Wikipedia talk:Identifying and using primary sources

Sampling of university-based sources that address the question:

• "A newspaper article is a primary source if it reports events, but a secondary source if it analyses and comments on those events."
• "Characteristically, primary sources are contemporary to the events and people described [e.g., like a newspaper article on a current event]... Examples of primary sources include...newspaper ads and stories. In writing a narrative of the political turmoil surrounding the 2000 U.S. presidential election, a researcher will likely tap newspaper reports of that time for factual information on the events. The researcher will use these reports as primary sources because they offer direct or firsthand evidence of the events, as they first took place."
• "There can be grey areas when determining if an item is a primary source or a secondary source. For example, newspaper journalists may interview eyewitnesses but not be actual eyewitnesses themselves. They also may have completed research to inform their story. Traditionally, however, newspapers are considered primary sources…. Examples of common primary source formats can include...contemporary newspaper articles…. Newspaper articles, although often written after an event has occurred, are traditionally considered a primary source…. "
• "Examples of primary information: A current news report that is reporting the facts (not analysis or evaluation) of an event."
• What are 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."

Which "university-based sources" is this section quoting from exactly? —Sangdeboeuf (talk) 01:49, 28 January 2017 (UTC)

WhatamIdoing probably knows. Flyer22 Reborn (talk) 07:17, 29 January 2017 (UTC)

I see now where the citations were removed. The source pages for the first and fifth quotations no longer exist, even in archived form, and the fourth quotation is taken out of context; the source also names "a newscaster's commentary on the day's events" and "Articles from magazines, journals, newsletters, newspapers, etc." as types of secondary sources.^[1] I would suggest trimming these parts and rewriting the section, with references, to more accurately represent the intent of the sources, for example:

Several academic research guides name contemporary newspaper accounts as one kind of primary source.^[2] Other university libraries address newspaper sources in more detail, for instance:

• "In the humanities, age is an important factor in determining whether an article is a primary or secondary source. A recently-published journal or newspaper article on the Brown v. Board of Education Supreme Court case would be read as a secondary source, because the author is interpreting an historical event. An article on the case that was published in 1955 could be read as a primary source that reveals how writers were interpreting the decision immediately after it was handed down".^[3]
• "Characteristically, primary sources are contemporary to the events and people described [...] In writing a narrative of the political turmoil surrounding the 2000 U.S. presidential election, a researcher will likely tap newspaper reports of that time for factual information on the events. The researcher will use these reports as primary sources because they offer direct or firsthand evidence of the events, as they first took place".^[4]
• "There can be grey areas when determining if an item is a primary source or a secondary source. For example, newspaper journalists may interview eyewitnesses but not be actual eyewitnesses themselves. They also may have completed research to inform their story. Traditionally, however, newspapers are considered primary sources. The key, in most cases, is determining the origin of the document and its proximity to the actual event".^[5]

—Sangdeboeuf (talk) 20:27, 6 February 2017 (UTC)

WhatamIdoing and Bluerasberry, any thoughts on what Sangdeboeuf stated above? Bluerasberry, Sangdeboeuf linked to a change you made. Flyer22 Reborn (talk) 07:23, 7 February 2017 (UTC)

I say replace the old text with this new text. The ideas are clearer and the citations are better. This is written for the humanities but it works. Ideally the quotations could be replaced with free text but that is not urgent. Blue Rasberry (talk) 13:47, 7 February 2017 (UTC)

The proposal is okay, but I think that the simplicity of the first one ("A newspaper article is a primary source if it reports events, but a secondary source if it analyses and comments on those events") is extremely helpful to editors who are trying to do the right thing. In a similar vein, I recommend leaving out the bit about eyewitnesses, because it's likely to mislead editors by conflating independence with primary-ness. It doesn't matter whether the reporter saw the car wreck with his own two eyes; the three-sentence newspaper story about the car wreck last night is still a primary source for the fact that a car wreck happened last night. WhatamIdoing (talk) 23:13, 7 February 2017 (UTC)

Yeah, I agree that we should retain that first sentence from the original piece. Flyer22 Reborn (talk) 00:08, 8 February 2017 (UTC)

The part about eyewitnesses is simply an example of the "grey areas" regarding primary vs. secondary that the source is talking about. Without it, the quotation wouldn't make much sense.

