a valid point... editor's can get into huge arguments over whether source X, source Y or source Z is "the best possible source" for some bit of information - and that is a needless debate when all three sources would be appropriate. "Best" is often a matter of opinion.
WhatamIdoing's point is also valid. We are not required to reach perfection in our articles, especially when they are just starting out. ANY source that appropriately supports a given statement is OK. Sure, if someone thinks some other source would be better, they can always swap it in... but that is not a "must"... it's more of a "should".
Perhaps there is a middle ground... start with WAID's "Each source must be...an appropriate source for that content." then continue with something like "Editors are encouraged to cite the highest quality sources that do so." Blueboar (talk) 22:54, 30 April 2013 (UTC) (forgot to sign when I wrote this).
That still implies we're not encouraged to change an "adequate" source to a better one, but rather that the initial editor is encouraged to use the best source possible. Konjakupoet (talk) 11:19, 30 April 2013 (UTC)
Although I personally prefer to use the highest-quality source, I'm not fully convinced that it's entirely appropriate, even if we completely agree on what "highest-quality" means. For example: is it actually better for our readers to exclusively use expensive or print-only medical journal articles, even when we're sourcing simple statements ("The common cold is caused by viruses, not by bacteria") that could be adequately sourced to free, online sources that they can understand? We have people fuss about using only scholarly writings about history (in particular), but why not use something simpler and more accessible for simple statements, like "Abraham Lincoln was the 16th president of the USA"? "Fanciest source" might not always be best for our readers or for our future editors. The "highest-quality source" in some instances might actually be harmful to maintenance and less helpful to our readers than a second-best source or a good-enough source. WhatamIdoing (talk) 22:49, 30 April 2013 (UTC)
Realistically, though, statements like "The common cold is caused by viruses, not by bacteria" or "Abraham Lincoln was the 16th president of the USA" are not going to get challenged based on the relative reliability of their sources (in the latter case, I wouldn't even expect an inland citation). The proposed change is much more likely to affect more controversial statements where a barely adequate source contradicts respected, reliable sources. Konjakupoet (talk) 14:10, 1 May 2013 (UTC)
Realistically, requiring "the best such source" has been used by some POV pushers and perfectionists to remove even basic statements because the first source supplied wasn't perfect. We're not writing this guideline for people who are already good at this. We're writing it for people who need to be told these things, which means both people who are providing lousy sources (nearly all of us when we were new) and also people who are insisting on extremely high standards. See, e.g., multiple RSN discussions over whether all claims about historical events must be supported by sources written by a card-carrying, paid-up member of some mythical professional historians club, or whether it's just possible for basic and undisputed historical facts to be sourced to a biography or (far more shocking) to sources written by a journalist. We have to solve both problems here, not just the simple one. WhatamIdoing (talk) 21:30, 1 May 2013 (UTC)
Exactly right... using the "best" sources is a goal... not a requirement. The requirement is for "appropriate" sources. Blueboar (talk) 21:48, 1 May 2013 (UTC)
If someone tries to remove the statement that Lincoln was the 16th president because it is poorly sourced, please take it to WP:DRN or perhaps WP:RSN. Changing the guideline here is just going to provide more ammo to people who want to push a POV via the use of "adequate" sources. We as Wikipedians should be allowed to remove information that is backed up by barely adequate sources and replace it with (contradictory) information found in the best possible sources. I don't understand Blueboar's argument here: of course the best sources are a goal: why should we change the guideline to effectively ban the insertion of the best possible sources?? Konjakupoet (talk) 00:10, 2 May 2013 (UTC)
Konjakupoet, I am not sure where you get the idea that anyone is talking about "banning" use of the best possible source. If two sources both support the same material, then it is absolutely OK to replace an "only adequate" source with a "better" one. I think WhatamIdoing's proposal supports that. All he is saying is that this is not a requirement. Now, you raise the issue of replacing one source with a contradictory source. You argue that people "should be allowed to remove information that is backed up by barely adequate sources and replace it with (contradictory) information found in the best possible sources". Absolutely NOT... our WP:NPOV policy explicitly says we should NOT do this. If two sources are reliable (but contradictory), we must present both viewpoints. While we can add what the "best" reliable source says, we can not remove what the "only adequate" source says. We can give the better source a bit more weight (assuming we can agree on which source is "better"... but we can not replace one with the other. Blueboar (talk) 00:37, 2 May 2013 (UTC)
I'm concerned with what happens when two sources don't support the same material, because one is written by an expert and published by a university press, and the other is a simplified text written by a layman. I have already linked to the relevant dispute above. Basically, I don't want people to be able to say "WP:RS says we don't need to cite your 'scholarly' sources if we have adequate ones already. And you're not allowed remove material that is backed up by a reliable source." This would, effectively, ban the use of the "Your source is wrong, and is contradicted by these more reliable sources. Therefore, it should be replaced by the more reliable ones." argument. Further, WhatamIdoing has already stated that the purpose of the proposed amendment is to ban the use of obscure scholarly sources for simple statements where "adequate" sources will do, so your statement that it would still be okay to replace an adequate source with a better one is flawed. Konjakupoet (talk) 00:57, 2 May 2013 (UTC)
As I said, What happens is that we mention what both sources say, but give more weight to the expert than we do to the layman. We don't say "your source says something wrong"... we say "your source says something different". That's WP:NPOV. Blueboar (talk) 02:04, 2 May 2013 (UTC)

Sorry, no. Like in the example I already gave, WP:WEIGHT says what we do with a barely adequate source that contradicts a factual statement in 99.9% of our reliable sources. The pronunciation of 十手 (jitte) is NOT a POV issue. Konjakupoet (talk) 02:56, 2 May 2013 (UTC)

Or, how would you suggest dealing with a situation where 99.9% of sources say the word is pronounced one way and 0.1% incorrectly say the opposite. Should we say "A jitte<999 refs> (十手, also pronounced jutte<1 ref>) was a ..."? That seems somewhat unrealistic, although it's the only way to be "NPOV" and not violate WP:UNDUE: I would be more open to "A jitte (十手) was a ... [Last paragraph of intro] Some English-language sources incorrectly spell it as jutte.<as many refs as necessary>" But that would violate your interpretation of NPOV as not allowing us to say "your source says something wrong". Konjakupoet (talk) 09:08, 2 May 2013 (UTC)
We hardly ever write in an article that a source is wrong, we just don't use that source or the claim put forward by the incorrect source. This isn't very well described in policy and guidelines; the closest I can find is in WP:RS#News organizations which says
News sources often contain both factual content and opinion content. "News reporting" from well-established news outlets is generally considered to be reliable for statements of fact (though even the most reputable reporting sometimes contains errors).
Incorrect sources come up all the time, but this is mostly in the form of information being updated. No one would argue that the population of the US as reported by the 1900 census must be reported as a minority view when an article is stating the current population.
In any case, keeping "best" in the guideline does nothing to help the situation of a POV-pusher or troll wanting to present incorrect information from a marginal source. Such a person will simply insist the marginal source is actually the best source, and give some spurious reason for so stating. "Best" is a direct contradiction to Wikipedia's slogan, "the encyclopedia anyone can edit" because many editors are able to make good contributions but don't have access to the best sources. Jc3s5h (talk) 13:47, 2 May 2013 (UTC)
Anyone can edit Wikipedia as is. The "best" or "adequate" distinction doesn't make a difference to that. Someone who doesn't have a source they can easily identify (good or great, doesn't matter) should probably not post something on Wikipedia. If they have an adequate source, they can cite that. If other editors don't like the source, they can change it to a better source. If the content they are trying to add is in some way controversial (or exceptional), then an "adequate" source won't be adequate enough. This policy is already made quite clear on this page. The proposed change won't "fix" anything; and as I have already said will only serve to justify using "adequate" sources to back up exceptional statements. Konjakupoet (talk) 17:10, 2 May 2013 (UTC)
As you note below, only an extraordinary source can support the weight of an extraordinary claim.
As for the pronunciation problem, you address it as DUE weight. If 0.1% of sources claim X, and 99.9% claim Y, then you say Y (and with sources that lopsided, even mentioning X may be WP:UNDUE). That is not a verifiability dispute; we can easily verify that most sources say Y and that a few say X. That is purely a due weight/NPOV dispute. WhatamIdoing (talk) 19:44, 2 May 2013 (UTC)

I think that the above-described change made by WhatAmIDoing is good and fixed a significant problem. The difference is that what they wrote is structurally sound, vs. just putting "generally (but not always) good advice" put in a place that will get used and misused as being policy. North8000 (talk) 11:33, 30 April 2013 (UTC)

Arbitrary break

WhatamIdoing, what on earth do you mean by "some mythical professional historians club"? Of course there are professional historians. There is a scholarly community of historians with its journals, academic publishers and courses at reputable universities. Maybe you think all that stuff is elitist, but you're not going to wish it away. I also object strongly to "POV pushers and perfectionists". Good sourcing is an essential weapon for those of us committed to NPOV. We have seen that time and again in disputes around nationalism. Itsmejudith (talk) 21:42, 1 May 2013 (UTC)
I agree completely. Konjakupoet (talk) 00:10, 2 May 2013 (UTC)
Of course there are professional historians. And there are a couple of Wikipedians, including you Itsmejudith, who have tried to exclude all viewpoints and all sources except those presented by professional historians, from Wikipedia. The proposal at HISTRS failed because this professionals-only-no-biographers-need-apply approach to sourcing violates community practice, the NPOV policy, and good editorial judgment. WhatamIdoing (talk) 19:52, 2 May 2013 (UTC)
Not really. If you look closely at the discussion both at Talk:Jutte and Talk:Kyahan, Darkness walks was trying to block out the view of people who aren't "professional historians", and I was trying to emphasize the view of linguists. "Professional historians" don't necessarily speak the language of the culture they are discussing, and might make mistakes. In this case they appear to have been deliberately following usage in martial arts magazines. (Isn't a "professional historian" just someone who makes their living from "history", rather than someone who necessarily has a qualification or any credibility? When asked if he was a professional historian or a theologian, Bart Ehrman, slightly befuddled, basically said "Well, I get paid to teach history, and don't get paid for teaching theology. Go figure.") Konjakupoet (talk) 00:56, 3 May 2013 (UTC)
I apologize for the ambiguous you, which I have clarified above. Judith and I spent a long time discussing what I perceive as the serious flaws of the proposal she helped develop, so it was obvious to her and to me, but not necessarily to anyone who hasn't followed that proposal or discussions at RSN in which its proponents have tried to put its POV forward as accepted policy. WhatamIdoing (talk) 05:02, 3 May 2013 (UTC)

