Wikipedia talk:Reliable sources

To discuss reliability of specific sources, please go to Wikipedia:Reliable sources/Noticeboard.

Archives

WP:MOS subguideline, anyone?

Imo, this guideline could be far more usefully handled and improved if it where a dedicated WP:MOS subguideline. The applicable policy (WP:V) appropriately handles the required minimum threshold, while this page could explain various scenarious in greater detail than a policy, including the ideal case, or how to proceed in the many suboptimal cases where high quality sources are not easily available etcpp. User:Dorftrottel 14:52, February 15, 2008

What happens if a reliable source publishes an unreliable article?

This is particularly a problem with citations from newspapers, but what happens if a reliable source publishes an unreliable article? For example, if an article from a major newspaper quotes a census figure, but gets the number wrong? Is the "original research" of looking at the census data acceptable in this instance? Is it worth having a section on this problem in the project page? Even the most reliable of reliable newspapers have filler and wire articles where the standard of journalism is not as good as the rest of the paper. --Surturz (talk) 01:52, 22 August 2008 (UTC)

• I think it makes it a dubious source in that example. To give an example of a non-hypothetical instance, in the article on the Australian plebiscite, 1977 (National Song) the source for the plebiscite figures was clearly wrong with its numbers despite it being the Parliamentary Handbook, reprinted several times - see the annotation to the References . The citation wasn't removed altogether but qualified. (copied my response from Wikipedia_talk:No_original_research#newspaper_erroneously_quoting_poll ) --Matilda ^talk 01:56, 22 August 2008 (UTC)

First, we need to acknowlege that ANY source can contain an error... barring a deliberate misstatement of the facts, an error could be a simple typo, or due to faulty research on the part of the author. Reliable sources are those who keep such errors to a minimum. Discovering an error does not affect whether the source is reliable or not. However, discovering an error might affect whether the source should be used in support of a particular statement in a particular Wikipedia article. Such decisions are a matter of editorial judgement... and have to be made at the article level.

Second, in your situation both the newspaper and the Census are reliable sources... so we need to ask, which is more reliable of the two. We need to search for collaboration. What do other sources say? If there are no other sources for the fact... the census is the more reliable of the two. Blueboar (talk) 14:07, 22 August 2008 (UTC)

I think we need to look at multiple sources. We need smart editors who can identify mistakes and correct them. That's all we can do. AdjustShift (talk) 17:03, 16 September 2008 (UTC)

Reliable sources not in English?

Not sure if this is the right place to ask about this, but what is the guideline for citing sources that aren't in English? I can't imagine that we can use automated translations, right? But if a major newspaper in another country publishes something, there's no reason I see that such information wouldn't meet WP:RS. I also imagine that such sources could be used to demonstrate notability, correct? Oren0 (talk) 23:16, 23 August 2008 (UTC)

See: Wikipedia:Verifiability#Non-English sources for the policy on this. Blueboar (talk) 12:24, 24 August 2008 (UTC)

locatable?

A discussion of a type that sometimes comes up in technical (physics/mathematics) articles occurs when one person has access to a source that almost no libraries appears to carry. Is the fact that no or few libraries carries a book good evidence that it is a less reliable source? Pdbailey (talk) 14:29, 1 September 2008 (UTC)

No. Consider an antique book, with very few known copies that have survived through the centuries. Consider an extremely expensive book on an esoteric topic (How many libraries would spend a significant fraction of their budget to acquire something of limited general use?). Consider a book that had a deliberately limited publishing run. Consider a book whose printing has been suspended due to contract disputes between the author and the publisher. Consider any very recently published book. WhatamIdoing (talk) 21:42, 1 September 2008 (UTC)

WhatamIdoing, you appear to be arguing against it being proof, not "good evidence" as I asked. Certainly, many of these examples have good reason for low circulation. You didn't really answer my question, is it good evidence that the source is less reliable? Consider, i.e. a book that is not in the largest university library in a populace state. As far as a book on an esoteric topic that is expensive, you have to ask yourself, what was the publisher thinking, why would they publish such a book? It must be that the topic has a significant audience willing to pay for the book. Pdbailey (talk) 23:21, 1 September 2008 (UTC)

No, as I thought I indicated with five examples of perfectly legitimate reasons why any given book would not be available at, say, the University of Maryland library, -- to which I will now add a sixth example, the unlikeliness of a library buying a book in a foreign language, and a seventh, the mismanagement of budgets so that a public university library may be the least likely place to find an expensive reference work -- the fact that a book is not easily available at some libraries by itself is not good evidence of a book being unreliable. It is, in fact, almost as irrelevant a criteria as I can image.