I have no problem with including the statement A newspaper article is a primary source if it reports events, but a secondary source if it analyses and comments on those events – but if this is a quote from a copyrighted source, then that source would need better attribution than a dead URL. —Sangdeboeuf (talk) 03:42, 8 February 2017 (UTC) (updated 00:28, 15 February 2017 (UTC))

Update: Pending further input, I have inserted the text as it appears above. —Sangdeboeuf (talk) 19:18, 10 February 2017 (UTC)

That quotation doesn't "need" better attribution than a bibliographic citation that includes a dead URL. This isn't a Wikipedia article, so WP:CITE doesn't apply (according to WP:POLICY itself).

One of the problems with the "grey area" quotation (which I added a few years ago) is that some editors read "journalists may interview eyewitnesses but not be actual eyewitnesses themselves" as the example and jump to the conclusion that everything except an actual eyewitness news article is always a secondary source – including "the journalist copied stuff straight off the police blotter", "the journalist interviewed the police officer about the mayor's arrest", "the journalist mindlessly repeated the facts that the preacher claimed about last night's ice-cream fundraiser at the church", "the journalist copied two sentences out of a press release that says WhatamIdoing's Gas Station has changed its opening hours", and all of the other primary stories that appear in a typical newspaper.

I think that the shorter quotation is clear enough without the intervening example, and the issue of eyewitness news is already addressed in another section. WhatamIdoing (talk) 07:55, 12 February 2017 (UTC)

WP:CITE applies to all copyrighted material in Wikipedia, even in information pages (see Wikipedia:Non-free content criteria). Specifically, the quotation I mentioned needs attribution unless it was explicitly published under a free content licence. The existing citation doesn't identify the source enough for readers to find the source themselves. Without this, it's not possible to verify the quotation and establish the context in which it was written – it could be over-simplified or otherwise misrepresent the intention of the author. —Sangdeboeuf (talk) 19:35, 12 February 2017 (UTC)

That's not actually true. CITE says (second complete sentence) "Wikipedia's Verifiability policy requires inline citations for any material challenged or likely to be challenged, and for all quotations, anywhere in article space." This is not "in article space", and therefore CITE does not apply. As I said earlier, POLICY, specifically the WP:NOTPART section, says that content policies don't apply to these pages. Look for the sentence that says "It is therefore not necessary to provide reliable sources to verify Wikipedia's administrative pages...".

Copyright status is irrelevant for this; we cite things in articles to show that it's verifiable, and everywhere to avoid plagiarism. But, in this instance, it doesn't actually matter how complete the bibliographic citation is: "readers" aren't going to be able to find this source themselves, because the library took that page off their website. Even if we find the original title for the article, people are still not going to be able to read the page. You are pretty much going to have to trust that if I were lying about the contents of that page, that someone would have mentioned it in the archives before now. WhatamIdoing (talk) 00:36, 13 February 2017 (UTC)

Please have a look at Wikipedia:Non-free content criteria, which is policy:

Articles and other Wikipedia pages may, in accordance with the guideline, use brief verbatim textual excerpts from copyrighted media, properly attributed or cited to its original source or author (as described by the citation guideline), and specifically indicated as direct quotations via quotation marks, <blockquote>, or a similar method.

While certain pages are exempt from this policy, project pages such as this one are not among them. Indeed, it makes little sense to argue otherwise; US copyright law doesn't care whether a page is an article or an information page; they are both equally accessible to the public. The Wikimedia Foundation's licensing policy states that

All projects are expected to host only content which is under a Free Content License, or which is otherwise free as recognized by the 'Definition of Free Cultural Works' [...]

