Wikipedia talk:Identifying and using primary sources/Archive 1

Archive 1

Things to think about

I do like the idea of noting that "secondary is not always good"... but I think it more important to note that "primary is not always bad". Blueboar (talk) 19:39, 10 May 2011 (UTC)

I have added something to that effect. Blueboar (talk) 19:58, 10 May 2011 (UTC)

What excellent work! One note: I'd recommend avoiding getting into the article deletion stuff. For starters, it's a different subject. Sincerely, North8000 (talk) 10:50, 12 May 2011 (UTC)

If by "article deletion stuff" you mean "don't discuss how primary/secondary sources impact WP:Notability", I disagree. I think that is something this essay should discuss. To me, the focus/topic of this essay should be inclusive... "Primary/secondary sources - what they are used for and how to use them appropriately - also how they can be misused and how to avoid doing so". Blueboar (talk) 14:12, 12 May 2011 (UTC)
I think that we need to address secondary-for-notability directly, because the odd, non-standard definition in play at AFD is where most editors get their first training on what a secondary source is. Then they wander out to the rest of the encyclopedia and get completely confused and often upset when someone points out that yesterday's 'eyewitness news' story actually isn't a secondary source. WhatamIdoing (talk) 04:35, 24 May 2011 (UTC)

Notability

The major use of secondary sources in Wikipedia is to show something has been noticed. As it is currently written a yellow pages directory of companies would be considered a secondary source, do we really want to consider that as source of notability? I think a lot more has to be made of the noticing and evaluating function of a secondary source. Dmcq (talk) 12:00, 30 May 2011 (UTC)

I think that the yellow pages would be considered a tertiary source, rather than a secondary source.
Also, the yellow pages are paid advertisements. As such, they are entirely non-WP:Independent sources and completely worthless for showing notability. WhatamIdoing (talk) 05:11, 21 August 2011 (UTC)

The yellow pages is not a secondary or a tertiary source. A secondary sources builds on the material from primary sources. The secondary source material must be somehow transformed primary source material, and is consequently a product of the primary source material (the facts) and the creativity of the author of the secondary source. In a simple directory, there is no creative content deriving from the "author" of the directory. A logical ordering or sorting of data is not creative transformation. --SmokeyJoe (talk) 13:36, 23 September 2011 (UTC)

On a second look, I can't find any independent assertion (i.e., not directly or indirectly from the English Wikipedia) that claims a telephone directory is a tertiary source. I did find one that defined two other interesting axes: factual vs analytical sources and objective vs subjective sources. WhatamIdoing (talk) 17:17, 30 September 2011 (UTC)

Initial thoughts

I think there is a lot of related material found at Wikipedia_talk:Primary_Secondary_and_Tertiary_Sources. I personally recall writing "I’d suggest instead writing a new guideline “Interpreting WP:PSTS”"

On the section "Characteristics of a secondary source"...

RE: "A secondary source is based on primary sources" "is based on" is used a bit too much. A secondary source is not so much based on primary sources as it is a product of the mind of the author. A secondary source makes reference to at least one primary source, even if implicitly, sometimes using the assumption that the reader is familiar with the facts, or may even assume that the reader is a personal witness to an unnamed event making primary sources virtually untraceable. It is the commentary, analysis, criticism, etc, that defines a secondary source.

Secondary sources need not necessarily be significantly separated in time or space (although they usually are).

I find it easy and useful to tie the word "story" to secondary sources, and and "report" to primary sources. This works quite well with newspapers and other newsmedia. If it tells a story, the must be some transformation of the basic information. (Of course, if the subject is the story itself, then the news story is a primary source)

"A secondary source is usually based on more than one primary source" is debateable. How do you count secondary sources to establish "usualy"? Private commentary, such as your housemate talking to you, happens a lot and may use a single primary source. "Good" or "useful" secondary sources are usually based on more than one primary source. Reputably published secondary source material is usually based on more than one primary source.

On the section "Secondary sources for notability" RE: "One rough rule of thumb for identifying primary sources is this: if the source is noticeably closer to the event than you are, then it's a primary source." I don't think that is so correct. While usually true, it is often not. A discussion of the event of January 1, 1800, published the next day, was, and is, a secondary source with regards to the event. It may not be the most useful secondary source, but if it presents opinion and analysis and does not repeat the facts verbatum, it is a secondary source. The reason that it is likely treated as a primary source is because today's story probably goes beyond discussion of the event to include discussion of local reaction at the time. It thus comes down to "usage"


It may be best to ignore non-historiographical usages, but just to note that they exist: In the sciences, there is often a single capital P Primary source. It will be the first, the original, the most authorative source for some highly specific thing. It may well be the first account of an eyewitness. It may be the the technition's labbook. It may be the paper that first speculates the explanatory theory. This very narrow definition of a single Primary source seems to encourage people to think that all other sources are secondary. However, in the sciences, there is no concept of a "secondary source", and "Primary" simply means "first".

The above concept of a single Primary source applies readily to eyewitness accounts, and seems to be a familiar usage in journalism. In journalism, a "secondary account" is an account of what someone else said. The information is therefore "second hand". It is also less realiable, but it has some relationship to notability in that if something is worth repeating, then it is more notable than something not worth repeating. This use of secondary seems completely independent of the historiographical "secondary source".

--SmokeyJoe (talk) 15:10, 16 August 2011 (UTC)

Thank you for your useful comments. I've made some changes already based on them.
Since Wikipedia has decided to use PSTS for everything, I think we're stuck with trying to shoehorn the non-history fields into this framework. It's always going to be inelegant, but at least we'll can tell editors what Wikipedia they can realistically expect when they are working on (e.g.) medicine-related articles. WhatamIdoing (talk) 19:14, 17 August 2011 (UTC)

Secondary sources for notability

  • One rough rule of thumb for identifying primary sources is this: if the source is noticeably closer to the event than you are, then it's a primary source. For example, if an event occurred on January 1, 1800, and a newspaper article appeared about it the next day, then Wikipedia (and all historians) considers the newspaper article a primary source.
  • Typically, very recent newspaper articles are mis-labeled as a "secondary source" during AFDs, by way of trying to finesse the general notability guideline's requirement that secondary sources exist, when no true secondary sources actually exist.

An 1800 newspaper is an extreme example. I don't think that we'd consider a 1985 newspaper article or book to be a primary source. I'm not sure that calling all newspaper articles primary sources is accurate either.   Will Beback  talk  00:16, 17 September 2011 (UTC)

I gave the extreme example on purpose, because it's a clear-cut example. An article from 20 years ago might go either way. A rough rule of thumb is not guaranteed to work well for borderline cases.
All newspaper articles are certainly not primary sources, and this page makes no such claim. However, all "eyewitness news" reports are primary sources (by definition), and basically all the newspaper articles that appear within hours of an event are primary sources (for purely practical reasons). WhatamIdoing (talk) 18:36, 17 September 2011 (UTC)
That's not true. If a news report from hours after a political debate includes analysis from experts then that's clearly a secondary source. I think a lot of the text in this section is misleading and doesn't address the topic of the section.   Will Beback  talk  22:21, 17 September 2011 (UTC)
I think that such a story is one that even experts would disagree on how to classify. The existence of some bit of analysis does not automagically make the source into a secondary one. If you presented a historian with a very old newspaper article, such as one describing a debate in the Continental Congress, s/he might very well tell you are calling "anaylsis from experts" was actually individual opinions issued by insiders rather than proper analysis, and that it is certainly a primary source for the initial reaction by contemporary experts. (Or s/he might not: it would depend on the details of the source and the historian's professional view of source classification.)
Even such a piece, however, is likely to be largely a primary source, because the first reports about a political debate have to provide basic descriptions, like who said what. All of that is unquestionably primary source material. The existence of a few sentences of on-the-spot analysis does not transform the whole piece into a secondary source. WhatamIdoing (talk) 02:05, 18 September 2011 (UTC)
I think we're basing too much here on our own speculation. I'm going to trim some of the assertions out of that section and focus it more on the simple issue of how secondary sources are importawnt for establishing notability.   Will Beback  talk  03:23, 18 September 2011 (UTC)
I'm not basing any of this on speculation. This is what the sources actually say. See, for example, "Primary sources include original manuscripts, periodical articles reporting original research or thought, diaries, memoirs, letters, journals, photographs, drawings, posters, film footage, sheet music, songs, interviews, government documents, public records, eyewitness accounts, newspaper clippings, etc."[1] (emphasis added)
Someone's original thoughts about a political debate, as reported in a newspaper, falls into the standard definition of a primary source. WhatamIdoing (talk) 03:27, 18 September 2011 (UTC)

Here are some more sources that say the same thing:

  • "Primary sources are original records created at the time historical events occurred or well after events in the form of memoirs and oral histories. Primary sources may include letters, manuscripts, diaries, journals, newspapers, speeches, interviews, memoirs, documents produced by government agencies such as Congress or the Office of the President, photographs, audio recordings, moving pictures or video recordings, research data, and objects or artifacts such as works of art or ancient roads, buildings, tools, and weapons."[2]
  • "Primary sources: These are contemporary accounts of an event, written by someone who experienced or witnessed the event in question. These original documents (i.e., they are not about another document or account) ... may also include published pieces such as newspaper or magazine articles (as long as they are written soon after the fact and not as historical accounts)" [3]
  • "Periodicals - magazines, journals, and newspapers - written during the time period under study are excellent primary sources."[4]
  • "Primary sources are original sources created at the time a historical event occurs"[5]; "newspapers" are listed as an example.
  • "Primary sources: Published materials (books, magazine and journal articles, newspaper articles) written at the time about a particular event. While these are sometimes accounts by participants, in most cases they are written by journalists or other observers. The important thing is to distinguish between material written at the time of an event as a kind of report, and material written much later, as historical analysis."[6]
  • "Primary sources are the historical documents used by historians as evidence. Examples of primary sources include diaries, personal journals, government records, court records, property records, newspaper articles, military reports, military rosters, and many other things…. The key to determining whether an item may be considered to be a primary source is to ask how soon after the event was the information recorded."[7]
  • This names "news reports" twice as examples of primary sources in its list of what's considered a primary source by each discipline.