I wonder whether In ictu oculi was reading "best such source" as (1) the best source that appeared in recent revisions of the article or has been mentioned on the talk page, or (2) the best source available in the world. Some editors will interpret it as the best source in the world, and be discouraged from editing because they can't afford a $200 book, can't afford to fly to Europe to inspect a particular archive, or can't read Russian and instead rely on a translation. If it means "best source in the world" editors of modest means or who are mono-lingual will be unable to edit certain articles. Jc3s5h (talk) 01:09, 2 May 2013 (UTC)

Hi Jc3s5h, both (1) and (2). The best of (1) competing revisions when there is a potential edit war in the offing is important, (2) of course is more idealistic, but still important. I give you an example of a WP:BLP stub I dePRODed and sourced a few days ago Toshimitsu Tanaka - a Michigan pdf gives DoB as 1927, a local Aomori education source in Japanese gives DOB as 1930. I have given both sources in footnotes but followed the Japanese source in the lead DoB bracket. I did that not because it was "an appropriate source for that content" (WhatamIdoing), but because it was "the best such source for that context." (WP:IRS stable version).
As it stands WhatamIdoing's edit doesn't have consensus support. Though everything that has been raised so far could in fact be covered by one simple change "best" (absolute) to "better" (relative). In ictu oculi (talk) 01:49, 2 May 2013 (UTC)
I support WhatamIdoing's edit. The previous wording indicates that if an editor wants to create an article, or add useful information to an article, but the sources the editor has access to, although adequate, are not the best sources in the world, the editor should suppress the information and either not create the article, or omit the useful information from an existing article. Jc3s5h (talk) 04:51, 2 May 2013 (UTC)
Have any of you actually read the policy you are trying to amend? The policy already says "Reliable sources must be strong enough to support the claim. A lightweight source may sometimes be acceptable for a lightweight claim, but never for an extraordinary claim.", so I really can't understand your core argument. You have only presented me with highly unlikely hypotheticals of "POV pushers" and "mythical professional historians" who don't think Lincoln was the 16th president, while I have provided a recent, concrete example of where your amendment could be destructive. By adding the proposed wording and muddying the water, you are only going allow disruptive editors to continue citing fringe sources and prevent us from fixing it. Konjakupoet (talk) 09:08, 2 May 2013 (UTC)
Yes: the guideline already says that editors are not required to use the single best source in the world. You may cite a children's book about the 16th US president or you may cite a peer-reviewed scholarly journal article for that lightweight fact.
Editors are separately required to observe DUE weight, which is where your pronunciation dispute comes in. You may not cite any source that claims the 16th US president was a lizard person from outer space, no matter how fancy the author's credentials. Tiny minority viewpoints (e.g., a pronunciation claim supported by 0.1% of [at least minimally] reliable sources) are normally excluded in favor of mainstream views (e.g., the pronunciation given by 99.9% of reliable sources). WhatamIdoing (talk) 19:52, 2 May 2013 (UTC)

propose adding as questionable: children's, adult new reader, and maybe large-print sources

I propose to add to the subsection Questionable Sources that websites and publications written for children and adult new readers are included as questionable. (Adult new readers are adults learning how to read, for whom books are often simplified or oversimplified just as are children's books, so they'll understand content by retaining as they read despite difficulty and slowness with the process of reading.) They are often so simplified to ensure comprehension by readers that accuracy is sacrificed and to ensure acceptability by parents and schools more concerned about avoiding controversy.

This seems to have been touched on significantly only once in this guideline's talk archives, and at that somewhat tangentially. This proposal would not bar use where appropriate, such as in reporting on what children's books say on popular culture and labeling such sources as such. But it would reduce a problem I saw recently when editors disputed whether content was adequately sourced when sourcing was (allegedly) written for children, where the article was about a fairly advanced legal controversy mainly discussed among adults.

I'm also thinking about adding large print publications, although not websites. Of course, by itself, font size does not matter. However, because publishing in larger print generally requires more paper and more costs for shipping and handling and is for smaller audiences, it's likely some large print publications are edited to have less text and readers are probably not told this, lest they feel cheated; and I suspect the editor is often not the editor of the regular-font edition, but a cheaper, less-skilled, editor, who may edit more for style than to preserve, say, intellectual subtlety. The concern about large print does not apply to websites because the cost of enlarging a displayed fontsize is between negligible and zero, inclusive, and is unlikely to lead to chopping of content.

I'll wait a week for any comments, especially respecting large print physical publications.

Nick Levinson (talk) 16:41, 19 April 2013 (UTC)