The only exemptions to this policy are those covered under WP:NFCC. Infringing copyrights is also prohibited under the Foundation's Terms of Use. —Sangdeboeuf (talk) 14:36, 13 February 2017 (UTC)

As to the second point, this is asking not just me, but every reader of this page, to trust that the James Cook University quote is accurate and fairly represents the author's view. Since the Troy University quote was taken out of context as I mentioned, and the Indiana University quote (beginning with "Characteristically..") was somewhat oversimplified, with editorializing added (see difference between the two versions above), both of which were originally added at the same time as the JCU quote, I think it's reasonable to ask for some better verification of the latter quote, or else to remove it. —Sangdeboeuf (talk) 23:18, 14 February 2017 (UTC)

Despite what I wrote above about not having any problem with including it, the wording of the quote might be confusing; it's not clear what the difference is between "reporting" on events and "analysing and commenting" on events. Most of the other sources I've cited say that secondary sources analyze and comment on other sources, not events. I think the information I just added based on Yale's comparative literature guide covers these points more clearly, so the JCU quote isn't really necessary. —Sangdeboeuf (talk) 00:28, 15 February 2017 (UTC)

Update: I removed it here. —Sangdeboeuf (talk) 00:51, 19 February 2017 (UTC)

If an editor seriously cannot grasp the difference between reporting on an event and analyzing or commenting on said event, then WP:Competence is required and that editor needs to find something else to do. If you'd like, we can provide editors with a link to Analytic journalism to help them figure it out.

More relevantly, this brief quotation has been in the page for years now, and nobody has ever expressed any confusion over it. It is, of all the quotations in that section, the one that seems to have been most useful to editors. It should remain here, even though you don't like it. WhatamIdoing (talk) 21:52, 22 February 2017 (UTC)

Let's not jump to any conclusions. I happen to like the quote fine; I just don't believe it represents informed scholarly opinion on the subject, and without a reliable citation, I don't see any reason to change that belief. I see that the quote was added back.

As for the quote remaining on the page for years without comment, that doesn't prove anything. Maybe not many people bother to read this information page in the first place. Also, many outright hoaxes have gone undetected on Wikipedia for years at a time.^[6][7] I'm aware that project pages are not subject to the same content standards as the rest of the encyclopedia. Nevertheless, it shows that false or misleading claims can and do persist until someone takes the trouble to correct them.

Regarding "analyzing and commenting" vs. "reporting", I don't think the distinction is so clear-cut. Consider a hypothetical breaking news story that begins, "The festive spirit of the season was interrupted by a scene of chaos this afternoon, as frantic pedestrians fled the path of a careening vehicle..." Do the words "festive spirit", "frantic", and "scene of chaos" represent factual reporting or editorial commenting on this hypothetical event? —Sangdeboeuf (talk) 23:16, 24 March 2017 (UTC) (edited 15:56, 30 January 2018 (UTC))

Considering these three factors: the questionable sourcing, the fact that secondary sources comment on other sources, not events, and the existing summary description of secondary sources based on Yale's comparative literature guide, I suggest removing the JCU quotation again – it's both questionable and unnecessary. —Sangdeboeuf (talk) 23:16, 24 March 2017 (UTC)

One more point: I don't think it matters how "useful" a statement appears if it's flawed to begin with; confusing events with sources seems a pretty important misunderstanding of the role of secondary sources. —Sangdeboeuf (talk) 07:58, 4 April 2017 (UTC)

I still agree with WhatamIdoing that this is helpful, but let's see if she has anything to state about your latest points. Flyer22 Reborn (talk) 00:08, 5 April 2017 (UTC)

• I don't think that it's possible to completely "represent informed scholarly opinion" in any 23 words. (It could, however, be done in two: "It's complicated".) But I do think that this is a useful (=the primary duty of an explanatory advice page) first approximation of what's happening. It is true that it's an oversimplification; however, it's an oversimplification that sets editors on the right path. To be candid, after producing this huge wall of text, you'll convince me otherwise when you produce actual examples of multiple editors getting it wrong while citing this sentence, and not a moment before then.
• The "festive spirit" example is irrelevant, because an encyclopedic summary should not normally need to consider such a trivial detail.
• Neither Yale's comp lit guide – nor any source from any single academic field – is the arbiter of the One True™ Definition of secondary sources. But if I were going to pick a single academic field to decide how to classify and use journalism, it would frankly not be the field of comparative literature, which ignores questions such as "Can we learn anything from this short news report?" in favor of questions like "Does this book actually 'count' as proper literature, or is it just unimportant junk?" and "What universal human truths are conveyed in this work?" WhatamIdoing (talk) 01:20, 6 April 2017 (UTC)