All of these are academic institutions. All of them agree: A newspaper story written immediately after a historical event is a primary source. WhatamIdoing (talk) 03:46, 18 September 2011 (UTC)

The question of whether particular sources, like newspaper articles, are primary or secondary is separate from the issue of the need for secondary sources for notability purposes. So I've split the section. Also, I'm concerned that you keep adding assertions implying bad faith on the part of people who are not historians. That's really not helpful to a would-be guideline like this.
Getting back to the issue of whether contemporary newspapers and magazines are primary or secondary sources, we have to remember that Wikipedia's standards and definitions are not necessarily the same followed by academic historians. Guidelines and even policies are descriptive of how Wikipedia actually works, not how it should work. I think it'd be hard to show that Wikipedia editors routinely regard newspaper articles as primary sources. We shouldn't use this guideline to try to change Wikipedia behavior, just to help editors follow community standards.   Will Beback  talk  04:18, 18 September 2011 (UTC)
That's exactly why these belong in the same section: We have a weird definition that makes old newspaper articles be routinely (and correctly, according to the academic standard) considered primary sources, but brand-new ones (except sometimes breaking news and eyewitness reports) be considered secondary sources—but only for notability purposes. If you're in a dispute over what's DUE, most types of recent newspaper reports (but not most types of magazine articles and not certain special types of newspaper articles, like stories on the 50 year anniversary of WWII) are going to be handled as primary sources.
We need to tell people what the academic standard is (thus the example from the 1800s), and how notability differs from it for recent events (thus the description of the "couple of years" grace period). Both of these are about notability, not just about newspapers. WhatamIdoing (talk) 23:02, 19 September 2011 (UTC)
The implications of your proposed definition (for Wikipedia purposes) are far reaching. NOR says that articles should be based on secondary or tertiary sources. If we define newspapers as primary sources then thousands of articles on contemporary topics would become non-compliant, including scores of Featured Articles, such as 2010 Lamar Hunt U.S. Open Cup Final, Barack Obama, Richard Cordray, South Australian state election, 2006, Déjà Vu (Beyoncé Knowles song), Tropical Storm Nicole (2010), Brad Pitt, 300 (film), 200 (Stargate SG-1), J. K. Rowling, and Assassination of Robert F. Kennedy. In some cases, this proposal would mean that those featured articles should be deleted outright. As such, it is at odds with usual practice on Wikipedia, which treats newspaper and news magazine reports as secondary sources.   Will Beback  talk  00:17, 20 September 2011 (UTC)
If you're talking about FAs from back when some of our policy geeks thought that "secondary" was a fancy way of spelling "independent", then you might be right that they could use some work.
But it's not all newspaper articles that are primary, and many of those articles cite non-primary media. For example, the LA Times report titled ""Occidental recalls 'Barry' Obama'" is a secondary source for the name of the college Obama attended, because it's writing about events that happened nearly three decades before. Three decades is significant separation.
This section is dealing with the notability problems of recent events: events that happen on Monday afternoon and turn up in Tuesday morning's newspaper. Those articles are (almost) always primary sources, and those articles are not considered proof of notability in the end. WhatamIdoing (talk) 21:18, 21 September 2011 (UTC)
Tropical Storm Nicole (2010) was passed as an FA in July 2011. Has the policy been changed since then? Please take a moment to review that article and see if it is based mainly on secondary sources, according to your definition of the term.
I don't doubt that there are any number of terms used on Wikipedia that have different definitions in academia. But this isn't academia. We create our own manual of style and we have our own functional definitions of "primary" and "secondary" too. It's the job of this essay to describe Wikipedia practice, not to change it to conform to academic practices. (That may be a valid goal, but this essay isn't the place to do it unless it's recast for that purpose).   Will Beback  talk  22:41, 21 September 2011 (UTC)

Yes, academic usage differs from what we do. I think the key issue is what the newspaper is reporting. A press release, in a newspaper, is a primary SPS. A news report featuring an interview with a reporter about what they saw would be a primary source. But I would argue that a news report filed by a reporter investigating a criminal matter would be secondary, regardless of how soon after the crime it occurred, since the reporter is not reporting what they say, but what the police and witnesses said. Separation need not be temporal. --Nuujinn (talk) 00:01, 22 September 2011 (UTC)

Technically, nothing published as an article in a newspaper is self-published, even if the article is word-for-word the same as the self-published press release. What makes something self-published is if the author and the publisher are the same entity. "Big Marketing Company, Inc" is not the same as "Smallville Times' publisher". Therefore the newspaper's publication constitutes proper publication—even if the source is still essentially lousy for Wikipedia's purposes.
I agree that separation need not be temporal, although temporal separation is the simplest concept to explain, and the most relevant to notability issues (since what's wrongly touted as a "secondary source" in the weeks after an event will be derided as merely primary at the successful AFD years later). An investigative report can be a secondary report. A report about a crime may also be a primary report: merely repeating the statements made by involved parties is not sufficient separation. It's not just a matter of counting links in the chain. As someone else said recently, a secondary source is a work of the mind, not simple regurgitation of what you saw or what someone else told you. WhatamIdoing (talk) 04:58, 22 September 2011 (UTC)
Tropical Storm Nicole (2010)?   Will Beback  talk  07:38, 22 September 2011 (UTC)
Press releases are marked as such in news papers precisely because they are word for word what the company said and are thus primary sources regardless of where they appear. Newspapers are responsible to some degree for what they print, but they distinguish press releases to establish that what is said is something for which they are not responsible.
If a reporter is reporting what involved parties said, that report is still secondary--the statements are those of involved parties, and the reporter has the necessary degree of separation. Editors are responsible for vetting the content, if the newspaper is reliable, and thus we can take that report as a secondary source for what was said, as opposed to taking an involved party's later statement later about what they said--that's a key difference. Person X says Y. If a reporter reports that, it is a secondary report of what X said. Person X says they said Y, that's a primary source. Whether Y is true or not is a different matter. I agree that if a reporter is a witness to, say, a demonstration in the street in Syria, they are acting as a primary source, but if they are reporting what others have claimed, they are a secondary source for those claims. You've referenced Wikipedia:Identifying and using primary and secondary sources, but that is an essay, not a policy or guideline. What policy states is "Secondary sources are second-hand accounts, at least one step removed from an event. They rely on primary sources for their material, often making analytic or evaluative claims about them."
As for " a secondary source is a work of the mind", that's not always true, and a primary source is often a work of the mind as well. Research papers drawing conclusions about the results of an experiment, for example, as primary sources the way we look at it, while an analytical paper drawing conclusions across a number of research papers would be a secondary source, but equally a work of the mind.
In regard to temporal separation, it is not so simple as you make out. A report from 1901 about an event in 1601 we would generally take as a primary source for what people in the early 20th century held about that event, but not as a secondary source for the event in 1601. For that we would look for more recent historical works, as we assume that historians build upon one another's work.
Will, I don't have time to look at that article right now, but I'd be happy to later. If you'd like to identify the couple of sources you consider to be the most important for the article, that would save me some time.
Nuujinn, I have never seen an article (not a paid advertisement) marked "press release" in a printed newspaper. Have you? Note that I'm not talking about something merely appearing on a website: actual ink on newsprint.
Also, your claim that quoting someone else makes your paper secondary is simply wrong. Reprinting someone else's words does not magically transform your publication into a secondary source. (And if someone reprinted yours, would that be tertiary? And when I cite that third source, do you propose making up names for the classification?) The policy is trying to provide a simple overview. It says "second-hard" to indicate a significant degree of separation, not to imply that secondary is a fancy way to spell independent. If you post X on your blog, then your blog post is a primary source for your words. If I quote your post, my quotation is also a primary source for your words (but now an independent primary source).
Please: go read real sources about this subject. A number of them are linked above, but there are even more detailed sources available. WhatamIdoing (talk) 16:05, 22 September 2011 (UTC)
I have, and I'm not convinced, since, as Will notes, our goals are not the same as academic goals. I think your argument is counter to the intent of the policy, and that you are making gross generalizations about when and how analysis occurs in reporting in newspapers and magazine. A report in a newspaper that references what someone says and analyses it, especially if it does so in reference to what others have said about the same topic, is very definitely a secondary source as we use the term regardless if appears a day or two after the event. --Nuujinn (talk) 21:58, 22 September 2011 (UTC)
Nuujinn, if you will look over your earlier comment, you will see that nothing like "and analyses it" is present there. Merely repeating the words that someone said to the journalist—what you originally described—does not constitute a secondary source. WhatamIdoing (talk) 01:14, 23 September 2011 (UTC)
@WhatamIdoing: My point with Tropical Storm Nicole (2010) is that you seem to be defining these terms differently than Wikipedia has done traditionally, even recently. It's up to you to show that you are using definitions that are consistent with actual practice on Wikipedia.
Regarding your revert of my edits, you wrote, " This isn't over-reach. It's supported by reliable sources, and the example is obvious".[8] Again, the only "reliable sources" for an issue like this would be Wikipedia discussions. Which discussion or other Wikipedia page are we using as the basis for this assertion: "Some editors—especially those with no training in historiography—incorrectly call these newspaper articles 'secondary sources'."?   Will Beback  talk  23:18, 22 September 2011 (UTC)
Will, my point is that if you don't know what sources the article is primarily based on, then it's going to be hard to figure out whether those sources are, and are being used as, secondary sources. There are 60 sources, and we both know that the article is not 1/60th "based primarily" upon each one of them equally. WhatamIdoing (talk) 01:14, 23 September 2011 (UTC)
Which of those 60 sources would you consider to be genuine secondary sources?   Will Beback  talk  01:15, 23 September 2011 (UTC)
This question amounts to "please analyze every single source and every single sentence in this long article." Classification of a source depends on not only the inherent characteristics of the source, but also on how the source is used. A meta-analysis is normally called a secondary source, but you can turn it into a primary source by the way that you use it. WhatamIdoing (talk) 19:46, 23 September 2011 (UTC)
the point I'm trying to make with that featured article is that it would appear to be composed entirely or almost entirely of primary sources, according to your view. If so, either there is something very wrong with the FA process, or with your definition (for Wikipedia purposes). How can we resolve this question?   Will Beback  talk  21:52, 23 September 2011 (UTC)
The point I'm trying to make is that if you don't know which sources the article is "primarily based upon", then you definitely don't know that those mysterious, unidentified sources are not secondary sources. It is possible—easy, even—to base an article primarily upon one good source, and still cite a dozen others. It's not just a simple matter of counting up the number of sources. WhatamIdoing (talk) 17:24, 30 September 2011 (UTC)