I think your proposal to include large-print editions, based on nothing more than speculation that someone might go to the significant expense of abridging the text to save a trivial amount of money on paper, is offensive.
I don't believe that children's books or adult new reader texts are unreliable. They are relatively lightweight sources, so they should only support relatively lightweight claims, but they are fully reliable for those claims. Even this little children's book is a good enough source for something simple at Whale shark, like 'Whale sharks live in warm water' or 'Whale sharks are the biggest fish in the ocean'. There are far better sources, but the purpose of this guideline is to help editors differentiate between a (perhaps barely) acceptable source and an unacceptable one, not to identify the best possible sources. WhatamIdoing (talk) 17:24, 19 April 2013 (UTC)
I don't agree that this is the purpose of the guideline. It should do what it says on the packet, give guidance on "identifying reliable sources". Not "identifying barely acceptable sources". More generally, I think we spend too much time trying to ensure fairness between editors of different points of view and not enough focusing on sourcing quality in the encyclopedia. Itsmejudith (talk) 18:26, 19 April 2013 (UTC)
The definition of a reliable source is one that meets the standards put forth in this guideline, even if it only barely meets those standards. A barely acceptable source is a reliable source. A barely unacceptable source is an unreliable source.
More pointfully, we don't really need a guideline to tell editors how to figure out the obvious cases. The best and worst sources are obviously reliable or unreliable. The guideline exists for the purpose of helping people figure out whether or not a marginal source (e.g., a children's science book or a gossip rag) is good enough to be acceptable. WhatamIdoing (talk) 20:28, 19 April 2013 (UTC)
One problem with enumerating unreliable sources is that some editors may treat the list as exhaustive. Any source not listed as unreliable will be assumed to be reliable. Instead of a largely arbitrary list of unreliable sources, we should clearly describe what makes a source reliable so that editors can evaluate any work for reliability. Pburka (talk) 03:58, 20 April 2013 (UTC)
I think children's books are a large category and listing it would not turn the guideline into a list of sources.
My two local libraries don't have the book on whale sharks and Amazon didn't show, for the copy I checked, the author's qualifications in their back-cover image. The author has a degree in English (as probably self-reported), but that's for his ability to write, and I don't see a credential or background that suggests he's knowledgeable about whale sharks or any more general field. I'm not sure I'd count that as equivalent to, say, a well-known encyclopedia, which would usually be tertiary but at least likely credible.
We already accept almost any source as potentially useful for very particular uses. The guideline, which is not as restrictive as a policy would be, is meant for general guidance, not to preclude all exceptions. That context matters is already in the guideline.
Being offended is reasonable; I agree, except that I think it's directed it at the wrong target. I've been involved in publishing and I saw that production and distribution costs are sgnificant and that they're higher for unabridged large-print editions. If paper and distribution costs were not substantial, prices of e-books would generally be close to those of physical books and physical newspapers would not be going out of business. Whether abridgement occurs is a reasonable question to raise and it is unreasonable to discourage it from being raised. It may be that publishers tend not to abridge but in the U.S. freedom of the press generally means publishers can unless constrained by contract or laws on misrepresentation and each publisher can approach the problem in their own way. A public librarian thought they probably tend not to be abridged but was not sure. I took three nonfiction books off of library shelves and found that two (in self-help and history) said they were not abridged although they said other aspects of the books might vary. The third (in sports) was about 750 pages long when the library's catalogue listed the regular edition as 502 pages, so it might be unabridged, although it didn't say so and the publisher was one of the best known (Random House), so why they don't say so if it is unabridged, that being an obvious selling poimt, is puzzling. I have not searched for industry discussion or whether anyone has criticized such a practice. Nor do I need to unless I have to establish the problem's prevalence. If large print abridgements are too rare, then we simply need not include them as unreliable.
If children's books are now agreed upon as being at or near the margin of reliability, then I argue that we should treat them as normally unreliable (even if barely so), not as normally barely reliable. Without more to support it, the whale sharks book, for example, should usually be considered unreliable for a discussion of whale sharks, although it may be reliable as an indicator of what children are taught about them.
I'm not concerned with media directed at all age cohorts, just those directed mostly at children. If a TV program (with a published transcript) is about whale sharks, it may be for children and adults. And it may be assumed that any children's book has to be readable by adults, who otherwise may refuse to let their child have it. But when it's unlikely that, say, a marine biologist even looked at it before publication because it's only a children's book, I doubt it's reliable. I recall Richard Feynman volunteering to review school science textbooks for a Los Angeles school district and finding things like a list of stars with their temperatures with a directive to students to add up the temperatures but not to average them so that the sum might be useful. (If ice cream is 32 degrees Fahrenheit and you scoop it into 10 bowls, the temperature is not 320 degrees F.) He found lots of these problems, found that reviewers often accepted what publishers told them (some reviewers managed to review books that hadn't been published or delivered yet), and that his method of working probably had or would have no lasting effect on reviews performed after he left. The New Yorker a few decades ago had a long article on school history textbooks; one of the problems was that Texas might buy 50,000 copies and specify what will be said in the textbook and another problem was that the authors listed on the cover or title page were not the real authors, who usually were not historians but simply professional writers given specifications. And that's for high schoolers, not toddlers. Allowing for rare exceptions, I don't see reliability there.
Nick Levinson (talk) 17:39, 20 April 2013 (UTC)
I'm sure we've talked about textbooks a lot before. We certainly have on RSN. Generally, the level of textbook that you can be sure is reliable is introductory postgraduate. Having said that, some undergraduate textbooks are excellent, and so are some for the final years of high school. Below that the aim is simplification and whetting the young reader's appetite. Itsmejudith (talk) 17:46, 20 April 2013 (UTC)
By the way, would anyone care to venture a guess on what perentage of children's books rely on fact research performed in Wikipedia? I'm just speculating on that, but if many high school students use Wikipedia for writing their assignments and since children's books are for readers less demanding than most high school teachers, then we can guess. If the percentage is high, chidren's books as a class are not reliable. Nick Levinson (talk) 18:52, 20 April 2013 (UTC)
  • "We already accept almost any source as potentially useful for very particular uses"—but you are proposing that children's books not be considered potentially useful, even for simple facts like the name of the current US President or the temperature at which pure water normally freezes. Why do you want to exclude non-fiction books aimed at school children? This 700+ page reference work is officially a "children's book". It's call number at most libraries is JR 578.76 AQU—"JR" as in "Juvenile reference". The recommended categorization is "Aquatic biology -- Juvenile literature.". Do you really want to say that's an "unreliable source" for basic information about fish and other animals in the ocean?
  • Why are you checking an author's qualifications for a children's science book, but not for newspapers? Haven't you heard the definition of "journalism" as being the career chosen by people who thought science and math is too hard? Reliability is not wholly determined by the author's formal credentials on Wikipedia.
  • My guess for why unabridged large-print books are not listed as being unabridged is because people now assume that works are unabridged unless told otherwise. As you (should) know, the cost of the paper is minor compared to the overhead of printing and distributing it at all. Furthermore, why do you assume that "abridged" means "unreliable"? Do you expect an abridged dictionary to get its definitions wrong? Do you think that an abridged history of WWI will say that Britain lost, or that the treaty was signed in Moscow?
The main point is this: a source must be strong enough to support its claims. A children's reference work is strong enough to support lightweight claims. It's not the best possible source, but it is strong enough for that purpose. WhatamIdoing (talk) 19:21, 20 April 2013 (UTC)
We accept self-published sources for a few purposes; we accept blogs on occasion. I added a Twitter tweet, it was deleted with a challenge, I replied with why it had been there, I put it back, and it's still there. All of these are exceptions. Children's books should be similarly classed. We probably cite autobiographies that are not self-published and I can imagine that someone wrote an autobiography as a children's book only; perhaps that's reliable. The facts you suggest (the current President's name or the freezing point) are reasonable cases (I sometimes misspelled "Barack" when he was running) but I probably wouldn't use one children's book; I might use a few children's books as cross-checks on each other or else I'd find one adult reference, and we hardly want to cite in a Wikipedia article a bunch of children's sources to prove that we cross-checked. I'm finding it hard to think up a fact of the sort you mention for which we could rely on children's books and for which adult sources would be hard to locate. Cheap year-old well-known almanacs would often serve well enough. Using children's books is just too risky. I didn't look up the 700-page work you mention because I'll take your word arguendo that it would be reliable; presumably it would be one of the exceptions we should recognize. But children's books almost never list references; a few might suggest a handful of further readings for the entire subject but not to prove a paragraph's content. At best, they're tertiary, and we don't want to depend too much on tertiary sourcing.
I didn't hear that about journalism but it is often said that it is the first draft of history. I probably wouldn't take an eyewitness contemporary description of a WWII battle from the N.Y. Times unless it was better than one in a modern book because newspapers tend not to provide retrospective editing of the sort that books regularly incorporate. Granted that Wikipedia uses journalism often, and I've cited it. But major news media often, I think typically, especially broadsheet print, hire as journalists people who know a subject besides journalism; if they cover science they probably studied science and did not find it too hard, and if they found it too hard and are journalists they probably practice journalism in another field that was not too hard to learn. Maybe I'm more selective than are other Wikipedia editors; I often check credentials before reading a book. And I edit controversial articles. But we only need sourcing for points that might be challenged; for everything else, sourcing just has to exist. For the kinds of facts you suggest, perhaps no source need actually be cited and we can rely on knowing that an almanac could be cited in the future if someone does challenge a statement.
Which of paper, printing, or distribution is the biggest share of the higher cost is something I have not checked and it doesn't matter here, because a publisher doesn't incur one without the other two. I use unabridged dictionaries and almost always avoid abridged ones except when I have confidence in the abridged one for a particular purpose; and when I cite dictionaries in Wikipedia or Talk it's usually or always the unabridged. That a work is assumed unabridged unless otherwise labeled is a good reason for excluding the large print books that don't affirmatively say they're unabridged for text.
I wouldn't require that only the best sources may be cited; we only need a floor. We're just discussing where the floor should be. And, overall, exceptions may apply. The guideline is meant to express a norm. Within a norm, children's books serve very limited purposes. I wouldn't say that we should say "must not use"; I'd probably say "should not use".
Nick Levinson (talk) 21:30, 20 April 2013 (UTC) (Added a sentence: 21:41, 20 April 2013 (UTC))
Sources must be judged on their individual merits. Trying to broadly classify them is a mistake. I suspect that it's easy to find children's books which are more reliable than, say, the New York Post, to pick a timely example. Of course editors should try to find the best possible sources, but let's not categorically exclude sources based on their target audience or choice of typeface. Particularly, excluding books aimed at young readers could inadvertently end up excluding a large number of potential editors. If a student adds a sentence, or a topic, sourced by one of their school text books, should we delete that as being improperly sourced? We need to encourage these editors. Blacklisting the sources to which they have access is counterproductive. Pburka (talk) 21:50, 20 April 2013 (UTC)
Agree with Pburka, every sentence of that last paragraph. --GRuban (talk) 22:03, 20 April 2013 (UTC)
I also agree with Pburka. I also suspect that this proposal would be relatively harmless in certain areas, e.g., modern history of the Middle East, but just imagine trying to apply it to articles like Skateboarding. Most of the sources published about that sport are intended to be read by kids. If you take out all the sources "for kids", then you're going to be left with medical reports on injuries and not much else. WhatamIdoing (talk) 01:17, 21 April 2013 (UTC)
Skateboarding is an interesting example. There are various kinds of claims in our article. Medical claims have to meet WP:MEDRS; statistics for injuries should be taken from official sources. The development of skateboards should come mainly from specialist magazines, supplemented with manufacturers' technical information sheets and such like. If those magazines are read mainly by under-18s that is not a problem, because the magazine editors and their young audience form a knowledge community that is obsessed with accuracy around details that outsiders find trivial. On the aesthetic grounds, we already have a principle for sourcing of subjective judgements about cultural forms. We look for film reviews in the magazines, newspaper and websites that regularly review films. So we look for reviews of developments in skateboarding styles in those sources that cover skateboarding as an artistic form. Again, it doesn't matter if those are mainly read by under-18s.
Claims about science or history are different. Yes, there are basic facts that are so well known that they are in all the textbooks. In that case, choose a good textbook and not a poor one. I don't see any point in discussing whether a basic textbook is better than a newspaper for science or history claims. We use news sources for news, and replace them with better sources when those become available. We are aware that there can be problems with news sources; see WP:RECENT. Itsmejudith (talk) 08:45, 21 April 2013 (UTC)
On this particular proposal I am more in agreement with Pburka. That does not mean that I think, it would not be good to more fully explain, here, we prefer better sourcing to less better sourcing. I proposed such additions in this edit [2] that were gone over in this discussion [3]. More general wording along these lines, then the present proposal, would be appropriate and helpful, (and replacing with better sourcing is the way these things are done on the Pedia). Alanscottwalker (talk) 09:46, 21 April 2013 (UTC)
Subjects mainly of interest to children and that are covered mainly in children's literature make a good point. They should be an exception, which can be stated, and that would effectively leave children editing articles for their age cohort undisturbed as to how they source, in that this guideline would remain essentially the same as it is now for sourcing their subjects. That also would preserve their role in contributing to Wikipedia and encourage them to continue, but if anyone writes about the speed of light or neutrinos both children and adults should use more suitable sourcing, which amply exist.
Type size was not the key issue; rather, it was abridgement, which likely results from a larger type size driving costs up. This probably would not affect e-books and no one has objected to using unabridged sources in any readable type size. Invidious discrimination occurs against people with visual impairments and it is partly in the availability of literature, in that many books are probably not available in large print and the costs of for them accessing literature are higher on average, but that doesn't affect what we should use for sourcing.
I mostly don't use the N.Y. Post but most editors would assume it is generally reliable, so, unfortunately, it'll generally be considered as above the floor. Children's books that are reliable should continue to be allowed, because they'd be exceptions. The older the evident subadult audience of the source, the more that reliability should be imputed, if all else about it apparently qualifies it.
The Feynman story came from a book by him.
I can imagine a world-class physicist relying on Wikipedia for the speed of light and then publishing replicable research in which that velocity is essential, and even that another physicist replicates the research successfully, but that would likely be because the first physicist either verified the data in Wikipedia's source or already had known the speed of light but had forgotten it momentarily and, upon checking Wikipedia, recognized Wikipedia's figure as correct. In either case, we wouldn't want a children's book to have been the source.
Nick Levinson (talk) 17:52, 21 April 2013 (UTC) (Corrected spellings and syntax: 18:02, 21 April 2013 (UTC))
It seems to me that it boils down to this: children's books are generally less reliable than books for lay adults, which are generally less reliable than academic sources; abridged books are generally less reliable than unabridged books; and we should always strive to use the best available sources. One would hope that editors don't need to be told this, as it seems self evident. Pburka (talk) 21:06, 21 April 2013 (UTC)
Not quite. The main point is that children's sources are generally too unreliable for use in articles, with exceptions already discussed, and some of the posts above show that this is not well understood. And we don't always use the best available sources; we often use sources that are just good enough, and that's permitted now, nor is anyone proposing to require the best for every statement, or most of Wikipedia would have to be deleted. But the floor should be raised a bit or, if that's consistent with what's already understood, the floor should be clarified.
Abridgements are acceptable, by the way, when they're cited as such. Implicitly they already are, when, for example, an unabridged dictionary and a collegiate dictionary from the same publisher have different titles and are cited accordingly. The problem with abridgements comes with works unknown as abridgements because two works have the same bibliographic data and the fact of abridgement occurred but is not stated.
Nick Levinson (talk) 21:41, 21 April 2013 (UTC)

It's been my experience that children's books are very well checked by the publisher and almost completely reliable, particularly if they have been accepted into school libraries. I wonder what evidence the proposer has that this is not true. GeorgeLouis (talk) 22:35, 21 April 2013 (UTC)

(edit conflict) Well, probably none. There are occasionally some simplifications, although modern science books for kids tend to avoid the "Lies to children" type of simplification in favor of simply omitting complex material or glossing over something with weasel words ("most fish..." rather than "all fish..."). Things have changed in the almost five decades since Feynman encountered the first crop of New Math textbooks.
Simplification also happens in sources for adults. A magazine article or a pop science about HIV testing is not likely to explain that the positive predictive value for an HIV test depends on your demographics (with the practical result that a positive screening test for a straight Caucasian-American woman is dramatically less likely to be accurate than the same result, produced by the same test, received by an African-American gay man), and a children's book on HIV is not likely to get into that, either. But in all of these cases, that's really a matter of skipping over details that the editor considers tangential, rather than providing factually inaccurate information. WhatamIdoing (talk) 00:11, 22 April 2013 (UTC)