Sorry for reopening this old thread. But I see the shortcut WP:PRIMARYNEWS being often misunderstood. Many editors seem to interpret it as "News is always primary" which is definitely not the goal of the paragraph. As User:WhatamIdoing said "it's complicated" is the best assessment of the topic. Maybe the text could be clarified by starting with a sentence explaining that "it depends" and the shortcut could be changed to clear things up? Maybe something like "WP:NEWSPRIMARY?" or "WP:NEWSPRIMARYORSECONDARY" (not much of a shortcut :-P) {{u|Gtoffoletto}}  ^talk 16:27, 1 April 2024 (UTC)

@Gtoffoletto, it sounds like you've identified another candidate for the list at Wikipedia:UPPERCASE.

I don't think that changing the shortcut will solve the problem. This is mostly because we won't "change" it (it's in hundreds of talk pages and edit summaries, so RFD won't want to delete it). Instead, we'd just "add another shortcut", and the familiar old one will continue to be used, and editors will continue to fail to read the page. Because part of the problem is that we teach Wikipedia's ruleset through a telephone game (because Wikipedia:Nobody reads the directions), I don't think that changing the content of the section would actually help. They're almost never reading it now, and they will almost never read it after any changes are implemented.

You could provide these editors with a link directly to Wikipedia:Identifying and using primary sources#Examples of news reports as secondary sources, which might broaden their understanding. WhatamIdoing (talk) 16:54, 1 April 2024 (UTC)

One needs historical knowledge to understand propaganda (starting from ancient emperors to Hitler and Stalin). This Wikipedia isn't for Western academicians only.

I find the Nazi propaganda present in many pages digusting.

Propaganda pictures are acceptable in Nazi propaganda, where they are explained, not as an image of Nazi Germany or Adolf Hitler or some other idol.

The same pictures are being used several times in connected pages. "repetitio est mater studiorum" which means here "Wikipedia indoctrinates you."

Xx236 (talk) 14:11, 10 February 2017 (UTC)

It's not clear what action is being suggested here. Problems with specific articles should be discussed on those articles' talk pages or at the related WikiProject. However, please also remember that Wikipedia is not censored, so simply finding some material "disgusting" is not adequate grounds for removing excluding it. —Sangdeboeuf (talk) 19:42, 10 February 2017 (UTC)

Propaganda pictures are useful in many kinds of articles. An educated understanding of the subject will know what all the sides thought of it: the image that the Nazis chose to project for themselves, the image that their enemies used of them, the images that we now use, etc. There's nothing wrong with explaining what those images represent in those other articles: "Official portrait of <person>" or "Photo of military parade by famous Nazi propagandist, Watts Hisface" or "Wartime poster by <enemy>". WhatamIdoing (talk) 08:02, 12 February 2017 (UTC)

I would like to get some clarification on this guideline "Wikipedia articles should be based on reliable, published secondary sources and, to a lesser extent, on tertiary sources and primary sources." because it is not easily applied to the basic sciences. Here's why.

The best sources in basic science are peer reviewed research articles in reputable journals i.e. primary sources by the WP's definition. The basic science equivalent to secondary sources, i.e. reviews of a developing research field, often just contain a brief summary of the primary research and can be colored by the authors' own view point. This often makes them a worse source than the primary research and it is different to medical research where meta-analyses and systematic reviews are often superior to primary research. But these types of publications are often not possible in basic research since at the bottom of the research tree it does not make sense to repeat the analysis on a small root many times to get enough data for a meta-analysis.