Now if you could demonstrate that this particular book is widely unavailable for a reason that is relevant to its credibility, then that would certainly matter.

Ultimately, the question of how the ref is used matters more than the number of extant copies. (You could take your question to WP:RSN or to Wikipedia:WikiProject Librarians if you wanted more opinions.) WhatamIdoing (talk) 03:13, 2 September 2008 (UTC)

WhatamIdoing, I actually think the Johns Hopkins library might be the library with the most extensive University collection in the state of Maryland, but I'm not sure. Certainly, the national archives and library of congress dwarf the two. In any case, I would ask you to remove your slanderous claim against the UMD library if you can't reference it, it's just not classy. The situation comes up many times (i.e. if University of Chicago owns the only library copy of a book, is it really a great reference if nobody but a UC student/alum can actually check the facts?) Pdbailey (talk) 01:56, 4 September 2008 (UTC)

The number of copies of a book that exist has nothing to do with it's reliability as a source. What you seem to be discussing is the issue of ease of Verifiability. On that issue... As long as the source is able to be verified, it meets our requirements. Blueboar (talk) 03:35, 4 September 2008 (UTC)

Paul, if the Free State hasn't tried to cut university budgets in response to the economic problems, then it is the only state in the Union to skip that easy target. You and I both know that capital budgets, such as for library acquisitions, get squeezed in tough economic times. UMD has really never been comfortably funded in the first place, and it's worse right now. WhatamIdoing (talk) 06:28, 4 September 2008 (UTC)

(backdent) I think you two are being a little obstinate. The question I raise is this, "if this book is so great, why doesn't everyone own it?" If there are two text books on a topic conflict, one that every university library has in it's collection and one that even those with huge libraries (like Harvard) do no have, there are two problems. (1) only one editor will have access to the rare book, (2) the conflict must be resolved on which is more reliable. I as suggesting that the lack of interest library collection managers express in the book is great evidence on which is more reliable. Pdbailey (talk) 18:53, 4 September 2008 (UTC)

You are making two erronious assumptions... (1) As long as a book is either for sale to the public or can be found in a library that is open to the public (not nescessarily a "Public Library"), then more than one editor "has access" to it. It may not be easy or cheap for the average editor to gain access, but it is possible for them to do so. (2) Actually, the conflict does not need to be "resolved"... if two reliable sources disagree, then WP:NPOV tells us we should discuss the conflict in a neutral way, and mention both views. Blueboar (talk) 20:16, 4 September 2008 (UTC)

Why try to duke it out between these two sources? Why not find a third, fourth, or fifth? Surely four textbooks that agree on a specific point will trump any one. Any text could contain an error, no matter how easily available it is. WhatamIdoing (talk) 20:14, 4 September 2008 (UTC)

WhatamIdoing, excellent point. Thanks! Pdbailey (talk) 23:40, 5 September 2008 (UTC)

Special care on rumors

This is added in response to the clearinghouse for rumors that the Washington Post, CNN, and New York Times seem to have become in the case of Sarah Palin. I have made it as general as possible. In an ideal world, reliable sources would take the time to verify rumors rather than pass them along. patsw (talk) 01:07, 4 September 2008 (UTC)