We can not answer this in generalities. Reliability really all depends on the specific book that someone wishes to use as a source, the specific author who wrote the book... and the context in which the source is used in particualr article (does it appropriately support what is said in the article). The same children's book can be reliable for one statement, and completely unreliable for another statement. If there is a problem with a specific source (as used in a particular context), that problem can be raised on the article's talk page of the article and discussed, or raised at WP:RSN. Another option is to try finding another source that supports the information (it is never wrong to simply swap one source for another that you think is even more reliable). Blueboar (talk) 23:58, 21 April 2013 (UTC)
The argument that "[w]e can not answer this in generalities" and that "[r]eliability really all depends on the specific book ...." is an argument against almost everything in this guideline and a call to repeal almost all of it, leaving only a generality, namely, that editors should use only reliable sources, with no specific definition of reliability. If that was meant, anyone may go ahead and propose it, but I think the result will be far more arguments over which sources may be used. Companies, for example, would happily cite their own websites as authority for everything about themselves.
I think we've all agreed that exceptions in either direction can be accommodated. Also, I think we all agree that some adult sources normally highly reliable sometimes have errors we should not repeat and I think we all agree that some normally subminimal sources are reliable for particular content. Evidently, however, we're now talking about whether there should be any minimum standard at all. I don't think there'll be consensus that there sould be no minimum more specific than that people should apply an undefined reliability. Please, let's try to stay on the specific topic that was proposed.
The Feynman case was, I think, about astronomy and not about new math and the error was not an error mostly limited to new math. I mentioned the article in The New Yorker, although beyond that I didn't provide a citation. I can look it up but it was several decades ago and if you think that disqualifies it (analogously to the dispute of the Feynman case posted above) then I guess we'd need fresh evidence of unreliability. But that's like arguing that we can use self-statements for any claims because evidence of their unreliability is old.
Doubtless adults check every book in a school library or bookstore but they're far more likely to check for legal liabilities like defamation of character, readability, lessons to be gleaned, and the most obvious errors (probably only very few misspell Obama's first name and a book on care of kittens probably has at least minimally acceptable advice not likely to lead to lawsuits or anti-cruelty protests). If they're not giving authors' credentials, they're unlikely to hand a biology book to a professor of biology for verification of biology (if they did, they'd likely say so). I'm unaware that industry practices have changed other than for legal standards or concerns about adult objections likely to impede sales.
I recently saw a young children's biographical book about a fast Olympic swimmer who, it said, eats nearly 10,000 calories a day, a lot of it in vegetables (and a drawing showed a vegetable truck). Fast Olympic swimmers tend to be thin and so is this one. If most of us consumed nearly 10,000 calories a day we'd probably get quite thick in the middle, but it's possible his exercise regimen and physical practices kept him thin. Without another source, should I believe this book? Is it likely a publisher checked the figure? Or would it pass adult scrutiny because it teaches a good lesson for children even if it's not true? Children are not going to check; most wouldn't know how. Should I add the claim as a fact to Wikipedia (if it's not there from another source or contradicted by another source)? I think if I did there's a good chance it would get challenged on the ground that it's only a claim and, if to remain, needs an in-text attribution (as in, "according to ..."). The fact if true is certainly interesting but I'd hate to have to defend a children's book as reliable for that. In general, we should not be using children's books except for, say, children's games and other subjects almost no other source will cover. Exceptions aside, are children's books reliable enough to cite for facts? We on this talk page seem to be split on that.
Nick Levinson (talk) 15:51, 22 April 2013 (UTC)
  • The Feynman story is about a story problem in math texts he reviewed in 1964.
  • Most libraries rely on various types of library review services to identify books that they want to include. School Library Journal is the biggest that I've heard of. Different services have different strengths, but libel isn't important: weeding out libel is the publisher's job, because the publisher is the one who'll have to pay for the lawsuit. WhatamIdoing (talk) 23:07, 22 April 2013 (UTC)
I read Feynman. It was about stars, and that's astronomy. The book he reviewed and to which I indirectly referred said to add up temperatures. The math was not the problem. The lack of a reason to do that math was a problem. The misleading of students (into believing that adding up temperatures when not averaging is useful) was a problem. Maybe you're referring to another Feynman story, and that would make sense, since he reviewed multiple books and found problems were not rare.
Agreed, libel is not important for our purposes. I didn't say it is. I said essentially that libel is one reason why adults review books and that therefore we cannot depend on that kind of review for our purposes. The fact that adults in libraries and schools review books before inclusion does not mean they do so to determine intellectual reliability, which is what Wikipedia needs. Which of the reviewers would have checked the caloric intake? Most children's books I've seen don't tell us the authors' credentials and don't tell us whether a subject expert has reviewed subject content; only a few do. That makes a difference in whether the content is reliable. We don't need a publisher or editor to check every fact but if the author is not reliable as an author then the publisher or editor should check factuality or we have not much to rely on. Someone has to be reliable in creating the work on which we plan to rely. Otherwise, anything in print becomes a reliable source, and we know where that would get us.
Wikipedia "only publish[es] the opinions of reliable authors".
Nick Levinson (talk) 15:05, 23 April 2013 (UTC) Added the last paragraph, corrected a misspelling, and clarified a phrase: 15:15, 23 April 2013 (UTC))
You've provided anecdotal evidence about an an anecdote Feyman told (you haven't identified where Feyman told this or the text Feyman reviewed) and about a children's book you read (you haven't identified the book). Is there any research published in reliable sources supporting your claim that children's books are generally unreliable? Pburka (talk) 15:35, 23 April 2013 (UTC)
It's in his autobiography, Surely You're Joking, Mr. Feynman! - a great book, but an anecdote. The problem with anecdotes is that only the unusual ones are memorable. If a hundred physicists are asked to review a hundred textbooks, and 99 are found mostly correct, but basically uninteresting, and 1 is found riddled with errors, which gets the press? There is no class of sources that can't have some errors found in them. Yes, I absolutely believe Feynman, and loved the book (though am not sure I would enjoy Mr. Feynman - even writing about himself he comes off as rather harsh). But that was one set of texts in the 1960s. We have no evidence that applied to other texts, or is still true now. --GRuban (talk) 16:56, 23 April 2013 (UTC)
As I recall, Feynman gave that as an example of what he found, an eye-catchng one but not an isolate. He would have known the difference between a formal observational study and a single anecdote and what he wrote came in between. He was reviewing the books passed to him by the schools, reflecting the schools' presumptive willingness to rely on the books and reducing his role in selecting books and therefore in bias in himself. He wrote of multiple problems. The article in The New Yorker was to like effect for history. I'm unaware that the school textbook publishing industry has had a methodological overhaul since then and would appreciate evidence of that. And, regardless, authors have to be reliable and merely having authored any book is already not sufficient. Nick Levinson (talk) 16:24, 24 April 2013 (UTC)
The burden is the other way. The guideline already requires that authors be reliable, as I've quoted above. Merely having two hard covers around an author's writing is not enough. Most children's books do not establish reliability. And Feynman found a general pattern. I can go look for his book again but there isn't much need (and he may not have given the title of the book with the bad summing), since the burden is on those who want to use sources with no known author reliability to show their reliability. That's usually impossible for children's books. I don't want to look for the book with the calories and supply the citation because I suspect someone would go cite it in Wikipedia on my say-so when I'm saying that children's books are categorably not reliable, exceptions aside. Can anyone show their general reliability and thereby meet the guideline? Nick Levinson (talk) 15:51, 23 April 2013 (UTC)
I propose we ban all sources written by lefthanded redheads. "Sinister gingers" as they are euphemistically known, are commonly known to be disgruntled members of society, just itching to inject false information into the world's main if not only free online encyclopedia. The burden of proof is now on anyone who wants to use sources written by left handed redheads to prove that they are generally reliable. --GRuban (talk) 16:49, 23 April 2013 (UTC)
We're discussing primarily children's literature. Instead of going off point, please consider whether all types of sources are presumed reliable or only those by reliable authors and excluding some types enumerated in the guideline. If your view is the former, you disagree with the guideline and you may wish to propose amending it. Most children's books I've seen make no effort to show the reliability of their authors. Please advise on how we determine the reliability of an author relevant to content if not from the book or other information about an author. Nick Levinson (talk) 16:24, 24 April 2013 (UTC)
We determine the reliability of an author the same way we always do, by looking at their experience, education, other writings, publisher, and how other reliable sources treat them, all with relevance to the exact statement we are trying to cite. (Oh, and we should also look at their dominant hand and the colour of their hair.) --GRuban (talk) 18:29, 24 April 2013 (UTC)
Then determine it. (The last sentence is already answered.) Nowhere did any of us claim that we should reject reliable children's literature. We should, however, look for reliability and should presume that books for children lack it until found. Nick Levinson (talk) 14:28, 25 April 2013 (UTC)

I picked three books from a library children's department shelves.

  • One had a relevant credential for a consultant. It could be generally reliable; I'm not sure but I'll grant it arguendo.
  • Walker, Kathryn, Mysteries of the Bermuda Triangle (St. Catherines, Ont.: Crabtree Publishing (Unsolved! ser.), 2008/2009) (original text by Brian Innes) (32 pp.), which tries to present both sides with facts, has a Further Reading section at the end, listing four titles but not as referents for any specific part of the book. It has a glossary and an index. Various people are credited with involvement with it. However, no statement is supported with a citation and no credential of any kind is given for Walker or Innes. It is possible to use it to inspire further research; one could take statements and Google them. But, if successful, that would lead to sources that would obviate any need to cite this book.
  • More advanced is Parker, Steve, with Census of Marine Life, Ocean and Sea (Schlastic (Scholastic Discover More ser.), 1st ed. [1st printing? printing of [20]12?] January 2012 (ISBN 978-0-545-33022-0)). It has a glossary and an index. Parker's credentials are not stated. The Census is cited for why one person became a biologist and that it catalogues species and their whereabouts. Various people are credited with involvement with it, including as a consultant Darlene Trew Crist, Director of Communications, Census of Marine Life; a communications director is likely mainly a publicist or a manager of publicists. Skimming it, I found no citation for a fact. It, too, can be used as inspiration for research, but again that would lead to more useful sources obviating this one. A statement at p. [16] says, "[a]bout 250 million years ago, there was just one ocean, a superocean called Panthalassa." The book by itself is, in my opinion, not reliable for that statement.