break

WhatamIdoing invited me back here to have a look. Will Beback doesn't seem to be saying anything disagreeable. I have trouble here finding the focus of the discussion. One question is: What is "SPS" in "A press release, in a newspaper, is a primary SPS."? --SmokeyJoe (talk) 13:43, 23 September 2011 (UTC)

primary SPS

I suspect that Will's concerns will be addressed by him reading the footnote in NOR that gives examples of primary sources:

"Further examples of primary sources include archeological artifacts, census results, video or transcripts of surveillance, public hearings, investigative reports, trials (including material — which relates to either the trial or to any of the parties involved in the trial — published by any involved party, before, during or after the trial), editorials, opinion pieces or interviews; tabulated results of surveys or questionnaires; original philosophical works; religious scripture; ancient works, even if they cite earlier lost writings; and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos and television programs.

[....]

"Primary sources may include newspaper articles, letters, diaries, interviews, laws, reports of government commissions, and many other types of documents"."

Every single mention of how to classify news reports in NOR says that they are (normally) primary sources. Despite his belief that this page restricts these sources far more than the policy, we're actually doing the opposite and pointing out that some newspaper articles are secondary sources. WhatamIdoing (talk) 19:43, 23 September 2011 (UTC)

Scholarly papers

One problem I have encountered is the scholarly paper = primary source view of many editors (see Talk:Weston_Price/Archive_3#Planned_reworking_of_second_paragraph_of_introduction and Talk:Weston_Price/Archive_3#Price_and_FIT-inspired_.27holistic_dentistry.27 for two such examples). Certainly the actual words of the person the biography is about regarding one of the two subjects (focal infection theory-root canal) that has made him notable to the general populous would be worth mentioning--especially as all the current sources talk the man's 1923 work and act like he didn't say a word on the matter afterword.

Yet direct nearly entire paragraph quotes from the Journal American Medical Association and a book by Paul B. Hoeber, Inc; Medical Book Department of Harper & Brothers that show that Price's views on the matter were far more complicated then the more recent sources (based on a RS flawed interpretation of Price's 1923 work) show have been kept out simply because they are viewed as primary even though they are in fact a mixture of both primary and secondary.--BruceGrubb (talk) 19:21, 22 September 2011 (UTC)

It is true that "scholarly paper = primary source" is a common default assumption. I think it is generally a useful default assumption. Of course, it is not always true. The nature of a source, whether primary source or secondarary source, can depend on how it is used. The rule of thumb that I find most useful is that if it is a report, it is a primary source; if it is a review, a story, an analysis or a commentary, etc, then it is a secondary source. This works equally well of scholarly papers as for gossip magazines. Often, whether in a magazine or an article there is a sudden change from from primary source material to secondary source material across a single paragraph division. First, the facts are presented. Then, comment is made. --SmokeyJoe (talk) 13:53, 23 September 2011 (UTC)
I agree that Bruce has found Wikipedia's approach to primary sources inconvenient for adding his beliefs about Price's work to that article. WhatamIdoing (talk) 19:49, 23 September 2011 (UTC)
Here we see an example of editor hiding behind claims of belief to keep stuff out:
"The relationships between dental infections and degenerative diseases, if such exist, should be demonstrable by other means than the establishment of simply an association of the two in the same person, or the development of such lesions in experimental animals with cultures taken from focal infections."(Price, Weston A. (1925) "Dental Infection and related Degenerative Diseases" J Am Med Assoc 1925;84(4):254-261)
"In my search for the cause of degeneration of the human face and the dental organs I have been unable to find an approach to the problem through the study of affected individuals and diseased tissues. In my two volume work on "Dental Infections," Volume I, entitled "Dental Infections, Oral and Systemic," and Volume II, entitled "Dental Infections and the Degenerative Diseases," (PRICE, W. A. Dental Infections, Oral and Systemic. Cleveland, Penton, 1923) I reviewed at length the researches that I had conducted to throw light on this problem. The evidence seemed to indicate clearly that the forces that were at work were not to be found in the diseased tissues, but that the undesirable conditions were the result of the absence of something, rather than of the presence of something. This strongly indicated the need for finding groups of individuals so physically perfect that they could be used as controls. In order to discover them, I determined to search out primitive racial stocks that were free from the degenerative processes with which we are concerned in order to note what they have that we do not have." (Price, Weston (1939) Nutrition and Physical Degeneration: A Comparison of Primitive and Modern Diets and Their Effects by Paul B. Hoeber, Inc; Medical Book Department of Harper & Brothers)
"It is very important that in the consideration of the dental caries problem it shall be kept in mind continually, that it is only one of a large group of symptoms of modern physical degeneration and when teeth are decaying other things are going wrong in the body. Fluorine treatment, like dental extractions, cannot be a panacea for dental caries." (Price, Weston (1939) Nutrition and Physical Degeneration: A Comparison of Primitive and Modern Diets and Their Effects by Paul B. Hoeber, Inc; Medical Book Department of Harper & Brothers)
This NOT my belief as WhatamIdoing claims but three exact quotea that meet Verifiability and yet by playing the "oh they are primary sources" card editors insist on keeping them out even though they provide a more NPOV regarding Price's actual view rather than the distorted version that is largely the product of George E. Meinig.
In fact, if not for Gunnar Hasselgren's 1994 Annals of dentistry: Volumes 53-54 New York Academy of Dentistry pg 42-43 review we editors wouldn't even know just how distorted Meinig's presentation really is--Meinig uses only Price's self published 1923 work and makes some conclusions that have no basis in anything (least of all Price).
The consensus of the editors (Talk:Weston_Price/Archive_2#Weston_Price_cautious_about_focal_infection_theory.3F) was that there were serious problems with the modern sources as there was evidence that they were going on how Price's work was currently being used rather than what Price himself actually said and did. For example, it was shown using Price's own words that Stephen Barrett had no idea what he was talking about (Talk:Weston_Price/Archive_1#Weston_Price_and_Stephen_Barrett_in_their_own_words) and yet a reference to that nonsense is still in the article. We keep out peer reviewed papers but allow blogs. SAY WHAT?!!--BruceGrubb (talk) 21:25, 23 September 2011 (UTC)

Dictionaries

There is some disagreement among scholars whether dictionaries are secondary or tertiary sources. Personally I don't see where the disagreement comes from as they are clearly secondary sources. The writers of dictionaries look at word usage across thousands of examples, and for their purposes each of those examples is a primary source. Some of those examples might be secondary or tertiary sources for other uses, but in the way dictionary writers use them they are primary. The only counter-example is when/if dictionary writers look to other dictionaries as a source. That's the only way in which dictionaries can be seen as tertiary. Anyway, as there is confusion about this we should not be stating flat out that dictionaries are tertiary sources. Mystylplx (talk) 05:13, 9 December 2011 (UTC)

Here's your list:

  • [9]—talking solely about a particular use (foreign language acquisition), not the general concept. Even a tertiary source can be made primary by the way you use it.
  • [10]—The convention in a single field of study (law) that does not happen to use tertiary classification for anything.
  • [11]—Ditto.
  • [12]—Another guide for law students.
  • [13]—Yet another source specific to the legal profession.
  • [14]—Another website that doesn't admit that any source could be tertiary.
  • [15]—Yet another website that doesn't believe in the existence of tertiary sources.