So, while I agree that secondary sources are best for history and medical research, I think this is not the best guideline for basic research. And not summarizing basic research would deprive the Wikipedia of most of its material in the sciences. — J.S.^talk 10:51, 15 March 2017 (UTC)

You wrote: "And not summarizing basic research would deprive the Wikipedia of most of its material in the sciences.". I assume that when you wrote "basic research" you meant "primary sources". Is that what you meant? If so, what about the loads and loads of actual secondary and tertiary sources that are published all the time? I see no basis in reality for this claim, that generally using secondary and tertiary sources and using primary sources rarely and with care would deprive WP of anything. Jytdog (talk) 19:18, 15 March 2017 (UTC)

This discussion is probably more appropriate at WP:SCIRS or WP:MEDRS btw. Jytdog (talk) 19:23, 15 March 2017 (UTC)

Hi J.S., and thanks for the comment. Are you talking about basic research (as in what happens at SLAC) or the basic sciences (as in biology, physics, etc.)? WhatamIdoing (talk) 01:28, 6 April 2017 (UTC)

Hi Jytdog and WhatamIdoing, thanks for getting back to me. Yes, I think it's problematic to cut down on primary sources and promote this as a guideline or even a rule of the Wikipedia. Secondary sources have an element of Chinese whispers. Information invariably gets altered, even distorted at times. You may say a news piece on a primary research publication is secondary and therefore more desirable as a Wikipedia source. I would say it's not. Better to go to the source like a detective and not rely on hearsay. Of course secondary sources often make the information easier to take up because they summarize and evaluate. So if there is a good review/newspiece/press release/highlight, why not. Often there isn't. The accelerator article is a good example. Many pieces of information entirely appropriate for the article do not have a good secondary source. This article would be much poorer without primary sources. I completely agree that data shouldn't be published in the Wikipedia but once it's peer reviewed, I think it's the richest source of information for the Wikipedia. When a new drug is discovered, double-checked, and published in a reputable journal, should we really wait a year until a review article comes out, or 5 until a meta-review is published? Or should we use the news piece of a journalist that understands the research less than many of the better Wikipedia editors? Let me know what you think. If this discussion is more appropriate for another page, feel free to move it. All the best, — J.S.^talk 15:12, 7 April 2017 (UTC)

Thanks for your note. What you write, is the way that many scientists try to approach working here, in Wikipedia, when they first come here. This is a discussion that many of us have had with people like yourself, many times. But scientific writing here is not like scientific writing elsewhere. This is explained somewhat in WP:EXPERT which you might find helpful. The emphasis on secondary sources exists for many, many reasons. Those reasons are pretty easy to explain and pretty easy to understand, if you are willing to listen. Let me know. Jytdog (talk) 17:49, 7 April 2017 (UTC)

I have a couple of separate thoughts about this:

• You sound like you're the target audience for WP:10SIMPLERULES. You probably also want to take Jytdog's suggestion and look at WP:SCIRS or WP:MEDRS, if you haven't already, since they deal with scientific sources in much greater detail. I've included only a brief paragraph on scientific literature here, because editors really need to look at the longer one.
• About whether you can cite a primary source:
  • The "guideline" that you quote at the start of this section is a widely supported "policy". WP:The difference between policies, guidelines, and essays is subtle (and sometimes arbitrary), but this particular statement is a widely accepted best practice that everyone should follow appropriately.
  • The net effect of this particular page is to walk that policy statement (slightly) back towards common sense. We have some editors who interpret the statement that "articles should be based upon" secondary sources as "Thou must not cite a primary source, lo, not even for the smallest sentence in an entire article". When the policy says that "articles should be based upon" secondary sources, the policy means that a majority of the content and the major themes for an article should be cite-able with reliable secondary sources. It doesn't mean that you can't ever cite a primary source for anything.
• About whether editors should prefer the primary literature:
  • As every scientist knows, especially in the range of biology/medicine/psychology, there's a peer-reviewed journal article that supports just about any viewpoint. The replication crisis in behavioral sciences is well-known, even for widely trusted studies in the most reputable journals, and some days, it seems that peer-reviewed journal articles about altmed treatments are just as (un)reliable as a sales brochure. So the odds of a Wikipedia editor adding accurate scientific information while citing a primary source is lower than we'd wish.
  • The secondary literature does not necessarily include news articles. A press release about a journal article is still primary. For the sciences, we tend to prefer peer-reviewed review articles and meta-analyses to either press releases or news stories (which often amount to little more than the press release anyway). You are correct that this can, in some instances, result in a delay. OTOH, that delay means that we're less likely to publish flawed results or information that can't be replicated.
  • When editors cite the primary literature, we tend to see a lot of unrepresentative or unimportant studies (e.g., people citing the one study that claims that bacon improves health, and ignoring the thousands that say it doesn't, or people citing the one study that's in the news this week). It's hard to end up with WP:DUE attention to the majority and minority viewpoints when you're reading this week's popular primary sources.