Nick Levinson (talk) 16:24, 24 April 2013 (UTC)

OK, you listed three, and granted the first one. I'm going to disqualify the second, since it's about the paranormal, an inherently dubious topic, so the most reliable book imaginable would be dubious at best; we might as well demand that fiction authors be proven to be reliable. That leaves the last one. A bit of research seems to indicate http://www.scholastic.com/teachers/book/ocean-and-sea - correct? If so, that page at least gives at least some of Parker's credits: "Steve Parker works at Britain's Natural History Museum, is a Scientific Fellow of the Zoological Society and is author of "The Encyclopedia of Sharks."" I don't see the book itself, but are you not sure that blurb isn't on it, perhaps on the back cover? While it's not perfect, but clearly at least indicates there might be something there. A few more minutes research finds Mr. Parker's web site. It's loaded. http://www.steveparker.co.uk/about: "STEVE PARKER is an author, editor and consultant specializing in information about the natural world, biology, technology and general sciences. Steve has a First Class Honours BSc in Zoology and is a Senior Scientific Fellow of the Zoological Society of London. He worked on the staff at London’s Natural History Museum, and in house at various publishers including Dorling Kindersley and Haymarket." He seems to have the necessary training and on the job experience to write a basic book about ocean and sea. http://www.steveparker.co.uk/books "Steve’s title list as author has now topped 250 – all illustrated non-fiction books. Some focus on individual areas such as sharks, space or [sic] Others cover general-interest sciences for the younger and family markets." He certainly has plenty of experience writing children's books. (Even if not proofreading his web site. :-P ) http://www.steveparker.co.uk/publishers: "Steve’s publishers include Dorling Kindersley (more than 30 titles including various Eyewitness series), BBC, Smithsonian, Open University, Encyclopaedia Britannica, Science Museum, WWF, Oxford University Press, Heinemann, Reader’s Digest, Gareth Stevens, Rodale, Grolier and A&C Black. Award nominations include BBC Blue Peter Book of the Year, Times Educational Supplement Information Book of the Year (twice), and Rhône-Poulenc Junior Science Book of the Year." So quite a lot of respected publishers seem to think a lot of him. If I ran into an author with those qualifications being brought up on WP:RSN I'd unequivocally say he meets WP:RS for this simple overview of this topic. We have an article for Panthalassa that says much what Mr. Parker does, just in more detail. If that's the typical example, I'm afraid you haven't done a very good job of proving your case for unreliability. --GRuban (talk) 18:06, 24 April 2013 (UTC)
The second book is partly intended to debunk the mystery (I said it "tries to present both sides with facts") and therefore, to that extent, is not categorizable as fiction, and I was not presenting it for the parts that would have been intended as fictional. However, I could have offered it for both, since Wikipedia has an article on the Bermuda Triangle, for which sources would be needed for both sides if they weren't cited already. Just for the factual side, many writers specialize in debunking the paranormal; they often oversimplify but they intend to be factual and they should not be classified as fiction writers, and Walker should not be, either. Stating that which one is criticizing as fiction does not make one an author entirely of fiction. The libbrary's call number for Walker's book is J 001.94 W, which is among nonfiction. I am reading a college-level economics textbook that debunks some misunderstandings; it is not thereby a book of fiction (it's Krugman, Paul, & Maurice Obstfeld, International Economics: Theory and Policy (Reading, Mass.: Addison-Wesley, 5th ed. [2d printing? printing of [19]99?] copyright 2000), and debunking is offered at p. 23).
The third book: Such a blurb wasn't found on the cover (a small library label may have covered what it was stuck to but nothing appears in Amazon's rendition of the back cover). The credential you pointed to is fine but you thereby made my point: you found a credential. You established the author's reliability, which is what is required. That's my point: If that is not done, the source should not be relied upon. And the Panthalassa article does not rely upon Ocean and Sea or Parker, thus making our related point that reliable sources can be found without relying on children's books.
If you'd like to try again on the two boks, please do. So far, the second is published partly as a factual authority and you have not addressed that part of it and the third was not acceptable as reliable until shown otherwise and you did so by not dismissing the need to show it. Please tell us how even the third book should be used in Wikipedia's Panthalassa article if reliability had not been shown.
Nick Levinson (talk) 14:28, 25 April 2013 (UTC)
Oh great. You list a book to prove your claim that publications for children should be viewed as questionable. I do research to prove it's fine. "Aha!" you cry. "You did the research, thereby proving my point!" Tell you what. Tell me what I should have done with that book to disprove your point. Otherwise it seems like you're offering a tautology - the book is questionable unless someone proves it isn't, but once someone proves it isn't, that proves your point anyway. --GRuban (talk) 14:48, 25 April 2013 (UTC)
Probably many editors look at a book to decide whether to use it in Wikipedia and don't research the book's authority elsewhere. If you had looked only at the book (and perhaps also considered whether your prior knowledge of the subject was agreed with by the book) to show its reliability, you'd have done what many or most editors already do. For children's books, that's generally not enough. By putting it as questionable, we wouldn't bar its use, we'd just emphasize or strengthen the using editor's obligation to find the reliability of a referent that's so often lacking in the genre. Nick Levinson (talk) 15:28, 25 April 2013 (UTC)
Let me repeat the question. You gave a specific example of a book, presumably intending to use it as specific evidence for your proposed addition: "publications for children are questionable". I did my best to find evidence that that same book was fine. You then said that actually proved your point. So, is your proposal falsifiable in any way? You don't give convincing evidence for your proposal, and you refuse to accept evidence against it. What is your purpose here? You're trying to convince us? How? With what evidence? With what argument? Just repeating your proposal over and over? Children's books are questionable because you say they're questionable, and because you once read somewhere (that you couldn't remember at the time) that someone else once said a children's textbook was questionable? And if someone tries to show that they can be fine, well that proves your point too, because all you're trying to do is bring more attention to the subject? Which is fine, because we can always have more rules, they don't hurt anything? I'm hoping I'm mistaken, but the number of logical fallacies in that summary is impressive. --GRuban (talk) 16:02, 25 April 2013 (UTC)
Falsifiability would mainly be by showing that for most children's nonfiction books reliability is evident in or on each book. That would mean that editors wanting to cite such a book need not take time to do research into an author's credentials that they wouldn't need to do for most adult-level books. For instance, if a book about the moon says on p. 10, "hey, kiddies, I'm an astronaut and I flew to the moon", and the author is not someone like Captain Kangaroo, it's probably at least minimally reliable, and that meets the guideline.
Where we disagree seems mainly to be on the norm on these books. I think most are unreliable on their face, i.e., till shown otherwise from research elsewhere; whereas you think most are, if I understand you correctly, self-evidently reliable. While I plucked three books off shelves and found two wanting, three are a very small sample and you could validly argue that a much larger sample is needed. I'd counter that the norm is so obvious that it's unlikely anyone would in recent years have spent a few thousand dollars on wages to study it scientifically with double-blinding and only a small margin of error or journalistically by interviewing authors, editors, publishers, and reviewers across enough of the industry to be reasonably representative to discover what I think almost any adult already knows. Maybe someone has; if you know of any such study either pro or con, please let us know. Wikipedia is sometimes compared for accuracy to Encyclopedia Britannica. Wikipedia would look quite superior if compared to a roomful of children's books but, to my knowledge, not one has praised Wikipedia that way. High school students looking for shortcuts for writing an overdue term paper on, say, China today would hardly turn to a 6-year-old sibling's books on that very subject even if handy; and if they did their grades would almost certainly suffer.
Respecting the second book in the trio, I add that a nonfiction book about fiction can be useful for Wikipedia if it does not invent new fiction but relays widely-recognized existing fiction. This one apparently purports to, so we would need to evaluate its reliability even for the reporting of the myths. Pretend the Bermuda Triangle article was new and had no sources and you undertook to find sources. That a myth is widespread does not make the myth a fact but it does make its being widespread a fact and for that you'd need a nonfiction source that stated the myth. Unless someone can refute my point about the second book, it supports my point and the count is two out of three were unreliable until shown otherwise from external research satisfying WP:RS, external research not usually needed when citing sources by grownups.
Logical fallacies were not mine; most of your questions are not my statements, so criticizing them as logically fallacious does not mean that my statements were logically fallacious. The evidence against my proposal that I "refused", as you characterize my response, includes that you chose not to review one book (please try to review it if you can refute my argument about it), you made my case for another, and someone said that some of what I said was old or anecdotal but where there's no evidence of the industry having changed or, I think, needing a systematic study. On that last point, significant disagreement might be reasonably grounded but I think almost no academic or parent would agree that books for 6-year-olds are even nearly as reliable as books for adults.
We're not merely repeating. We're responding to different things said and trying to arrive on what is agreeable among most of us.
The quantity of guideline provisions would hardly change percentage-wise by stating this one. Add policies and the percentage drops.
Nick Levinson (talk) 15:06, 26 April 2013 (UTC) (Clarified a phrase and corrected a missing period and two misspellings: 15:19, 26 April 2013 (UTC))
A large public library's children's department senior librarian just told me today that they select books to buy on criteria such as listing further readings and agreed that probably many libraries have similar criteria, so one could argue for Wikipedia's purposes that a children's book being available at a public library leans toward the book's being reliable. (This library discourages the public from donating books, which would keep shelves oriented toward those criteria, but I don't know how many libraries have similar discouragements.) She also suggested a couple of search terms for finding research on comparative reliability, and I plan to follow up on that soon. Nick Levinson (talk) 19:56, 27 April 2013 (UTC)
I searched for articles likely to be relevant (my search term was "accuracy in children's nonfiction" without quotation marks, on the librarian's recommendation) in education journals from 2005 to the present (with the caveat that the latest journal issues are often excluded in databases). My concern goes beyond whether the books have features that support research by students, although that's a good sign, and my concern doesn't include picture quality even though pictures inform, since we don't copy them into Wikipedia so they don't matter here, but books' textual content needs to be accurate and we're talking about being able to judge that by looking only at the book, not authors' or publishers' websites or reviews published elsewhere. What I found: A review of prize-winning children's books (most, of course, are not prize-winning and prizes are often awarded after publication so that first editions usually will not reveal having won a prize) suggests that accuracy is increasing but that significant problems persist. Another article described a method for teachers in choosing books for accuracy in science that depended largely on external evidence, such as asking an expert or searching for reviews; criticized judging books by their availability in libraries or bookstores; and raised a concern about teachers who themselves are not expert enough in science to judge books' accuracy. (Sources in order of discussion here: Prize-winners: Gill, Sharon Ruth, What Teachers Need to Know about the "New" Nonfiction, in The Reading Teacher, December, 2009–January, 2010, esp. pp. 262–264 (in JStor & Eric); Science: Atkinson, Terry S., Melissa N. Matusevich, & Lisa Huber, Making Science Trade Book Choices for Elementary Classrooms, in The Reading Teacher, March, 2009 (in JStor).)
Whether books are in libraries that choose carefully is only external evidence, since Wikipedia editors are not guided by our views of availability as a guide to reliability, since we haven't stated any such consensus views. Whether reviews are positive on point or prizes were won is only external evidence. External evidence can be valid, but editors need to look for it or in some other way judge the reliability of children's books before citing them, and not simply assume the presence of two covers is proof of reliability. That's what the proposal is for.
Nick Levinson (talk) 20:40, 28 April 2013 (UTC) (Corrected hard-space markup: 20:47, 28 April 2013 (UTC)) (Corrected hyphen to en-dash: 20:52, 28 April 2013 (UTC))
I surveyed a sample of 14 books on an open shelf in the same large public library's children's room where I recently spoke with a senior librarian as I discussed supra, thus in a library that buys books partly based on what we view as reliability. I did the survey last Tuesday morning. I tried to neutralize my bias. I didn't select a shelf until approximately the minute I began, the only conscious criterion being that it wasn't the shelf where I recently picked one of three books (see supra); I didn't remember where I'd gotten the other two. No one did anything at the shelf or spoke with or interrupted me while I did the sampling. I started at the far left end of the shelf. Books were positioned with spines out, thus reducing my being influenced by covers or by promotion by anyone by displaying covers. The first book was in call number 530 and the last was in 534; the general subject was science. The readership intended appeared to me to be 3–10 years old, although I don't have expertise on literacy norms by age cohort; one book explicitly said 3–6; others may or may not have had statements about intended readerships but I didn't notice any except the one. Each book stated in text (not considering illustrations) at least one claim of fact that is reportable in Wikipedia if the source is reliable; the claimed fact could have been very elementary; if I found no such claim when skimming, I did not count the book as being in the sample; I excluded no more than one book on that ground. Each book had only one author. I counted books that had the same authorship as separate books; the only instance I noticed of that was of nontentatively reliable authorship. I looked for evidence of each author's reliability. If the author was described as a writer but not as knowledgeable in a field substantively related to the major content, I counted the author as not reliable. I did not consider publishers or other people, except that I counted a consultant in a substantively-related field as establishing tentative reliability (it being unknown whether the consulting included editing a submitted manuscript). I did not research any book, author, or consultant more than in the book. I did not search for reliability within the book's main text, such as for an author's first-person declaration akin to "when I taught this subject at the University of London" (a hypothetical phrase). The result is that 6 were reliable and 3 others were tentatively reliable. Thus, presumptive unreliability was about 36% to 57%. A larger sample (for a low margin of error) from wider topical and literary ranges and across mainly-English-reading nations would be more scientific, but I don't have the money to do that. If the statistics hold closely for the larger sample suggested, a third to a half of children's books being facially unreliable means that the problem is so substantial across the class of children's books that editors need to be advised in advance to look for reliability in children's books. Nick Levinson (talk) 16:50, 2 May 2013 (UTC)