There were two other links that said that dictionaries could be secondary or tertiary, depending on the circumstances, which is not at all good support for your claim that they are secondary sources, especially compared to the many sources available that say they are exclusively tertiary. What you need to find—and haven't, apparently—is something that (1) deals in a three-part classification system and (2) defines dictionaries as being secondary (not maybe both secondary and tertiary). WhatamIdoing (talk) 19:43, 9 December 2011 (UTC)

Uh, actually sources that say they are both secondary and tertiary are quite sufficient to say they are both secondary and tertiary. And the rest of your objections are pretty strained. I could invent equally strained objections regarding the 3 sources you gave saying they are exclusively tertiary, but I'm not going to bother. Mystylplx (talk) 23:28, 9 December 2011 (UTC)
However, sources that say they are both secondary and tertiary are completely inadequate for your claim that dictionaries are always secondary, which is what you had written. WhatamIdoing (talk) 01:27, 10 December 2011 (UTC)

Affiliated sources

Are official published War Diaries of World War I (used by the British and Australian armies since 1907), and written by adjutants or intelligence officers, considered to be affiliated sources? This tag has been added to Sinai and Palestine Campaign and Battle of Magdhaba without any explanation by the editor and repeated cutting only invites its reappearance. As the war diaries are the only sources which have been contentious, although no longer so, I wonder if these may be the problem. --Rskp (talk) 02:02, 17 January 2012 (UTC)

We don't really use the phrase "affiliated sources" any more (were we ever using it much)? There are three distinct concepts to look at here: Independent sources, third-party sources, and primary/secondary/tertiary sources. WP:Party and person covers the distinctions. Quoting our article: "A war diary is a regularly updated official record kept by military units of their activities during wartime." So, that's clearly a primary source. All diary-type things are. As an official military record it's neither independent of nor third-party in relation to the army/government, if that's the question that's being asked in the dispute. Just because they're not primary and not independent (independent is a subset of third-party), doesn't mean they can't be used at all, just carefully per WP:PRIMARY.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:15, 28 May 2015 (UTC)

The "closer to the event" criterion is unreliable

I'm very much bothered by the sentence One rough rule of thumb for identifying primary sources is this: if the source is noticeably closer to the event than you are, then it's a primary source. If you think about the various possibilities, it doesn't make much sense. Scenario 1: an event occurs in the year 1800; a newspaper writes about it in 1880. The year 1880 is noticeably closer to 1800 than the present day is. Is it a primary source? Scenario 2: an event occurred yesterday; a newspaper writes about it today. The newspaper is no closer to the time of the event than I am; is it a secondary source? As rules of thumb go, I don't think this is a good one. Jowa fan (talk) 05:25, 22 August 2012 (UTC)

Some other examples: say some scholar publishes something popular in 1990, and you have a bunch of others critique it in 1991. They're way closer to the event than me, but still secondary. Alternatively, there may have been a fire in 1990 and new results of the arson investigation just came out last year. The results are much closer to me than the fire, but primary. I have doubted the authority of that rule of thumb as well. But I get its general point: usually, primary sources are published earlier than later. I would however caution people to pay more attention to the content of sources and less to timelines, at least for Wikipedia's purposes. NTox · talk 07:39, 27 August 2012 (UTC)
There are other problems with this "rule of thumb", too. It's only applicable to certain kinds of material, though we should err on the safe side. A newspaper quoting what someone said in a speech in 1882, or 1933, or 2015, is still a secondary source for WP purposes (cf. the Willie Thompson quote at #Academic historians, below.) We really don't care whether a historiographer would prefer to treat the 1882 one as a primary source, because it was past some arbitrary date line, for reasons peculiar to historiography. WP is not a historiographical work. We're much closer to academic publishers' conceptions of primary, secondary, and tertiary, though the WP interpretation of them doesn't match that exactly, either. And it can't, because our goals are different, and our approach to information and sources is different. We have a very different (much looser) standard for what constitutes adequate sourcing than a peer-reviewed journal does. It's a mistake to take primary vs. secondary source determination criteria from an external discipline and attempt to apply them robotically to WP.

Pre-modern material is always primary, because it pre-dates both journalistic ethics and scientific publishing standards; virtually everything surviving to us from the distant past in written form has a political, religious, or other POV to push, while the concepts of neutrality, fairness, accuracy, and reproducibility of results hardly existed (and tended not to last long).

As for old newspapers, the canonical example, they're treated as if primary sources for anything that could be biased, that could have changed, or about which understanding could have changed, in the intervening span, such as interpretation of historical facts, socio-political rationales, explications of scientific discoveries, etc. For much else, old newspapers, published in countries with broad freedom of press, remain secondary sources (e.g. for fixed, simple historical facts like who won a billiards match, whether a law passed, etc.), but we simply are not obligated to accept them as reliable ones. If there's any doubt at all, do treat them as primary sources. (And it is a treatment; it's a logic mistake to say they "are" primary; no magic changed them mystically from one into the other.) Such doubt about treating old news as reliably secondary can come from anywhere reasonable. If a flood destroyed a large part of a New Orleans neighborhood in 1870, and a newspaper of the era reported that 87 homes were destroyed, we might want to treat that as a primary source, and say The New Orleans Gazette reported after the flood that 87 homes had been destroyed. Why? Because racial prejudice at the time could easily have ignored the loss of over 100 shacks in a former-slave shanty town, declining to dignify them as "homes" to the largely-white audience of the publication. These kinds of blind spots exist in a lot of old reportage, especially with regard to women and ethnic minorities (which was a wider problem back then; e.g., in North America, the Irish, Italians, Poles, Cornish, and many other groups were subjected to more ethnic discrimination, but today are generally lumped all together with other European-Americans). It's always safest to take this treat-as-primary approach with old newspapers. But we need not do this, and it's irritating to readers, for trivial matters, like the bare fact of (not interpretation about) the election or resignation of a local politician.

Old newspapers were also primary all along, as are newspapers today, for novel material, such as an opinion-laden book review. Investigative journalism also generally constitutes original, primary-source research by a journalist, though it may mix in secondary-source aspects, like synthesizing what has come out of interviews in the course of the investigation. A reporter writing a piece in which he purports to have uncovered the cause of a series of warehouse fires or whatever, that's primary, whether it was published this afternoon or a century ago; it's his own "theory", unexamined by other, secondary sources.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:07, 28 May 2015 (UTC)

Thank you

. . for codifying WP:NOTGOODSOURCE. So important, but so seemingly invisible much of the time. . NTox · talk 07:53, 27 August 2012 (UTC)

Newspaper articles

I added and removed some material to make clear that current newspapers are mostly secondary sources. The page gave the impression that most are primary sources. This is true for historical documents, but not true for current affairs. I also removed some of the examples, because they seemed confusing. Calling a book review partly a primary source, for example, if it describes the book, seemed misleading. SlimVirgin (talk) 03:35, 27 September 2012 (UTC)