I did some copyedits to this well-needed additional section. I would suggest, however, that some allowance be made for rumors which, while still totally unsubstantiated, have been mentioned in a reliable source but also mentioned as having had a real-life effect on something else notable. I am thinking of the John Edwards story, where the rumors were flying for some time, but at some point it was (reliably) reported that the rumors affected his asperations to his Vice-Presidential nomination, as well as his role in the party convention, without the content of the rumors having been verified at all (yet). Clearly the level of impact justifying mention can be debated on a case-by-case basis, but this is a nuance which probably should be developed. Baccyak4H (Yak!) 16:52, 4 September 2008 (UTC)

Update I made another edit to make the reasoning clearer as to this part of the guideline. In doing so, I admit that the caution has been weakened somewhat, and perhaps the wording should be tweaked again to strengthen it a little. But surely the guideline as a whole is better for this new section. Baccyak4H (Yak!) 18:22, 4 September 2008 (UTC)

Update 2 I made another couple of edits to clarify the language; now it means pretty much what I think it was intended to mean when I started, only now is (IMO) clearer. There is one ambiguity: The section asks for "special care" when reporting on (only) rumors, but then says we should include only verified material. There is a subtle contradiction there. My thought would be to loosen the only to elaborate exactly what "special care" means, similar to what I suggested in my first post in this section. But in the meantime, since either interpretation option of this section is reasonable, I encourage further discussion as to the best way to clarify this point. Baccyak4H (Yak!) 20:02, 4 September 2008 (UTC)

Per WP:NPOV if a major source reports a rumour, then it should be represented as the view of that source (and as a rumour, if that is what the source says). It is not up to us to pick and choose. 06:05, 5 September 2008 (UTC)

This isn't a clear cut, black and white issue. The question of whether to include mention of a rumor in our articles depends on many factors... what sources report on the rumor? (Is it being reported by a tabloid or a high quality news outlet?). How extensive is the coverage? (Is this being reported by one newspaper, or has it been picked up by several outlets?) What do these sources say about it? (Do they report it as fact, or simply as an allegation?) How do they discuss it? (passing remark? disparaging of the rumor?). If a rumor has been or is being discussed by lots of high quality sources, and they take it seriously, then we should indeed mention it (as a rumor)... if it is beng discussed by tabloids and scandal rags we should not. Remember that this issue is at the core of WP:BLP. If there is any doubt about mentioning a rumor, we should be conservative and not include it. Blueboar (talk) 12:54, 5 September 2008 (UTC)

I'm not very happy with the "rumors" text. Rumors may have news value, yes. But Wikipedia is WP:NOT a news organization. The issue of news reporting is mentioned several times in WP:NOT, using terms like "scandal mongering". I think that we need to make a stronger effort to support WP:NOT here -- or at least to point the editors to the several sections of WP:NOT that are applicable. WhatamIdoing (talk) 02:02, 6 September 2008 (UTC)

Kolubarska bitka/kolubara battle

Concerning all the non confirmed data and doubts about facts listed, please consult for all the relevant data in the book
 Zivojin Misic "My memories" (unfinished because of Z.M. death 1921) and inclosed text by army historian Lt.colonel Savo Skoko.

In this book the list of all relevant historical documents from most the world war historian library's are available to support

every word in this Wikipedia page concerning Kolubara battle.
Moreover , the details of this battle  could be find as analytic
subject (as a sample of unique strategy) on many  World Army Academies's.  —Preceding unsigned comment added by Deleor (talk • contribs) 13:47, 7 September 2008 (UTC) 

Blogs, wikis and webforums

I've often seen it said (and I often say it myself) that blogs, wikis and webforums are not usable as reliable sources, but on coming to this page I find that these sources aren't even mentioned. Is there a page that specifies what internet material is suitable? I'd like to have backup for the debates I sometimes get into on this (eg here and here. If there isn't a page or section I'll write it you like. Totnesmartin (talk) 11:57, 12 September 2008 (UTC)

The initial paragraph of WP:V#Self-published sources, and the associated footnote. As with all things Wikipedia, this may change over time. -- Boracay Bill (talk) 21:16, 12 September 2008 (UTC)

Ah, that's great, now I know where to find it, cheers! 12:33, 13 September 2008 (UTC) Totnesmartin (talk)

Needs list of sample authoritative sources for medical articles

...that way, I can insert cites quickly without having to first determine what is an authoritative source. Medline is assumed to be okay by it's very nature. But e.g., what about an article on webmd? What about a university site? Yes, at first some good sources wouldn't be mentioned but over time a valid list would be built up - always a work in progress. My case in point: while reading up on Amicar, I came across a med animation at youtube. I see it as an illustration, rather than a reference, but the whole issue of what sources can be used is not obvious and I'd rather not spend a great deal of time in learning the ins and outs.