Here's the Feynman anecdote, for those that don't have the book. It is preceded by a discussion of the pointlessness of converting numbers from base 7 to base 5, and followed by a discussion that hints at his doubts that other commission members bothered to read the books, and then by an explanation that he resigned because after a year of distress over math books, he couldn't bring himself to risk next year's review, which would have been all science books. WhatamIdoing (talk) 19:27, 24 April 2013 (UTC)

The Feynman anecdote
The following discussion has been closed. Please do not modify it.

I understood what they were trying to do. Many thought we were behind the Russians after Sputnik, and some mathematicians were asked to give advice on how to teach math by using some of the rather interesting modern concepts of mathematics. The purpose was to enhance mathematics for the children who found it dull.

I'll give you an example: They would talk about different bases of numbers -- five, six, and so on -- to show the possibilities. That would be interesting for a kid who could understand base ten -- something to entertain his mind. But what they turned it into, in these books, was that every child had to learn another base! And then the usual horror would come: "Translate these numbers, which are written in base seven, to base five." Translating from one base to another is an utterly useless thing. If you can do it, maybe it's entertaining; if you can't do it, forget it. There's no point to it.

Anyhow, I'm looking at all these books, all these books, and none of them has said anything about using arithmetic in science. If there are any examples on the use of arithmetic at all (most of the time it's this abstract new modern nonsense), they are about things like buying stamps.

Finally I come to a book that says, "Mathematics is used in science in many ways. We will give you an example from astronomy, which is the science of stars." I turn the page, and it says, "Red stars have a temperature of four thousand degrees, yellow stars have a temperature of five thousand degrees . . ." -- so far, so good. It continues: "Green stars have a temperature of seven thousand degrees, blue stars have a temperature of ten thousand degrees, and violet stars have a temperature of . . . (some big number)." There are no green or violet stars, but the figures for the others are roughly correct. It's vaguely right -- but already, trouble! That's the way everything was: Everything was written by somebody who didn't know what the hell he was talking about, so it was a little bit wrong, always! And how we are going to teach well by using books written by people who don't quite understand what they're talking about, I cannot understand. I don't know why, but the books are lousy; UNIVERSALLY LOUSY!

Anyway, I'm happy with this book, because it's the first example of applying arithmetic to science. I'm a bit unhappy when I read about the stars' temperatures, but I'm not very unhappy because it's more or less right -- it's just an example of error. Then comes the list of problems. It says, "John and his father go out to look at the stars. John sees two blue stars and a red star. His father sees a green star, a violet star, and two yellow stars. What is the total temperature of the stars seen by John and his father?" -- and I would explode in horror.... There's no purpose whatsoever in adding the temperature of two stars. Nobody ever does that except, maybe, to then take the average temperature of the stars, but not to find out the total temperature of all the stars! It was awful! All it was was a game to get you to add.

I'm glad to see I correctly characterized what Feynman said even roughly 20 years after reading it. As I noted, it is somewhat beyond an anecdote, although no one said it was a full-sczale observational study. Thank you for researching and providing the text for others to see. Nick Levinson (talk) 14:28, 25 April 2013 (UTC)
Mr. Feyman has convinced me that we shouldn't use maths texts as sources for astronomy articles. Pburka (talk) 23:52, 25 April 2013 (UTC)
Especially not story problems in math texts. They also shouldn't be used to discuss how many candies a child has after giving some to his friend, how many holes for fenceposts the farmer needs to dig, or when two trains traveling at different speeds will pass each other. WhatamIdoing (talk) 02:15, 26 April 2013 (UTC)
I don't see any support here for Nick's idea that juvenile reference works should be treated any differently from any other non-fiction book. We could keep beating this dead horse, but given the complete lack of support, I think we should just stop. WhatamIdoing (talk) 02:15, 26 April 2013 (UTC)
There is no "complete lack of support". Roughly speaking, Itsmejudith and I appear to agree, WhatamIdoing, Pburka, GRuban, Blueboar, and GRuban disagree and GeorgeLouis disagrees but with a question that someone else disrupted before I answered, and Alanscottwalker has a split view. The purpose of discussion includes seeing if we can persuade each other and I think I'm answering points raised with the possibility of arriving at something consensus can agree on, and there has been movement, especially on subjects written of mainly to children. I understand the convenience of using first reactions as the decising factor and the desire to disrupt efforts that may take a week or so but let's try to see where else more of us can agree. Nick Levinson (talk) 15:06, 26 April 2013 (UTC)
Nick: are you aware of cases where editors have added questionable material using children's books as sources? If so, please give examples. If not, is the purpose of this proposal simply prophylactic? Pburka (talk) 06:53, 28 April 2013 (UTC)
Yes, I know of one that occurred recently, as I noted in the opening post of this topic/section, in the second paragraph, with a link. The editor Itsmejudith wrote in this topic/section, "I'm sure we've talked about textbooks a lot before. We certainly have on RSN." The first matter seems to have settled down satisfactorily, but this discussion shows a split by which some editors clearly believe such sources are reliable in most cases for matters beyond children's subjects and apparently RS/N is to like effect. To see how often children's sources are cited in Wikipedia would require sampling articles and checking for that, a very time-consuming trek. As shown by another editor, Feynman found severe problems in the books he looked at; I already mentioned that The New Yorker wrote of history books for high schools; I suppose some editors now would dismiss both as decades old. While I consider the claim of general reliability an extraordinary claim requiring proportionate proof, I plan as I noted above to see whether there's recent research in either direction on the general reliability of such sources, especially for subjects normally of interest to adults (e.g., not childrens hobbies). My proposal would only require or recommend that reliability be established, not that children's books be barred for any subject even if reliable. Nick Levinson (talk) 18:24, 28 April 2013 (UTC)
That link is to another discussion. No evidence has been presented that there is a persistent problem of editors adding questionable information sourced from children's books. You've simply presented your assumptions with little evidence to support them nor any explanation of the problem you're trying to solve. Can you provide even one example of a questionable edit which used a children's book as a source? Pburka (talk) 07:56, 30 April 2013 (UTC)
I have indeed explained the problem I'm trying to solve, which is that children's books are generally unreliable on their faces and that editors need to be advised to look for reliability, just as we tell them that regarding self-published books (if we allow them at all). If they can find reliability through off-book research, such as publishers' websites, that's fine, but reliability has to be found before citing.
Asking about actual citations in articles is irrelevant, because a now-deleted past usage is an instance that has been deleted, a still-present single instance we find today can already be deleted and should be, if instances still present are too numerous we can ask for help in deleting them, and, because individual books are likely cited in only one or two articles per book, searching for them across Wikipedia requires a large-scale method, such as searching for prolific nonnotable authors lacking substantive credentials stated in their respective books, publishers of many titles some of which lack stated credentials, and large series in which some titles lack authors' credentials and where a series is indicated within book titles with a tolerable level of false positives (comparable probably to the For Dummies series). I plan to see if one of these searches would be feasible; I started to find some names; I did some trial searches and found books cited as sources for subjects not limited to children's interests, although we'd need to eliminate whether their reliability can be established outside of a cited book. This thread is evidence that many editors believe children's books are not a reliability problem, which suggests that children's books are getting in, and in nontrivial amounts.
I presented assumptions because I didn't expect editors here to think children's books were as reliable as adult books, an astonishing concept to encounter. Now I've been posting evidence and getting limited response to the evidence itself. A quotation of Feynman reinforced that I had described what Feynman said accurately. I picked out three books and found two facially wanting and that has not been refuted (another editor's refusal to look is not a refutation). I cited two modern studies that say that significant problems exist and mentioned a journalistic study of high school books' problems, including falsely identified authorship. Now, simultaneously with this post, I've added (inserted supra) about another search, a sample of 14 books. Consider what most parents do: they explain things to their children so the children will understand, and children's lesser intellectual capacity (they know less) needs simplification in ways that would cause failure at a university level. Our sources do not have to be at a university level but they are supposed to be reliable. A notion that children's books are just as minimally reliable as adult books even when they don't state authors' credentials fails the chuckle test.
Nick Levinson (talk) 16:50, 2 May 2013 (UTC)

Your definition of "reliability" is not the one that Wikipedia uses, so your personal opinion that they are unreliable is irrelevant. Here's our list of factors that you need to consider:

  • Does it has a reputation for fact-checking and accuracy?
  • Is it published by a reputable publishing house, rather than by the author(s)?
  • Is it "appropriate for the material in question", i.e., the source is directly about the subject, rather than mentioning something unrelated in passing?
  • Is it a third-party or independent source, with no significant financial or other conflict of interest?
  • Does it have a professional structure in place for deciding whether to publish something, such as editorial oversight or peer review processes?