Because of these problems, we tend to get better results, on average, when we encourage editors to stick to the secondary literature. Yes, that can result in a lag between, say, "this disease is universally fatal" and "as of last month, exactly one person has survived this disease so far". There are costs to this choice. But the choice tends, on average, to result in stronger articles. WhatamIdoing (talk) 15:04, 12 April 2017 (UTC)

Wow you said that so, so well. Jytdog (talk) 15:15, 12 April 2017 (UTC)

Hi Jytdog and WhatamIdoing, thanks for expanding on the secondary-preferred guideline. Ultimately the most important quality factor is the writer. Primary sources are not always reliable but the same is true for secondary sources. Reviews are great because they save us time but bad because review authors know less about the discoveries that they sum up from a distance. Primary literature, on the other hand, is bulkier but also less error-prone and easier to verify because you can directly go to the source instead of hoping from one review which cited another review which finally badly cited the original experiment. Ultimately, it's in the hand of the editor to pick the best source and check it thoroughly before summing it up nicely for the Wikipedia. What will definitely lower the quality of almost any Wikipedia article is the dogmatic application of simplistic rules (no research articles ever) but I'm hopeful from your comments that there are sensible editors around and maybe even in the majority? — J.S.^talk 16:22, 12 April 2017 (UTC)

You are writing abstractly. WAID and I are both explaining to you how the community thinks about these issues - the consensus has existed in the community for a long time, and is broad and deep. You are free to ignore us and to ignore the reasoning behind the consensus that we are explaining to you (and the reasoning makes a great deal of sense in the context of working in Wikipedia, which is not like other places) but you will find that your edits get consistently reverted. If you need to bang your head against the wall for a while, so be it. Jytdog (talk) 16:28, 12 April 2017 (UTC)

J.S., I fully agree with you that dogmatic application of simplistic rules is bad for Wikipedia. I believe that there are many sensible editors around. I know that we don't always get it right – especially not always on the first try (that's why we have talk pages), especially not when we're busy or distracted (we're all humans) – but I think that people are trying, as best as they can, to do what's best for the encyclopedia, even if that means not dogmatically following The Rules™. WhatamIdoing (talk) 22:39, 13 April 2017 (UTC)

• "The best sources in basic science are peer reviewed research articles"? This is not true. Research articles are too focused. The best sources are broad-audience publications. For "basic science", these are popular publications. If the facts are disputed, it is not basic science. --SmokeyJoe (talk) 16:38, 12 April 2017 (UTC)
• The fact that JS oversees students who edit is a bad thing, with their approach to WP. See User:TüBioc. Jytdog (talk) 18:04, 12 April 2017 (UTC)

There are many cases in basic science where the primary source is a conference paper, before the actual work is done. Often enough, the facts change by the time the later paper is written. There are some documented examples, but I forget them now. In any case, one still has to be careful with science journal articles. Gah4 (talk) 00:26, 21 July 2021 (UTC)

I have been wondering whether ancient historians, like Thucydides or Plutarch could be considered as reliable secondary sources and hence rely on them to edit an article. As I understand, these are not self edited (but I am not really certain, we don't know how the publishing industry worked then!), they seem more or less independent (even though there was a little pro-Athenian bias of Thucydides) and these historians were not witnessing the events they describe at there work. But considering them as secondary sources, seems awkward. Secondary sources in history are usually books or scholarly journals, from the perspective of a later interpreter, especially by a later scholar. So, is there a formal WP policy or guidance relevant to this question? Τζερόνυμο (talk) 07:21, 4 December 2017 (UTC)