Or maybe what I'm talking about exists somewhere else here on wikipedia? If so, please point me there. Thanks. Kenmcl2 (talk) 01:17, 14 September 2008 (UTC)

Please see Wikipedia:Reliable sources (medicine-related articles). Eubulides (talk) 04:08, 14 September 2008 (UTC)

That link is useful for me also! AdjustShift (talk) 13:44, 16 September 2008 (UTC)

JJ's restaurant

I have owned JJ's restaurant since 1991 & I dont remember Paddy Mcguiness working for me. Andy Thomas —Preceding unsigned comment added by 86.134.20.226 (talk) 10:37, 14 September 2008 (UTC)

I've copied this Talk:Patrick McGuinness. -- Rick Block (talk) 16:48, 14 September 2008 (UTC)

Given the above comment, I removed the mention of his working at JJ's as it is unsourced.Blueboar (talk) 17:13, 14 September 2008 (UTC)

Good job! False information should be erased as soon as possible. AdjustShift (talk) 16:55, 16 September 2008 (UTC)

Are court documents reliable sources?

On a Talk Page discussion, an editor has put foward the opinion that court documents are not acceptable sources for wikipedia. I looked through some policies and couldn't find any information referencing this. Is there anyone with experience who'd like to take a look at the discussion? McJeff (talk) 04:00, 15 September 2008 (UTC)

It depends on the circumstances. Court documents are typically primary sources, and as per WP:RS#Primary, secondary, and tertiary sources it's generally better to cite secondary sources. For specific facts, such as the date the judge ruled and so forth, court documents are reliable; but for other things (e.g., if a document contains an argument by a plaintiff, and if Wikipedia cites that argument in support of a claim) they are not particularly reliable. Looking at Tucker Max #Legal troubles, this particular case, it seems clear to me that we have a case of WP:OR; the first paragraph calls on primary sources to advance a theory that the primary sources themselves do not directly support. I'll follow up on the talk page and tag the section. Eubulides (talk) 04:15, 15 September 2008 (UTC)

It is always better to cite secondary sources. That's the best way to stay out of trouble. AdjustShift (talk) 13:41, 16 September 2008 (UTC)

It is important to note what?

This edit inserted the following text into WP:RS #Self-published sources:

It is important to note that some non self-published sources of the origins listed above can be used, if they do not break any other policy/guideline, with exceptions.

I'm sorry, but I don't know what that text means. It's trying to say something, clearly, but I can't make heads or tails of it. For now, I undid the change; perhaps someone can explain what's going on? Eubulides (talk) 06:09, 17 September 2008 (UTC)

Basically the cite note on WP:V on self published sources that people seem to gloss over, miss, and indubitably misrepresent the policy, I've seen it quite often, and WP:V says that its a policy and WP:RS should be updated accordingly, so basically adding it, sorry, not good at wording. ηoian ‡orever ηew ‡rontiers 02:56, 18 September 2008 (UTC)

About popular press and scientific facts

Ludwigs2 has repeatedly deleted text added eleven days ago by Eubulides that reorganizes existing information about news organizations and clarifies its applicability in several contexts.