Your criterion seems to be:

  • Did the publicity department claim that the author is an expert on the book jacket copy?

These are rather different lists, aren't they? WhatamIdoing (talk) 19:38, 2 May 2013 (UTC)

That's a misunderstanding. I'll clarify what we're discussing.
The same policy says Wikipedia "only publish[es] the opinions of reliable authors". We're talking about authors with no basis in the respective books for our assuming they're reliable. And the same policy urges us to "[t]ry to cite present scholarly consensus when available, recognizing that this is often absent." We've already agreed on subjects covered only (or almost only) by children's books, such as some hobbies, for which higher-level sources are unavailable; in those cases, we can cite children's books as the best available sources, but we should not for most other subjects. We're talking about subjects for which adult-level sourcing is available. It is rare that a children's book can be credited with presenting a scholarly view and that one or more children's books can present a scholarly consensus. It should not be assumed that a publisher that publishes both adult-level and children's books applies anywhere near the same standard of fact-checking; it wouldn't need to, given differing customer demand about what's good enough for each readership.
We're not talking about self-published children's books; self-published books of all levels are already covered by this policy and I was not proposing to change that. We're not talking about a publisher's independence or oversight structure; presumably, they're adequate for what they publish. I was not concerned with who wrote a book's jacket, frontmatter, or backmatter but with what the book as a whole says or doesn't say.
It appears that children's books are already being cited in Wikipedia's articles about subjects of interest to adults (whether also of interest to children does not matter). It therefore appears that this policy is not well enough understood as generally steering us away from children's books as generally facially unreliable. It therefore appears we need to be clearer than we have been. Do you have an alternative proposal?
Nick Levinson (talk) 15:51, 3 May 2013 (UTC) (Corrected missing apostrophe: 15:57, 3 May 2013 (UTC))
Of course we only add the opinions of reliable sources; to add our personal opinions would violate NOR. There is nothing in that guideline that says that the author of a children's book is unreliable. There is also nothing in our guideline that says non-scholarly people are unreliable authors. And I don't ever remember seeing a children's book being cited to support an opinion anyway. "The whale shark is the largest fish" is not an opinion, and could be sourced to any children's reference book on whale sharks. "Theater critic Judith Martin thought that the Star Wars film was lousy" is an opinion. I doubt that you will find any children's books being cited for opinions, and I further doubt that you will find a "scholarly consensus" on the merits of the Star Wars film.
My proposal is that we change "reliable authors" in the Overview to "reliable sources", so that you will no longer believe that the author's credentials are the sole factor in whether a source is useable. WhatamIdoing (talk) 16:27, 8 May 2013 (UTC)
No one proposed adding our personal opinions to articles and no one said all nonscholarly sources should be excluded. We're only discussing children's and related sources. We've been talking almost entirely about factual claims and hardly at all about opinions and the Star Trek point is completely irrelevant. Editor WhatamIdoing, if you really don't understand what we're talking about, please try to read what some of us are posting more carefully, because responding as if we're saying things we're not is a distraction from attempting to arrive at a consensus on the issues raised. If I'm being unclear, I'm glad to try to be clearer. Tell me what it is that you don't understand.
Children's books are being cited in support of facts in Wikipedia and, if reliability is lacking, as it often is for children's books, then we have a problem, and finding most of those citations is a daunting task, especially if we have to do the research into reliability that the original editors adding the sources should have done and probably didn't because the guideline didn't make the distinction. "The whale shark is the largest fish" (if true) is a statement of fact that needs to be sourced to a reliable source. Just because someone stated it in a book is not enough, as I think we would all agree. The book has to be reliable. It can be sourced to a children's book but the children's book has to be reliable. Many, on their faces, are not.
Changing what it says in the Overview from "reliable authors" to "reliable sources" would make no difference to the issue at hand (it also contradicts your view on that wording earlier this month but you're free to change your mind). Whether the Overview should say "reliable authors" or "reliable sources", reliability is needed and, for children's sources, tends to be lacking.
I agree with this posted statement: "There is nothing in that guideline that says that the author of a children's book is unreliable." We need to add that children's books are ordinarily facially not reliable and that editors need to find reliability, not simply assume it's there. The guideline says, "Reliable sources may be published materials with a reliable publication process, authors who are regarded as authoritative in relation to the subject, or both. These qualifications should be demonstrable to other people." An author of a children's book often has no credential other than being a writer, and in that case usually cannot be considered as authoritative on the subject until more evidence is found. Many of the publishers are not known for their reliability as to content, and in that case more evidence is needed. That evidence may be available online or in catalogues, but if it's not in the source it has to be found elsewhere, perhaps online or in a catalogue. That's not a terrible burden on Wikipedia editors who want to add from a children's source for an adult-level subject. If evidence of reliability is lacking, citing an adult-level source may be easier and would enhance Wikipedia's quality.
Most adult-level sources that I consider citing can be checked for prima facie reliability in less than a minute. That's my experience, albeit probably biased by the kinds of subjects and sources I tend to choose. There are exceptions but I find that most sources are quickly checkable. It takes me less time to check an adult-level source for meeting Wikipedia's guideline than it would to check a children's book if, for the latter, the evidence is outside of the source. I don't cite children's books for adult-level subjects because it would take too long. But if an editor is willing to use a children's source, they should look for reliability. Why would anyone think that's wrong?
Nick Levinson (talk) 15:45, 10 May 2013 (UTC)
Nick, children's books are ordinarily not lacking reliability. Children's books are ordinarily reliable for most basic facts.
Some children's book lack dust-jacket copy that claims scholarly credentials for the author. That's not the same thing as "lacking reliability". The author's credentials are not the definition of reliability. A reliable publication process—that's the ordinary process exercised by reputable publishers, including all reputable publishers of juvenile non-fiction—is adequate on its own. WhatamIdoing (talk) 15:57, 10 May 2013 (UTC)
Do you have evidence that "children's books are ordinarily not lacking reliability"? I gave evidence above that they ordinarily do lack it. That evidence still stands. Contrary sourcing or other evidence would help.
That a book happens to be correct about a fact is not relevant, since that depends on our knowing the fact from another source (itself presumably citable) and the coincidence of being correct does not, by itself, make the tested source reliable.
Reliability has to be shown through either the publisher's process or the author's credentials, according t the guideline as quoted above. If the author's credentials are lacking, the publisher's process has to be present. The book itself often shows neither. We should not assume that the book having a major publisher (and many don't) is the same as the book having a reliable publisher.
Nick Levinson (talk) 17:02, 11 May 2013 (UTC)
Major or minor, if they have a good reputation, they would still have a reputation to maintain. It is difficult to believe that a reputable publisher of high quality children's educational materials would just not care for the sake of their brand protection alone. They would have librarians, and professional reviewers clamoring not to buy their products. -- Alanscottwalker (talk) 17:33, 11 May 2013 (UTC)
But that requires checking both reputability and quality level and is limited to those sources known as educational (presumably not just any children's books but those meant for schools to teach from), and I'm proposing checking reliability, which is more to the point. Most books do not get reviewed except perhaps by buyers who don't tell us their findings. There isn't much review space available anymore. Your phrasing would still require validating and my proposal is simpler. Nick Levinson (talk) 21:00, 11 May 2013 (UTC)

Draft proposal

Consensus appears to be presently against this proposal, but, given the evidential nature of the respective claims, I am requesting comments on it. If supported by the result, I would add the following as the last subsections in the section Self-Published and Questionable Sources, after all other subsections, subject to editing of the following:

=== Children's sources and new adult reader sources ===

Sources written for children, especially for young children, generally tend to be unreliable. The main exceptions are subjects that are of little interest to adults, for which no other sources are likely to be better; these subjects may include children's games. To determine reliability, look for evidence that the author has expertise on the subject and not only as a writer.

New adult reader books, although written for adults, are written for adults who are first learning how to read and therefore need simplification so they can concentrate on the slow and difficult process of learning to read. Therefore, adult new reader books should be judged for reliability like children's books.

=== Large print books ===

Large print books may state on a cover, in the frontmatter, or elsewhere that the text is the same as that in a regular print edition. Without such a statement, unannounced abridgement should be presumed and reliability has to be established. If such a work is to be cited and sameness of content is not assured, include in the citation that the large-print edition is the one being cited.

This does not apply to websites and online documents, because their cost of font enlargement is negligible and they're unlikely to have abridged their content because of costs associated with larger font sizes.