WAID reverted, but I've restored the text I added with another source, a former professor of history at Glasgow: "Leaving the remote past for any other century, including the twentieth, the same principles apply no less. Newspaper reports, having been subject to editing, are secondary sources for the events they record; the original reporter's notebook, if it had survived, would be a primary source. A document can, as we have seen, be in different dimensions, either primary or secondary."
Can we base everything on academic sources to settle this? I've added two as a start. If the issue is nuanced, we can reflect that. SlimVirgin (talk) 04:48, 27 September 2012 (UTC)
(edit conflict) A few hours ago I reverted an edit on this issue. I don't think it's clear-cut—we certainly can't say that all newspaper articles are primary sources. However, I certainly don't think the opposite viewpoint is justified. Of the various recent versions of this page, I prefer this one. I hope a consensus can be reached in the talk page without this turning into an edit war. Jowa fan (talk) 04:59, 27 September 2012 (UTC)
WAID, could you post here some academic sources that use the definition you are using, for contemporary news reports? I only have undergraduate history, but we were definitely taught that current news items (and most old news items too) would not be regarded as primary sources; all the sources that talk about news items being primary sources are talking about historical sources, and even there many do not regard such items as primary. And anyone who has worked in research would not (I think) regard a current newspaper story as a primary source.
So it would be helpful to see some academic sources who say what you're saying, then we can include both perspectives. SlimVirgin (talk) 05:07, 27 September 2012 (UTC)
WAID is reverting and removing the academic source I added. [16] SlimVirgin (talk) 05:23, 27 September 2012 (UTC)

That whole section has serious issues, and really badly fails to understand the difference between secondary and tertiary sources. This whole document appears to, and it leads to a lot of errors. Just to speak to reviews, as one example, they can be primary, secondary, or tertiary (and most often a mixture). An opinional, aesthetic review is a primary source. An analytical review of a work, that refers to previously published material, is a secondary source. A "review" that is simply a neutral abstract with neither analytic nor emotive content, is tertiary. There's a similar error in the segment just before this; the hypothetical "A newspaper column lists the events reported in that newspaper on the same date from 25, 50, 75, and 100 years before", is a tertiary not secondary source. Even the TV show mentioned in the same line isn't necessarily a secondary source. As an example, let's say the History Channel does another of its quasi-documentaries, this time on mysteries surrounding the building of the Egyptian pyramids. If it's a totally regurgitative work, presenting no new analysis, commentary, interviews, etc., and simply repeating what mainstream, reliable secondary sources say (perhaps being a straightforward video adaptation of one of them), then it's a tertiary source. If adds in synthesis, brings in experts to comment, and suggests avenues for additional exploration, then it's a secondary source. If at any point it makes a novel claim, e.g. that "some mysteries about the pyramids can't be explained", or declares as mysterious and unsolved something which actually does have explanations in reliable, secondary sources the producers ignored, then for those statements it is a primary source.

Simplistic approaches to determining what are primary, secondary, and tertiary sources leads to errors. The medium or format of the publication is not in any way determinative, only the content is. While we can identify trends in one direction or another – material in encyclopedias is usually tertiary, etc. – there are always exceptions, some of them categorical. Before I knew this page existed (hardly anything links to it), I began addressing these questions with regard to tertiary sources, at WP:TERTIARYUSE.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:32, 28 May 2015 (UTC)

Academic historians

Starting a section here for the views of academic historians, so we can include all perspectives. SlimVirgin (talk) 05:15, 27 September 2012 (UTC)

  • Willie Thompson. What Happened to History?. Pluto Press, 2000, pp. 7, 79 (former Professor of Contemporary History at Glasgow):
  • "Leaving the remote past for any other century, including the twentieth, the same principles apply no less. Newspaper reports, having been subject to editing, are secondary sources for the events they record; the original reporter's notebook, if it had survived, would be a primary source. A document can, as we have seen, be in different dimensions, either primary or secondary."
  • Suraiya Faroqhi. Approaching Ottoman History: An Introduction to the Sources.Cambridge University Press, 1999, p. 2.
  • For while the available primary sources conditions the kinds of question an historian might usefully ask, it is also true that we read secondary sources such as newspapers and magazines, long before we ever embark on specialized training."
Thompson is using a minority definition of "secondary". This is most obvious on page 79, where he directly equates the term secondary with anything that is published:

[They] "will have as their first undertaking to read all feasible 'secondary'—i.e., already published—texts"

In that system, hand-written information in a lab notebook is primary, but the original publication describing the experiment is secondary; my initial notes on the car wreck I witnessed are primary, but my blog post on it is secondary; etc. This is not the definition that Wikipedia uses. We follow the much more common definition, which says that both lab notebooks and original publications are primary, but review articles are secondary.
As for Faroghi's point, IMO some magazines are primarily secondary sources. A newsweekly that focuses on who-what-when-where facts will contain a lot of primary sources, but one that focuses primarily on analysis, like The Economist, is largely a secondary source. Many feature-length articles in major newspapers are secondary sources. But the majority of newspaper articles (When was the last time you got newsprint on your fingers?) aren't. The majority of newspaper articles are short pieces about which local politician said what, where this week's charity event will be held, who got arrested, and what the outcome of the court case was yesterday. All of that is primary source material. WhatamIdoing (talk) 20:31, 28 September 2012 (UTC)
What I think we are missing here is the element of time. Sources can change their "status" over time. They can shift from being secondary to primary due to age. A newspaper from the 1700s would be considered a primary source by modern historians, even if it contained secondary analysis of previously published primary material. Blueboar (talk) 12:41, 29 September 2012 (UTC)
It's not just centuries-ago newspapers. The page quoted from Indiana University Bloomington gives newspaper reports from the year 2000 as an example of primary sources.
One example that makes the nature of news reports clear to me is the 9/11 attacks. Do you remember what the news coverage was like? Do you remember that nearly every day for weeks and weeks, we had a new, and almost always lower, body count? Even the "analytical" pieces were filled with factual errors that are obvious now (and often with formal corrections issued, not things that only a kook believes to be erroneous). We shouldn't be using those newspaper reports in our articles about the 9/11 attacks, because they are best treated as primary sources rather than as secondary sources. Something published a month after the attacks doesn't have the necessary distance from the actual events to be a secondary source. WhatamIdoing (talk) 22:58, 1 October 2012 (UTC)
Well, I would not go as far as saying we could never use those reports... we can use them... but in very limited ways. We simply need to treat them with the same caution we would give to other primary sources. (for example, I could see us using them for an article or section describing how the Media reported on the attacks ... describing what the initial reports said, and how they were corrected over time.) Blueboar (talk) 15:17, 2 October 2012 (UTC)
Yes, that's what I meant to say. They should be acceptable if used like any other primary source, but they should not form the basis of the article and they should not be used to contradict later secondary sources. WhatamIdoing (talk) 20:00, 2 October 2012 (UTC)

On the “closer to the event” rule (again)

Isn’t every published source closer (in time) to any given event than the present? Especially if it wasn’t published in the very recent past. It seems to me that this “rule of thumb” sorely needs explaining or rewriting, or even sourcing. —Frungi (talk) 05:40, 22 May 2013 (UTC)

I see what you mean. What we're aiming for here is that the source is closer to the event than you are to the source. So the event is in 1688, the source is in 1776, and you are in 2013: the source is primary. But if the event is in 1963, the source is in 2001, and you are in 2013, then it's at least possible that the source is not primary. WhatamIdoing (talk) 05:29, 26 May 2013 (UTC)
So the actual rule of thumb, then, is that a secondary source is no less than half as old as the event, yeah? I’d still very much like to know where that rule came from, but I’ll edit it to be more clear on that point. —Frungi (talk) 05:35, 26 May 2013 (UTC)
I think you'll find that it's more complicated than that. WhatamIdoing (talk) 06:00, 26 May 2013 (UTC)
Well, that was what you said, wasn’t it? Unless you’re using an odd measurement of time, or meant something other than time. I hate to keep repeating myself, but a source for that rule would definitely help my (and, no doubt, others’) understanding. —Frungi (talk) 06:04, 26 May 2013 (UTC)
It's more complicated than that. So the examples I gave hold true: if the Glorious Revolution of 1688 is mentioned in sources written during the American Revolutionary War, you must treat them as primary sources now. If JFK's assassination is mentioned in a source written shortly after the 9/11 attacks, then (depending on what the source says), that might be secondary.
But there are other examples that give different proportions and won't hold true: The Gospel of Mark was written sometime late in the first century; Matthew Henry wrote his commentaries in the early 18th century; and Henry's work should still be handled as primary sources. Similarly, if the Oscars are announced on Sunday evening, and Tuesday's paper contains a truly analytical article about the winners, then you may use that article on Saturday as a secondary source, even though the source is two days away from the event and you are four days away from the source.
If it were simply a matter of subtracting dates, then we could have supplied a link to an internet date calculator. But it's not simply a matter of subtracting dates. It's more complicated than that. WhatamIdoing (talk) 06:38, 26 May 2013 (UTC)
Hence it’s a “rule of thumb”. But it’s frankly sounding like that rule doesn’t have much value, so why include it? —Frungi (talk) 06:52, 26 May 2013 (UTC)
It's helpful to people who are dealing with old documents. It is a useful way of explaining the problem to people who have found a "review article" about a historical treatment from the 19th century, for example. It's not going to be helpful to everyone. WhatamIdoing (talk) 23:18, 29 May 2013 (UTC)
It's not actually helpful though. It's just leading to dispute, the reasoning in it is faulty in several ways (see earlier thread, where I explain why, though agreeing we should err on the safe side), and hardly anyone knows this page exists, so it's not providing actual advice to much of anyone anyway. (This essay needs its kinks worked out first, then to be "advertised" in the "See also" sections of all the relevant policies and guidelines and other essays.) I'll be happy to help work on this, as I've been writing WP:Use of tertiary sources, and they perhaps should be merged, though their approaches are presently very different.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:46, 28 May 2015 (UTC)
I've been teaching historiography for 50 years and never came across anything remotely resembling that strange "rule." Take Abe Lincoln--died 1865 = 152 years ago, half that is 76 years = 1941. The "rule" indicates that people in the 1930s writing books and articles about President Lincoln were creating primary sources! there is zero sourcing so we can drop it. Rjensen (talk) 22:41, 17 June 2017 (UTC)