In particular, it adds a link to WP:MEDRS, which is the crux of the problem, because Ludwigs2 is one of two people (the other a professional journalist) that objects to MEDRS's preference for citing scientific sources, such as peer-reviewed scientific articles and textbooks, instead of popular press reports. This has been discussed at great length at WT:MEDRS, but fundamentally the reasons are these:

• Journalists make mistakes. It's beyond foolish to assume that every newspaper and magazine article gets every single fact correct, the first time, and every time. On the common sense scale, this is not very far different from saying "Please check the original document instead of relying blindly someone's notes about the original document."
• Journalists leave out critical information. Most news articles fail to discuss important issues such as evidence quality, costs, and risks versus benefits.[1]
• Some journalists simply don't have the necessary background. Around my house, we refer to some particularly dreadful newspaper stories as having been written "by someone that majored in journalism because science was too hard." This results in things like every investigational new drug as the "discovery of the cure" of a disease. They also tend to express risk in useless terms, leading people to believe, for example, that breastfeeding prevents childhood leukemia (and it might: but you'd have to exclusively breastfeed four million babies for a year to prevent a single case of leukemia).

For these and other reasons, the editors at WP:WikiProject Medicine and WP:MEDRS have included a section in MEDRS that explains the relative strengths and weaknesses of the popular press. Overall, it prefers citing scientific sources for scientific facts (and non-scientific sources for non-scientific facts). This recommendation was appropriately and briefly summarized here as follows:

For medical and scientific facts and figures, it is typically better to cite the scholarly research behind a newspaper story, rather than simply citing the story itself. Newspapers tend to overemphasize the certainty of results, and often fail to adequately report methodology, error, risks, and costs associated with a new scientific result or medical treatment.

Ludwigs2 has repeatedly deleted this information from this guideline today, and despite his assertion at WT:Consensus (where he discovered this addition) that he thought it should be discussed, I find no discussion here. So here's the explanation; are there any objections by anyone other than Ludwigs2? WhatamIdoing (talk) 01:05, 18 September 2008 (UTC)

No objection here of course. The change reflects consensus at WP:MEDRS, improves the structure of this page, and causes this page to reflect WP:MEDRS's contents more accurately. Eubulides (talk) 02:03, 18 September 2008 (UTC)

• Now that I'm on the same page, I still have the same misgivings about focusing on a single area when the problem is universal to academic topics. I agree with the other bullet points. I made another draft that I feel serves the intended purpose (User:Vassyana/news). Thoughts? Vassyana (talk) 07:33, 18 September 2008 (UTC)

• Again, a nice job of rewriting. Some minor comments:

• "news media often fails" should be "the news media often fail", since "media" is a plural noun. Or, if that grates, say "news reports often fail".
• It'd be nice to mention medicine as an example here, as well. One way would be to replace "information about academic topics" with "information about academic topics such as medicine". This would help to avoid giving the implication that scholarly sources are preferred only when talking about pointy-headed topics like philosophy.

Eubulides (talk) 08:22, 18 September 2008 (UTC)

Revised per suggestions. I went with "physics or ancient history" for the examples, for the sake of breadth and variety. We want to be clear that this covers the whole spectrum of academic subjects, both "hard" and "soft". Medicine is used as the example in the other draft and medical treatment is mentioned in the closing statement. Physics and ancient history are both examples where fringe and pop understandings of the subject can be problematic. Any other tweaks you think it needs? Vassyana (talk) 10:33, 18 September 2008 (UTC)

• I would endorse this, it is an improvement over what we have. 16:47, 18 September 2008 (UTC)
• I also endorse. I like the "ancient history" (Wikipedia circa 2004, say?....). Eubulides (talk) 17:41, 18 September 2008 (UTC)

I have restored the deleted text, and hope that we can focus on Vassyana's #Scholarship proposal now. WhatamIdoing (talk) 05:21, 19 September 2008 (UTC)

About Scholarship

• Reading the summary version above, I was deeply wary of the change. Too much emphasis on individual studies (which can be often difficult to properly interpret and place in context, even for experts) was my primary concern. However, after checking the suggested version, I'd agree this is the proper direction. However, I do have a couple of misgivings. Textbooks and academic press books should be mentioned alongside review and meta-analysis literature. They often provide excellent overviews of the accepted mainstream in scholarly fields. Also, there seems to be a bit of undue emphasis on science, which comprises only a portion of scholarly study (though certainly a significant one). The changes should take a more neutral position addressing the whole of scholarship, rather than being so focused on "hard" science. It could also stand to be more concise, but I have no problem addressing that at a later point. Vassyana (talk) 02:34, 18 September 2008 (UTC)