Request for comments: Whether children's sources are nearly as reliable as adult-level sources

Are children's sources generally reliable enough for adult-level facts (besides in subjects mainly of interest to children) so that Wikipedia editors need not be advised to look beyond a source itself to find out whether a source meets WP:RS? A similar question applies to large-print books not described as full-text. Nick Levinson (talk) 19:56, 18 May 2013 (UTC)

This would not bar use of children's, adult new reader, or large-print sources not labeled as full-text for any fact, but simply advise editors to ensure that sources they intend to cite for adult-level facts are reliable. I posit that evidence shows that such a source is too often not visibly reliable if we check only the source itself; some editors posit that there's little or no evidence of lack of reliability. (This clarification was added after a comment infra.) Nick Levinson (talk) 16:32, 19 May 2013 (UTC)
Proposed guideline provisions appear on this talk page immediately before this request for comments. Nick Levinson (talk) 17:18, 19 May 2013 (UTC)

Survey

Please add threaded discussion only to the Threaded Discussion section below, not to this Survey section, which is for support, opposition, and the like. Nick Levinson (talk) 19:56, 18 May 2013 (UTC)

  • Oppose: for the many reasons given above. Bondegezou (talk) 21:04, 18 May 2013 (UTC)
  • Oppose. There's little evidence that the premise is valid. Pburka (talk) 22:13, 18 May 2013 (UTC)
  • Oppose Nick's ongoing campaign to declare that juvenile reference works are presumptively unreliable. There is no good reason for us to put textbooks and reference works written for schoolchildren and published by respected, reputable publishers in the same class as supermarket tabloids and personal blogs. His rationale for defining all large-print works as being unreliable—unless the large-print copy explicitly states that (1) a small-print edition also exists and (2) the contents of the large-print are exactly the same as the small-print edition—is even weaker and smacks of bigotry. If he'd really meant "abridged", he should have just said "abridged", rather than singling out a format that is accessible to people with vision impairments.
    Per long-standing consensus, a source needs to be strong enough to support its claims. For a lightweight claim, like the fact that Abraham Lincoln was the 16th president of the United States or the fact that frozen water is called ice, a lightweight source like a children's book or an abridged book, is good enough and perfectly acceptable under this guideline and all of our content policies. WhatamIdoing (talk) 18:45, 19 May 2013 (UTC)
Juvenile reference works and textbooks are only subsets of children's books and are more likely to pitch their believability by stating authors' credentials in or on the books; in that case, the proposal would permit their use without further investigation, thus not hindering any editor planning to cite one.
Reputation depends on knowledge outside of a book, and that's part of checking reliability. A book itself in naming its publisher does not establish reputability, although publishers wish it would. If the reputation is that of factual accuracy and that can be found, that is all we'd ask of editors: find out if the publisher is reputed for being factual such that the book is reliable.
Books that are labeled as abridged were not included because what I found is that some large-print books are labeled as full-text and others are not labeled on the point at all; and, absent the label, to read two editions of a book to discover if one is abridged is too much work to expect of most editors in most cases, as it would mean comparing editions page for page between often-unidentical pages. The bigotry would be such but would be a publisher's, if anyone's. Accessibility is not disturbed by our being careful in editing Wikipedia for reliability; if our guideline and editing would have any effect on the publishing industry, it would be to encourage nonabridgement, which would be of benefit to readers who are visually impaired.
Large-print editions that have no regular-print counterparts are not at issue, because they would be reliable on their own or not, and already the guideline covers them quite well. That distinction probably should be clarified in the proposal; thank you for bringing it up. That an adult-level book has large print is not in itself the issue; it is whether a regular-print book that got serious editing for accuracy and then was abridged (if abridged) for large print to save on paper and distribution was given sufficiently careful editing for what was left after abridgement, given that editing is also expensive.
Light-weight facts can be believed because we know them from other sources, which can themselves be cited.
Nick Levinson (talk) 20:20, 19 May 2013 (UTC)
You don't seem to be operating with the same reliability rules as the rest of the community. Merely naming the publisher is enough to establish a presumption of reliability: For example, the community normally accepts, automatically and without any extra questions about reliability, any publication whose publisher is a university press, a major newspaper, or a publisher who specializes in reference works (e.g., anything published by published by Encyclopædia Britannica, Inc.). That's all you need to know, because the publisher does establish reliability. WhatamIdoing (talk) 20:39, 19 May 2013 (UTC)
Those are fine and I use some of them without checking reliability further, but suppose a book is published by Mystery789 Press (a name I just made up). One should try to find out if the author or the publsher is reliable. The name "Mystery789 Press" being on the cover, if we don't know anything else about the publisher, is not enough to establish reliability. Nick Levinson (talk) 21:38, 19 May 2013 (UTC)
That's true, and it's true regardless of the book's intended audience. However, this section shouldn't be used for an extended discussion. Please continue this discussion, if necessary, in the section below. Pburka (talk) 01:09, 20 May 2013 (UTC)
"Those are fine", but you've proposed that this guideline declare that Britannica Illustrated Science Library, a "source written for children" and published by a highly reputable publisher, is presumptively unreliable solely on the grounds that it is written for children. (If we really want to follow Nick's own suggestion that he not carry on conversations here, then this could all be moved to the next section. I'm not sure that it matters, though.) WhatamIdoing (talk) 02:40, 20 May 2013 (UTC)
Continued below. My suggestion with sectioning here was per RfC instructions. Nick Levinson (talk) 14:29, 20 May 2013 (UTC)
  • Oppose, the premise that educational books marketed toward children, or lower than high school level text books, are less reliable sources has no basis in verified fact. If anything I would hope that the publishers of such books (especially text books) have a more stringent editorial process that ensure that reliability of their content. Although children's books may not be the "best" source for all content, that doesn't mean that they should be excluded as a source when their individual reliability can be determined through the normal course of inquiry (such as review at WP:RSN).--RightCowLeftCoast (talk) 08:47, 21 May 2013 (UTC)
  • Oppose per above discussion, as well as everyone and their brother. No reliable evidence has been presented that children's books or large print books, should to be presumed unreliable for the abridged information that they do cover. --GRuban (talk) 17:32, 22 May 2013 (UTC)

Threaded discussion

  • Comment. This RfC is so poorly worded that it is hard to see what 'oppose' for example actually means. Oppose the use of children's books as sources? Oppose advising editors to look for a better source? Or what? And then there is the large-print books issue tacked on the end. How is anyone supposed to respond to that? It is another issue, and contributors may have entirely differing opinions on the two. I suggest that before this goes any further, the RfC is closed as malformed, and it is replaced with two clearly-worded RfCs. AndyTheGrump (talk) 22:52, 18 May 2013 (UTC)
It's about whether an editor should check beyond the source to be sure it's reliable if the source doesn't give enough evidence of reliability, or the editor need not check beyond because, for example, we can assume that children's books are almost always reliable for adult-level facts even if the source does not state evidence of reliability. A proposed text for the guideline immediately precedes the RfC subsection. Similar reliability issues apply to all three kinds of sources described in the proposed guideline text but respondents with different opinions may certainly state them and the proposal is not a yes-or-no proposition but can be rewritten in response to what is said about it. Apparently, for me to withdraw and re-do the RfC gets too complicated, so I clarified above, pursuant to the procedure in Wikipedia:Requests for comment#Suggestions for responding. Nick Levinson (talk) 16:32, 19 May 2013 (UTC)
Sorry, but you are still trying to conduct a single RfC on two different issues, and I don't see your 'clarification' as actually clarifying anything. If the RfC is over adding specific text, it should explicitly say so. AndyTheGrump (talk) 16:47, 19 May 2013 (UTC)
I don't interpret this proposal the way you described it. An editor should always ensure that the sources being used are reliable, regardless of the class of source. This proposal is about categorizing books aimed at young readers as presumptively unreliable. Pburka (talk) 16:42, 19 May 2013 (UTC)
We classify questionable sources in about the same way, too: editors should check before citing. While editors should check all kinds of sources, it appears that children's books are presumptively unreliable unless statements within them show reliability such as an author being described as a medical doctor in a children's book on the common cold. The proposal can be rewritten. Nick Levinson (talk) 17:10, 19 May 2013 (UTC)
So is this about "whether an editor should check beyond the source to be sure it's reliable", or whether "children's books are presumptively unreliable"? You're contradicting yourself. Pburka (talk) 01:21, 20 May 2013 (UTC)
No contradiction. When a source is presumptively unreliable, by checking we may find it is reliable after all. Nick Levinson (talk) 14:29, 20 May 2013 (UTC)

(Copied from the Survey section:) "'Those are fine', but you've proposed that this guideline declare that Britannica Illustrated Science Library, a 'source written for children' and published by a highly reputable publisher, is presumptively unreliable solely on the grounds that it is written for children..... WhatamIdoing (talk) 02:40, 20 May 2013 (UTC)" (Corrected quotation abridgement by me: Nick Levinson (talk) 14:37, 20 May 2013 (UTC))

Information that it is reliable would overcome the presumption. We should treat children's sources with such a presumption but it is not unrebuttable; that's the point of a presumption. The younger the likely audience, the more important the presumption when it comes to adult-level facts; a book for 6-year-olds is likely less reliable than a book for high school students.Nick Levinson (talk) 14:29, 20 May 2013 (UTC)
Why should we have this presumption? Why shouldn't we instead go with the normal presumption, which is that reference works published by reputable publishers are presumptively reliable, no matter what the type size is and no matter what the intended audience is? WhatamIdoing (talk) 20:31, 23 May 2013 (UTC)
That would be the same burden or more, because the editor wanting to cite a source would have to check its reputability respecting reliability. If they already know that reputation, that may suffice to rebut the presumption, as it suffices now without the presumption. Sources include nonreference nonfiction works and the reliability of a reference work does not extend to nonreference nonfiction works. It's generally agreed that information for people with much less education than adults have, which includes almost all young children old enough to read, has to be simpler than information for adults in order that it be understood; and the information provided to children is often inadequate for adult reliance. Audiences matter for reliability of sources. Type size by itself does not but all of the kinds of books encompassed in this proposal tend to use large type sizes and audiences and editing do matter. Nick Levinson (talk) 15:30, 24 May 2013 (UTC)
It's generally agreed[by whom?] that information for people with much less education than adults have, which includes almost all young children old enough to read, has to be simpler than information for adults in order that it be understood;[citation needed] and the information provided to children is often inadequate for adult reliance.[dubiousdiscuss] Pburka (talk) 22:14, 25 May 2013 (UTC)
My statements that you quoted and tagged are of the sky-is-blue variety, at least for the U.S. and most other nations in which education for any given children's and early adult age cohort is advanced or at least organized toward children growing up to be more knowledgeable than most of their forebears. Neither of us needs a list of names or citations to support them; they're too obvious. The general agreement is among educators of young children, parents of the same, child psychologists, and others. Indirect evidence of this agreement among educators, parents, and psychologists can be found in books published for and distributed to children: they're not adult-level books. If we tell a child to look a word up in a dictionary, we don't usually mean the Oxford English Dictionary. There's probably a young children's book on cancer but we wouldn't expect a physician to apply that book and nothing else to treating a cancer patient. If you are making a claim (I'm not sure you or anyone is) that most children can comprehend most adult-level books just as adults can and that most adults agree, that would be an extraordinary claim requiring evidence. Of course, there are exceptions, but we're not talking about them. Most children are not that precocious. Nick Levinson (talk) 15:17, 28 May 2013 (UTC)