Subject bias

This essay, especially the final section, seems to lean too far toward the academic, historiographic uses of “primary” and “secondary”, rather than the definitions used by Wikipedia. In many places, it seems to be saying that the way Wikipedia internally uses the terms is somehow “wrong” because it’s different from how an academic field uses them, when really it should be about how Wikipedia uses them. I mean, the essay is in the Wikipedia: namespace, not the Historiography: namespace. —Frungi (talk) 05:54, 22 May 2013 (UTC)

So that section on media sources, is it about Wikipedia? Because if it isn’t—if it’s about how sources are classified in a particular field of study—it really doesn’t belong on a Wikipedia-namespace page giving general direction to general editors. —Frungi (talk) 05:46, 26 May 2013 (UTC)


The problem with classifying news according to "how we use it" is WP:Notability. In fact, the news section used to be a part of the section about notability.
Once upon a time, we had a few policy editors who (in good faith) thought that 'secondary' was basically the same thing as 'independent'. So the words were used pretty much interchangeably in various policies and guidelines. That's been fixed in most of them since then. The concept of "secondary" sources crawled into WP:N sometime in 2007 (e.g., March (don't miss the confusion evident in that edit summary), but not January).
Sometime later, some editors started figuring out what 'secondary' means among academics, and also figured out that we didn't actually want articles based on primary sources (as defined by most academics [there are some significant differences in academic fields]; previously, articles solely based on primary sources were explicitly permitted at NOR), and we developed a problem: if you go to NORN or RSN and ask about a source, you get a response that most academics would recognize, e.g., that eyewitness news reports are never secondary sources. But if you go to AFD and ask about the same source, you get a very different answer. The reason for this is that the WP:GNG requires secondary sources, and people "just know" that a stack of newspaper articles about whatever recent event is supposed to make their new article immune from deletion. So we get handwaving assertions that newspaper articles are secondary, even when the sources in question are plainly primary sources (and plainly not what the GNG wants from a secondary source).
Short of convincing people to re-write the GNG, I don't believe that we can produce a single answer of "how Wikipedia uses them". We have two definitions. One is fairly close to how historiography defines it, and the other is what early versions of this page called "Please don't delete this article sources". This second definition is actually wrong: it is not what academics say, it is not what our own policy at NOR says, and it is not what editors at WT:N claim they mean. But it is nonetheless a pervasive claim at AFD. WhatamIdoing (talk) 05:59, 26 May 2013 (UTC)
In other words, we have one definition, and then we have a misunderstanding of the concept. —Frungi (talk) 06:38, 26 May 2013 (UTC)
If you don't want to say that we have two incompatible definitions (NOR vs N), I would say that we have one definition (which is not exactly the same as the definition used by any particular academic field, although it's closest to the historians'), one dramatic misuse of that definition that is widely supported, and a serious lack of education.
Or, to go back to your original comment: yes, we're saying that the way Wikipedia internally uses the terms is frequently wrong, because it is frequently wrong, and wrong even according to Wikipedia's own official definition in the policy, not just according to academics' definitions. WhatamIdoing (talk) 06:43, 26 May 2013 (UTC)
I’m sorry, I thought you were saying the misuse in AFD was the second definition rather than WP:NOR/WP:N. That’s probably my sleep-deprived fault. Though “secondary sources” in the GNG links to NOR rather than giving its own definition, so I’m still not seeing it. —Frungi (talk) 07:01, 26 May 2013 (UTC)
The 'definition' actually used at AFD, and blamed on WP:N, has nothing to do with NOR. The link to NOR is simply ignored. WhatamIdoing (talk) 23:24, 29 May 2013 (UTC)
I strongly agree with the concern raised in the original post. It's okay for us to explain (more briefly) how some other contexts use these terms, but this essay should not be advocating or imposing them, or even dwelling on them. WP has its own meanings and criteria for these classifications, and if this essay continues to buck them, it should be userspaced. I think it has the potential, after a lot of cleanup and reworking, to be very useful, but the first step in this process is going to have to be normalizing it to Wikipedia's own internal voice. I appreciate WhatamIdoing's historical view of how problems have arisen, but agree with Frungi's "In other words, we have one definition, and then we have a misunderstanding of the concept." Our "job" with a page like this is to figure out what is clearly intended in policy, explicate and apply it better, and disabuse people of the incorrect interpretation. If we do this really well and still come to the conclusion that one policy or guideline or the other needs a wording tweak, it should not be an insurmountable challenge to get that taken care of.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:56, 28 May 2015 (UTC)

Ancient historians

I have just discovered an article on ancient history (Dacia) seriously (with a straight face, if articles had faces) citing ancient historical writings. Headdesk. It should be obvious that these are not citeable sources for Wikipedia (at least for everybody who has a clue about the methodology of historical research: any source older than about 1950 cannot be taken at face value, even if it was written by an academic historian), but obviously it is not. Perhaps this point merits an explicit mention somewhere. --Florian Blaschke (talk) 00:38, 8 September 2013 (UTC)

Yes, it should. Ancient writings can be used as primary sources in limited ways, but we would never literally cite directly to them, but to a modern, translated edition with analysis, or better yet simply to a secondary source that analyzes it as a topic. It would be perfectly fine to cite a secondary source at analyzed the poems, for assertions a WP article makes about on the meaning of a passage in "Sir Gawain and the Green Knight", while also using and properly referencing J.R.R. Tolkien's translation of the poem to provide illustrative quotations in the article, if we like his version better that some other translation. In neither case would we cite the ancient manuscript itself.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:40, 28 May 2015 (UTC)
Agreed. "Ancient" is perhaps too narrow, considering that, as the OP pointed out, most historical scholarship before the mid-20th century is seriously problematic, with respect to both factual accuracy and POV. I have seen an alarming number of articles on WP referenced entirely to 19th century sources, despite a wealth of recent scholarship from the past two decades. There is really no excuse, considering how drastically changed our views of many of these subjects are today. "Prefer recent sources unless the older source is still considered more authoritative/accurate by experts today" would be my take on it. In an ideal world, I'd also like to see a cleanup template for articles that rely too much on outdated scholarship. I want to see what others have to say on it first though. --diff (talk) 01:34, 28 July 2015 (UTC)

Use of primary peer reviewed sources from prominent minority views that seem to contradict secondary sources from majority view.

I am seeing some editor debate on articles I wanted to improve, and there seems to be a lot of push back on minority views in contentious medical areas. (Lyme disease)

I think I have demonstrated that the minority view has multiple prominent adherents, and that there are at least 100 peer reviewed papers that support some of the minority claims. By the very nature of being a minority, there are only some primary sources. But in order to make verifiable claims with appropriate weight, this requires apply some number of primary sources. There are usually multiple papers to cite, but of course none of them come from the medical societies that are driving the mainstream view.

All positions seem realistic, and the growing body of research is advancing the state of the disease on both sides, so it seems important to cite medical research in the context of the minority view.Bob the goodwin (talk) 08:37, 28 December 2013 (UTC)

Sorry but this is the wrong approach. Please see the intro to WP:MEDRS. We need to present in Wikipedia mainstream views on health-related issues. There are some who go so far to say (and I am pretty close to this position myself), that if health-related content cannot be supported by a reliable secondary source it should not be in Wikipedia at all - we are not a medical debating society - we are an encyclopedia presenting knowledge to the public. It is OK to represent significant minorities but they must be presented that way, and only as represented in reliable secondary sources. I don't know how closely you follow the primary biomedical literature but there is a huge percentage of primary studies that turn out not to be replicable. In other words, that are not reliable science. That is one of the big reasons why we need to hew closely to the scientific consensus wherever we can find it. Jytdog (talk) 03:18, 29 December 2013 (UTC)

medicine is different

Wikipedia:Identifying reliable sources (medicine) cuols be added somewhere here as it seems different — Preceding unsigned comment added by 49.49.40.207 (talk) 13:03, 14 April 2014 (UTC)

Spinach

This is a fantastic read regarding how important it is that when we use secondary (or later) sources, we should always attempt to trace back to the original source to ensure that it says what we're asserting it says. This is a staggeringly common problem on Wikipedia. Chris Cunningham (user:thumperward) (talk) 08:44, 31 July 2014 (UTC)

Thanks for sharing that Chris... it was fascinating. Blueboar (talk) 22:43, 28 May 2015 (UTC)

Plato and Socrates etc.