• I've made a sandbox revision to try addressing my concerns (User:Vassyana/scholar). Thoughts? Vassyana (talk) 03:03, 18 September 2008 (UTC)

User:Vassyana/scholar is a substantial expansion, and should probably be discussed separately. All I care about is stating a preference for scholarly works (when reasonably available, WP:IAR, etc.) "For medical and scientific facts and figures". Do you have any objection to this? WhatamIdoing (talk) 03:52, 18 September 2008 (UTC)

There must be some confusion here. User:Vassyana/scholar is a rewrite of WP:RS #Scholarship, but the changes Ludwigs2 reverted are to a different section, WP:RS #News organizations. The two sets of changes are pretty much unrelated, as far as I can see, so I inserted an #About Scholarship subhead above Vassyana's comments. Eubulides (talk) 05:23, 18 September 2008 (UTC)

Pardon me while I wipe the egg off my face. :-) I did confuse the sections, mainly due to the subject of the discussion and a lack of care (bad Vassyana!) in checking the diffs. Thanks for taking the proposed changes seriously nonetheless. Vassyana (talk) 07:33, 18 September 2008 (UTC)

Here are some comments on Vassyana's proposed changes to WP:RS #Scholarship:

• User:Vassyana/scholar is a very nice rewrite overall. It prunes away redundant text and it makes the section easier to read. Most of the comments below are about minor points.
• The phrase "cover academic topics and" is redundant and should be removed. That phrase might tempt editors into thinking that for many topics, academic scholarship is irrelevant and should not be searched for.
• "However, some material" should be "However, some scholarly material" to avoid confusion.
• "cover all significant published views" should be "cover all significant views published by reliable sources"
• Please don't add "philosophy" as an example field where single studies are less definitive. In some parts of philosophy, e.g., formal logic, single studies can be quite definitive (more so than in most fields). The current text's use of medicine is better as an example, and one example suffices here.
• "Care should be taken to avoid" should be "Avoid".
• Omit "academic press books" from the list of preferable sources. They aren't in the same league as meta-analyses, textbooks, and reviews. For example, Targeting Autism (ISBN 978-0-520-24838-0) is a fine book from a respected academic press about autism treatments, but Autism therapies is right to prefer Myers et al. 2007 (PMID 17967921), a reliable review: it's much more authoritative and is more carefully reviewed.
• A nit: "Meta-analysis" should be "Meta-analyses", right?
• The phrase "review literature" is a bit ambiguous; are we talking about the New York Review of Books? Perhaps "scholarly review articles" instead?
• Wikilink to Meta-analysis.

Eubulides (talk) 05:23, 18 September 2008 (UTC)

Good points. Draft altered accordingly. Vassyana (talk) 07:33, 18 September 2008 (UTC)

Thanks, the latest draft looks good. Eubulides (talk) 08:22, 18 September 2008 (UTC)

No problem. Thank you for the constructive feedback. Vassyana (talk) 10:33, 18 September 2008 (UTC)

• Given the points made in the previous discussion section, I see a potential problem with

  Reliable non-academic sources may also be used, particularly material from reputable mainstream publications. Wikipedia articles should cover all significant views published by reliable sources, doing so in proportion to their published prominence. The choice of appropriate sources depends on context and information should be clearly attributed where there are conflicting sources.

• As was discussed above, an announcement such as "Alzheimer's breakthrough" may be far more widely published than the actual research paper(s), which may make a far less confident claim. So could we think about whether the "in proportion" could be improved? 17:11, 18 September 2008 (UTC)

• Maybe reword it as "Wikipedia articles should cover all significant views, doing so in proportion to their published prominence among reliable sources."? That is, it's prominence among the reliable sources that counts, not overall prominence. Eubulides (talk) 17:41, 18 September 2008 (UTC)

The potential problem with that is, we have classified mainstream newspapers as reliable sources as well. 18:17, 18 September 2008 (UTC)

This is why MEDRS encourages people to cite the actual paper instead of the newspaper's version of the paper. I think we can consider these two proposals separately: if #News orgs points out (per the proposal above) that newspapers aren't necessarily the single best source for medical facts, then #Scholarship doesn't have to be as restrictive about the general case. WhatamIdoing (talk) 18:22, 18 September 2008 (UTC)

• That is fine for medicine, but I think any scientist would be able to offer examples from his or her field. This report on using salt water as a combustible fuel may be one such example; while I am not an expert on the physics of water, and you should take what I say with a grain of salt (pun intended), I have heard it argued that these reports failed to mention that the amount of energy that needs to be input to make salt water burn far exceeds that released by the combustion. I fear that generally, items are sometimes deemed newsworthy even if they have only a tenuous grounding in scientific fact, whether it's medicine or any other field. 22:15, 18 September 2008 (UTC)

• OK, how about this rewording instead? ""Wikipedia articles should cover all significant views, doing so in proportion to their published prominence among the most reliable sources." The new words here are "the most". Eubulides (talk) 22:20, 18 September 2008 (UTC)

• I think that could work, well done! 22:37, 18 September 2008 (UTC)

Revised accordingly.[2] Vassyana (talk) 23:10, 18 September 2008 (UTC)

In general, I like what Vassyana has written. The last bullet currently reads:

* Single studies are usually considered tentative and may change in the light of further academic research. The reliability of a single study depends on the field. Studies relating to complex and abstruse fields, such as medicine, are less definitive. Avoid undue weight when using single studies in such fields. Meta-analyses, textbooks, and scholarly review articles are preferred to provide proper context, where available.

and I'm inclined to suggest a few minor changes:

• It's not "single studies" (a systematic review can be considered a single "study", since it hopes to disprove a hypothesis) so much as it's "unreplicated experimental results" that are at issue.
• It might be worth naming both narrative reviews and systematic reviews along with meta-analyses.

Neither of these changes are truly important to me; they're just suggestions. Thanks for your work on this. WhatamIdoing (talk) 05:21, 19 September 2008 (UTC)

What about replacing "single studies" with "isolated studies"? Would that accurately convey the point? Vassyana (talk) 07:10, 19 September 2008 (UTC)

Works for me. (Thanks for your corrections today.) WhatamIdoing (talk) 17:48, 19 September 2008 (UTC)

The page later uses the more-standard term "primary sources"; how about if we use that here, with a wikilink, and then mention a single, isolated study as an example? Something like this:

A primary source, such as a paper reporting experimental results, is usually considered tentative and may change in the light of further academic research. The reliability of a primary source depends on the field. Primary sources relating to complex and abstruse fields, such as medicine, are less definitive. Avoid undue weight when using primary sources in such fields. Meta-analyses, textbooks, and scholarly review articles are preferred to provide proper context, where available.

Eubulides (talk) 17:51, 19 September 2008 (UTC)

Unfortunately, there's a lot of baggage with the term "primary source" that I would rather not drag here. For example, a common argument has been that only the lab notes are primary sources, while the study papers are secondary. I would prefer to avoid opening that can of worms here. Vassyana (talk) 21:18, 19 September 2008 (UTC)

Yes, I think we're better off avoiding that term here. WhatamIdoing (talk) 04:12, 20 September 2008 (UTC)

General question

Feel free to respond here or on the page the question came up, WT:WIAGA, or not. There's a question over whether this sentence should be re-inserted into criteria for nominating Good Articles: "It is generally acceptable for good articles to contain a small percentage of sources with borderline reliability; however, most sources should be reliable". I'm uncomfortable with trying to define a class of "not reliable, but not that bad" (IMDB was suggested) sources that might sometimes be acceptable in Good Articles; what do you guys think? - Dan Dank55 (send/receive) 15:50, 19 September 2008 (UTC)

I think it really depends on how the sources are used. When I think of "not-the-best sources" (my preferred wording) I think of self-publications by experts, or primary sources. I would never use such a source for information that doesn't seem in line with the reliable sources present, but I would use it to provide more context for something from an RS. Thus, I think editors have gone over the line if removal of all not-the-best sources leave not just little details, but a significant fact/claim unreferenced. Someguy1221 (talk) 16:25, 19 September 2008 (UTC)

One of these days we will actually put the RSN mantra in WP:RS: "No source is universally reliable. Whether or not a source is reliable depends on how you use it."