An interesting point has come up in Socrates. Is Plato a reliable source for the life of Socrates? Plato will have to be used of course, but should he be the main source, to the exclusion of modern sources. It has been argued that these later sources generally rely on Plato anyway, and so it would be better to use Plato directly. I would argue that modern sources are necessary because Plato might well be considered as a primary source for Socrates, and we need other sources to explain him. The same sort of argument might be made about eg. Boswell and Johnson. Myrvin (talk) 08:38, 25 April 2015 (UTC)

You are asking two separate questions... 1) is Plato a reliable source for the life of Socrates? Yes. An article on Socrates that did not mention (and thus cite) Plato would be incomplete. 2) Should Plato be the (main) source? Probably not. Certainly, where modern scholarship disagrees with Plato, we should highlight the discrepancy. The same would be true for Boswell and Johnson. Blueboar (talk) 22:02, 28 May 2015 (UTC)

conflict of interest of a source

WhatamIdoing back in March 2013 you made the following edit, adding the underlined text: "It is a third-party or independent source, with no significant financial or other conflict of interest. That sentence is part of a bulleted list, introduced by: "According to our content guideline on identifying reliable sources, a reliable source has the following characteristics:". WP:RS doesn't say anything about conflict of interest of a source. I am not sure what this means in the context of WP:PAG broadly speaking. Can you please say more, about what you meant? thx Jytdog (talk) 21:32, 3 June 2015 (UTC)

For a source to be independent, it has to be uninvolved in the issue, by definition. That would obviously include COI. Sarah (SV) (talk) 21:47, 3 June 2015 (UTC)
i understand that you are interested in getting clarity on issues around COI and sourcing. I am too.
I don't know what it meant to WAID to say that a source has a COI, so I asked her.
about your statement, SV... Authors can have a COI (they are required to disclose it). Funding of research used by authors to do research that is published in a source can influence the results/conclusions conveyed in the source; that funding is also disclosed in the source itself. A source can be SPS and not independent of its author... all three of those things are understandable to me. WP:RS only deals with the third of them. I don't understand what "a source has to be uninvolved in the issue" means, concretely, to you, nor what it means to you to say that a source has a COI. What does that mean to you? Jytdog (talk) 22:53, 3 June 2015 (UTC)
I see you restored the content. i still don't know what that means. this is just an essay, not a big deal. Jytdog (talk) 00:22, 4 June 2015 (UTC)
Hi Jytdog,
Actually, WP:RS stresses the need for third-party sources. The WP:BIASED section specifically mentions financial COIs. On the English Wikipedia, third party and independent are used interchangeably, so everything it says about third-party sources is also about independent sources. (There is a technical difference, but conflating the two pretty much only bothers lawyers.)
Given the relatively poor state of WP:INDY and WP:THIRDPARTY (which I still hope to merge some day, as soon as I find another 30 hours or so to spare), my goal in that edit was to help people understand what it means for a source to be independent/third-party. The main source of non-independence is having some sort of COI.
It's really important to remember that we're talking about COI, not WP:COI. This is about the sort of real-world COI that an ethical author would disclose in a research article, not the sort of COI that a Wikipedia editor has.
"Source" has three meanings on the English Wikipedia: author, publisher, and document. A document probably can't have a COI, but both authors and publishers can and do (even when they're not the same entity/self-published). WhatamIdoing (talk) 04:36, 4 June 2015 (UTC)
Thanks for writing here @WhatamIdoing:. I read BIAS before I wrote here. What BIAS says is: "Common sources of bias include political, financial, religious, philosophical, or other beliefs". In other words, "financial beliefs". Which is not saying financial COI. And which I found to be strange and strained. I'm also trying to reconcile your dif above, with this dif] at MEDRS, where you added to MEDRS the following: "Do not reject a high-quality type of study because you personally disagree with the study's inclusion criteria, references, funding sources, or conclusions." WP:MEDINDY seems to be focused on FRINGE stuff not on any kind of financial COI. In this thread in the WT:MEDRS archives you argued that a third party/independent analysis doesn't read on the composition of a journal's editorial board. I think it would be very helpful if our guidelines and essays on sources dealt clearly with what a COI or lack of independence in a source (publisher, authors, funding too) is and looks like. Am not asking for a definition to wikilawyer but rather just some clear and coherent discussion of it.
In my own editing, I strive to use sources that are what I view as independent as best as I can define that (again I think our guidance is weak). As an example, there is a discussion about sourcing for a statement about the relative safety of GM food. Two of the most recent reviews (published in peer reviewed journals) are authored by Monsanto scientists: PMID 25972882 and PMID 24579994; I wouldn't cite them on this, nor would I cite other sources by advocates with clear financial ties like Jon Entine or something put out by BIO nor an SPS by an advocacy group like this. The current content does cite a review in a solid, peer-reviewed journal by Pamela Ronald, a University of California scientist who is very well regarded and who does a lot of public outreach about GM food - that review cites funding by the NIH and the DoE. I think SV's question at WT:MEDRS about how we should handle funding of published research (especially of reviews) is a good one, and I would like our discussion of independence/third party/BIAS to present a coherent picture that everybody can make sense of and follow. Jytdog (talk) 11:47, 4 June 2015 (UTC)
  • If you read BIAS as applying to "financial beliefs" rather than financially motivated sources of bias, then I did a poor job of writing it (hardly surprising; I started with a bold effort and didn't get many suggestions for improvements in the subsequent discussions).
  • Independence (intellectual, financial, and otherwise) is good. It's not an absolute requirement. A systematic review with "tainted" funding (from your POV) is not unreliable, and is not worse than a case study that "pure" funding (from your POV). WhatamIdoing (talk) 17:15, 4 June 2015 (UTC)
i agree 100% that independence is good. it is just a matter of saying what that means relatively clearly. Jytdog (talk) 19:10, 4 June 2015 (UTC)
So far, I think that "relatively clearly" is going to take between two and three thousand words. This place on this page can reasonably include about half a sentence on the subject. Assume that the main goal is to help inexperienced people remember that WP:Secondary does not mean independent and that self-published doesn't mean non-independent. What would you tell them that independent means? WhatamIdoing (talk) 05:21, 5 June 2015 (UTC)
Thanks for asking! I think it would be more clear if we changed this from
  • from: third-party or independent source, with no significant financial or other conflict of interest.
  • to: third-party or independent source, where the authors or publisher have no significant financial or other conflict of interest.
Might it be useful to add to the beginning of the short paragraph following the list, something like: "A source that has all these characteristics is optimally reliable; a source that lacks one of these characteristics is less reliable; consider finding a better source, and if you must use the source, use it only with care and with attribution."? Jytdog (talk) 09:51, 5 June 2015 (UTC)
WhatamIdoing thoughts on that? I know you are busy working on BRD which may be promoted to a guideline but didn't want to let this fall by the wayside. Jytdog (talk) 02:23, 8 June 2015 (UTC)
If you are willing to change the where to something like for which (because a source is not a geographic location ;-) then I think that all of your suggestions here would be significant improvements. Be bold. WhatamIdoing (talk) 04:52, 8 June 2015 (UTC)
Jytdog, I would object to that change. Authors and publishers are not the only sources of COI, as we've discussed elsewhere. Although COI funding would normally give the authors a COI, a situation could arise where it seems not to. Another example is where an author's source has a COI unknown to the author (a history book written on the basis of a conflicted primary source). And on WP, if someone makes an edit by more or less copying over material written by a paid advocate, neither the editor/author nor WP has a COI, but the article has been affected by COI.
What is the purpose of changing the wording? Sarah (SV) (talk) 05:24, 8 June 2015 (UTC)
b/c COI of a source is not meaningful. a document cannot have a COI. Jytdog (talk) 06:06, 8 June 2015 (UTC)
SV, I'm not looking at this as the be-all and end-all of the definition. I'm just trying to get half a sentence or so to give people an idea of what we mean. How would you explain INDY sourcing to someone who has no idea what we're talking about? WhatamIdoing (talk) 06:30, 8 June 2015 (UTC)

Name change to Identifying and using primary sources

This essay was titled, "Wikipedia:Identifying and using primary and secondary sources". I just changed it to Wikipedia:Identifying and using primary sources.

Using secondary sources is the norm, and practically all guidance and policies on Wikipedia editing refers to secondary sources. As this essay says, "primary does not mean bad" and "secondary does not mean good", but usually it does, and if anyone has come to this essay they have gone through the basics and want more nuance. This essay does not try to describe the common case in Wikipedia of using "reliable sources". Instead it is a place for describing the use of exceptional cases - sources which are not reliable sources, and which are primary and under the rule of WP:PRIMARY.