I think it's acceptable for GAs to have some weak sources, particularly for non-controversial statements. It might be better to cite a textbook for some widely accepted fact, or a newspaper for a political event, but when the fact in question is not actually being challenged, it's probably okay to support it with, say, a website run by a small organization, or someone's lecture notes. Complete perfection of all sources is perhaps best reserved for FAs. WhatamIdoing (talk) 17:49, 19 September 2008 (UTC)

That resonates with my view exactly. Horses for courses. --Malleus Fatuorum (talk) 18:13, 19 September 2008 (UTC)

WP:NOT#NEWS vs.WP:RS#News_organizations

Having read the prohibitions regarding the use of trivial sources in WP:NOT#NEWS, I propose that we change the wording of WP:RS#News_organizations so that it addresses these issues. Whilst I am not proposing that WP:RS#News_organizations should repeat these concerns verbatim, at the very least I think there should be some sort of link to WP:NOT#NEWS. If you compare and contrast these two sections, would anyone agree that they need to reflect each other more closely?--Gavin Collins (talk) 14:29, 21 September 2008 (UTC)

They relate to completely different things. WP:RS#News organizations refers to the reliability of using a newspaper as a source. WP:NOT#NEWS refers to writing articles about recent news events. Apples and oranges. Blueboar (talk) 15:38, 21 September 2008 (UTC)

• Since the content of articles about recent news events is likely to be drawn from newspapers and similar sources, I would have thought they were dependent on one another. --Gavin Collins (talk) 17:31, 21 September 2008 (UTC)

I don't think they are dependant on each other... in the case of an article on a recent event, first comes the decision as to whether to create an article or not... that is where NOT#NEWS plays a role. If the decision is to create the article, then we have to write it, and use reliable sources to support what we write. That is where RS#News organizations might enter the picture. They apply to different parts of the article writing process. Also, RS#News organizations discusses using news media as a citation in any article, not just those on recent events. Blueboar (talk) 19:42, 21 September 2008 (UTC)

I agree with Blueboar. Any apparent problem is a matter of vagueness in N and NOTNEWS. A news article isn't rendered unreliable by reporting something new. Someguy1221 (talk) 07:26, 22 September 2008 (UTC)

• However, a news article is rendered unreliable if it is "Routine news coverage of such things as announcements, sports, and tabloid journalism" per WP:NOT#NEWS. --Gavin Collins (talk) 10:00, 23 September 2008 (UTC)

You are conflating distinct concepts. "Not appropriate for Wikipedia" is not the same as "unreliable". You have noted in notability discussions (no pun) that just because we can verifiably and reliably relate some information doesn't mean we should. To take two further examples from WP:NOT, WP:NOTDICDEF and WP:NOTGUIDE both exclude content that can be reliably sourced. Vassyana (talk) 14:16, 23 September 2008 (UTC)

Vassyana is right. Many reliable sources present information that is not appropriate for Wikipedia for some other reason (e.g., not being encyclopedic). WhatamIdoing (talk) 00:56, 24 September 2008 (UTC)

Third party

I cannot find where the term 'third party' is directly explained. It links to Wikipedia:No_original_research#Primary.2C_secondary.2C_and_tertiary_sources, which doesn't directly explain it either. --Matt Lewis (talk) 09:44, 22 September 2008 (UTC)

Try Wiktionary: wikt:third_party. The second definition is the relevant one. Note please that "third party" is used loosely and is not meant to imply that all disputes/other situations always have exactly two sides that are always represented by exactly one party each. WhatamIdoing (talk) 18:55, 22 September 2008 (UTC)