I would like to eliminate the suggestion that anyone should come here to learn to identify and use secondary sources. This is a guide for distinguishing between primary and secondary in confusing cases, but usually there is no confusion, and usually the situation is clear. I think this page's focus should be

  1. identifying primary sources
  2. distinguishing primary and secondary sources, if immediate identification is not easy
  3. using primary sources

I think this essays focus should not be

  1. Basics of using reliable sources, which is the common case covered elsewhere
  2. characteristics of secondary sources outside the context of distinguishing them from primary

The part most lacking in this essay is usage instructions and use case examples for primary sources. I think this page is the best place for that kind of content. Blue Rasberry (talk) 19:15, 19 July 2016 (UTC)

Bluerasberry, I'm not sure that I agree with the move since this essay is about identifying and using secondary sources in addition to identifying and using primary sources. Flyer22 Reborn (talk) 21:58, 19 February 2017 (UTC)
Agree. Bluerasberry and Flyer22 Reborn, this should be moved back. The page isn't only about primary sources. SarahSV (talk) 22:21, 19 February 2017 (UTC)
SlimVirgin Could either or both of you respond to the idea that using "secondary sources" is the default norm in Wikipedia and implied when sources are discussed, when this page seems to be talking about the different and special case of using primary sources? I can agree that this page includes more than discussion of primary sources, but I feel that this page is different from almost all other pages because it discusses primary sources when that is not the norm.
An analogy would be calling this page "Identifying and using a special sort of source". Technically it is correct to say "Identifying and using the usual sorts of sources and a special sort of source", but I feel that the second title is a bit misleading because this the focus of this page is the special case, and this is not the place to learn about the norm. What do you think? Blue Rasberry (talk) 03:13, 21 February 2017 (UTC)
I don't feel strongly about the title, but I doubt your assertion that secondary sources are "the norm" (in the statistical sense of being the most common in actual use, rather than the theoretically preferred type; check any current politician's or celebrity's article for the number of primary sources in use), and I think that WP:NOTGOODSOURCE, which is mostly about secondary sources, is one of the more popularly cited sections. Either of those could be reasons to move the page back. WhatamIdoing (talk) 21:43, 22 February 2017 (UTC)
Bluerasberry, I simply think that the previous title is more accurate. Flyer22 Reborn (talk) 15:05, 25 February 2017 (UTC)
I can confirm that many Wikipedia articles use primary sources. When I referred to "the norm", I meant to refer to Wikipedia guideline pages like WP:RS, which I feel talk about secondary sources as the norm and to not anticipate primary sources. Like for example, when notability guidelines talk about citing sources, they want a certain number of secondary sources and not for example self-published biographical information and accomplishment lists as commonly seen in politician and celebrity articles.
I still fail to understand why the page should be titled in such a way that suggests that this would be a place for people to come to learn about secondary sources. If someone wanted information about secondary sources, I think that I would send them to WP:RS or almost any of the other help pages talking about reliable sources, and if someone wanted information about primary sources, then I would send them to this page. Do you not share my perception that this page is unusual for addressing primary sources, and that in general, help pages like WP:RS talk about secondary sources? Blue Rasberry (talk) 15:24, 25 February 2017 (UTC)
As I said, I don't feel strongly about the title. Either way is okay with me. I agree that before I started this page, there was very few pages explaining primary sources (and most of what we have on secondary sources is an exhortation to use them, rather than practical information on how to identify them).
(Notability cares more about independence than historiographical classification. We could delete a significant number of articles on businesses and living people if the existence of two true secondary sources were actually required.) WhatamIdoing (talk) 04:11, 26 February 2017 (UTC)
I am willing to support the judgement of any new person to join this conversation, if that path to resolution would satisfy others involved. This would be like WP:3O. Otherwise, anyone else currently participating could propose their own path to resolution of this. I care a little but this is not worth so much conversation. Blue Rasberry (talk) 15:03, 26 February 2017 (UTC)

Primary sources in organizations

Previously the "primary sources from businesses" example said that it was okay to make some limited promotional claims if they were true. The example was, "sells more products than others in the region". I do not ever recall a time when this was allowed. I removed this.

Some people find it controversial to allow any primary sources from businesses, but I just listed some facts commonly taken from primary sources which I think usually are allowed. See my edit. The facts that I listed are these -

  • annual revenue
  • number of staff
  • physical location of headquarters
  • status as a parent or child organization to another

As a further explanation -

  • Having the annual review of an organization differentiates small organizations with a low annual budget from massive corporations with huge budgets. It is useful context to provide some measure of money, and this information often is only in an annual report
  • similarly, number of staff can be a measure of the size of an organization. Sometimes on Wikipedia it is difficult to differentiate companies with no staff from those with hundreds. List the number outright when it is available in primary sources.
  • Knowing the location of a company is useful like the nationality of a person. It can be misleading for multinational, multi-site organizations but in many cases country matters and GPS coordinates are welcome
  • If one organization is controlled by or controlling others then it is useful to note relationships, in the same way that primary sources for biographies often note family relationships not otherwise established in secondary sources.

All of these are grey area but in my experience this information is routinely included from primary sources. Furthermore - this sort of information is often desired to be hidden by many companies, which limits the potential for abuse that is common about sharing other organizational information. Blue Rasberry (talk) 19:33, 19 July 2016 (UTC)

Photos on Commons are primary sources

It is a quirk of Wikipedia to need to address this but photos on Commons are typically original research without verification and primary source material included in Wikipedia articles.

I just added a statement that this is okay. I am not sure where or how this has been addressed elsewhere, but now it is here. Blue Rasberry (talk) 19:35, 19 July 2016 (UTC)

All citations removed

I just removed all the citations to external sources for this article. There were 5, and they seemed to all go to university library websites describing the difference between primary and secondary sources.

4 of the links were dead. I did not bother to try to find their updates, archived version, or any replacement. One link was alive, but since it is published in the outdated format of a 1990s website, it is obvious that it is obsolete content.

This is an essay which will be used mostly by experienced Wikipedians. I have regularly seen young Wikipedians, high school and early college, easily outthink university librarians regarding the difference between primary and secondary sources from the perspective of good practice in Wikipedia editing. At this point, I think easier to understand, more comprehensive, and more applicable "primary versus secondary" explanations can be found in Wikipedia's own editing guides than in any homemade university writing center guide. This essay is kind of old - I also removed notes which said, "Wikipedia is not the real world", which is Wikipedia community jargon from the time when it was imagined that Wikipedia is a fringe community. Nowadays Wikipedia has its own extensive in-house documentation on these things.

If anyone finds a good external link, share it, but I think these are not necessary. Blue Rasberry (talk) 19:48, 19 July 2016 (UTC)

Unfortunately, not only is removing these citations (dead links or no) extremely unhelpful for readers who may wish to check the sources, but quoting a non-free source without any citation or attribution is likely to be a copyright violation (see Wikipedia:Non-free content criteria). I've suggested a possible way to improve sourcing for this section under § WP:PRIMARYNEWS below. —Sangdeboeuf (talk) 18:52, 6 February 2017 (UTC)
Sangdeboeuf Confirmed - if I had this to do over again I would have deleted the content along with the citations. I was careless about that. I will comment about changes below where you made a proposal. Blue Rasberry (talk) 13:43, 7 February 2017 (UTC)

"The first published source for any given fact is often considered a primary source."

"The first published source for any given fact is often considered a primary source" is not a true statement - it should be "often" not always. Disagreement?

A very common case is the publishing of research analysis (secondary source) while keeping the original study data unpublished. The presumption behind the above statement is that primary sources have to be published before secondary sources are published, and this is not correct. Secondary sources can be based on unpublished primary sources, leaving the reader to wonder if it is possible to find the primary source elsewhere.

Perhaps this entire sentence could be omitted just because it makes a sweeping debatable generalization. Blue Rasberry (talk) 10:48, 22 August 2016 (UTC)

I don't disagree with saying that it's "always", but it seems to be one of those things that depends upon the field. Regardless of whether such a source actually "is" primary (according to whatever definition is preferred by any given speaker), the fact remains that on Wikipedia such a source "should be treated as" primary (i.e., used carefully, without giving too much weight to it). WhatamIdoing (talk) 19:30, 27 October 2016 (UTC)

Illustration

I used a familiar example. It was replaced with a copyrighted work that is hardly known.

Can the example be something that can be seen in Wikimedia Commons? Wiki favors free content. I do not understand the bit about removing pictures to increase diversity, and think that if art is mentioned at all then it should be shown. Blue Rasberry (talk) 10:51, 22 August 2016 (UTC)

I'm not sure that there's any advantage to showing the art work (we don't give excerpts from the other primary sources, after all), especially when the title itself provides all of the description that's needed to understand the example. I also have a bias in favor of using examples that don't conform to the "white male as default", especially when the examples are relatively unimportant.
I'm also not sure that we're helping users optimally by showing a picture of antique travel diary. A modern blog or a travelogue in a magazine is equally a primary source. WhatamIdoing (talk) 19:46, 27 October 2016 (UTC)