Primary. And you cannot cite them directly, per WP:SAYWHEREYOUGOTIT. You are not holding in your hands an ancient manuscript of Caesars Commentarii de Bello Gallico or whatever, so you have to cite what you're actually reading, something like: Caesar, Julius; McDevitte, W. A.; Bohn, W. S., trans (1869). The Gallic Wars. New York: Harper. p. 9. ISBN 978-1604597622. Retrieved 8 January 2017. {{cite book}}: ISBN / Date incompatibility (help)CS1 maint: multiple names: authors list (link) Generally speaking any source older than about 100 years should be treated as primary, and for news sources, any source close to the date of what's being addressed by the journalist should be as well, since it's just reguritation of off-the-cuff reactions by talking heads, maybe with some journalistic investigation on the basis of too little data. The problem with sources is that they reflect the undersetanding of the time in which they were written. A 1920 scholarly analysis of the global impact of what we now call World War I (then, the Great War) is not a secondary source, because that analysis has been vastly superseded by better analyses and by actual changes in what the effects were, and by more significant later events like WWII and the Cold War. An analysis of events by someone writing in 32 BCE or 524 AD is utterly primary, because – on top of the too-close-to-the-events problem – virtually no one made any effort whatsoever to produce neutral analysis before the advent of modern scientific writing and proper journalism.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  12:06, 4 December 2017 (UTC)

Re: "A 1920 scholarly analysis of the global impact of what we now call World War I (then, the Great War) is not a secondary source" - I would disagree... it both a secondary source and a primary source (at the same time). It remains a secondary source for it's analysis (although it might not be a good - or "useful" secondary source, because it probably is outdated, and has likely been superseded by more modern sources)... but it is also a primary source for an analysis of what scholarship said about the war, back in 1920.

Regarding ancient sources... Yes, they are considered primary... and there are lots of caveats and restrictions that apply to primary sources. As for whether you can cite them... SMcC is correct that you can not cite them directly (unless you are holding the original in your hands)... but you can cite them indirectly (by citing a modern translation). Primary sources do have their place, and can be cited... but see our WP:PSTS policy on how we should use Primary sources. There are restrictions. Blueboar (talk) 14:49, 4 December 2017 (UTC)

Thank you both for your responds. Is it possible to change the wording of Wikipedia:Identifying and using primary sources, so to clarify it even better? We are having a dispute in Greek WP 1 and some fellow Wikipedians seem to be more literalists than contextualists. According to them, it is clear that Thukydides, Plutarch and so on, are secondary sources coz they were not witnesses etc. Their argumentation ends, as expected like that: "if you want ancient historians to be considered as a primary and not as a secondary source, go chance the guideline". So...how do we fix this? Τζερόνυμο (talk) 20:52, 5 December 2017 (UTC)

Suggest you read our Primary source article ... it nicely explains the nuances of classifying sources. In the case of Plutarch and similar classical material, I will quote one important sentence from that article: "In some instances, the reason for identifying a text as the "primary source" may devolve from the fact that no copy of the original source material exists, or that it is the oldest extant source for the information cited." In other words... a source that may have originally been secondary (at the time it was written) can become a primary source, if the older sources it was based upon have disappeared. In that situation, even though the source may not have been an actual eye-witness account, it is as close to an eye-witness account as we can get today. Blueboar (talk) 22:07, 5 December 2017 (UTC)

Whether we call them primary or secondary or whathaveyou, ancient historians cannot be considered reliable, unless filtered through a modern scholarly source. Here's a great writeup: Wikipedia:Identifying_reliable_sources_(history). EEng 23:32, 5 December 2017 (UTC)

I tend to agree with EEng. Whether they "are" primary or secondary or even tertiary in any given system, you need to "use them" as if they are primary sources. WhatamIdoing (talk) 01:23, 27 September 2018 (UTC)

Eric Walberg has contributed to "the Palgrave encyclopedia of imperialism and anti-imperialism" where there's a short bio of him along with the bio of other contributors. Is that bio considered as a primary source? --Mhhossein ^talk 19:14, 13 March 2019 (UTC)

Hello there...? --Mhhossein ^talk 12:36, 28 March 2019 (UTC)

Almost always these are WP:SPS. EEng 14:32, 28 March 2019 (UTC)

I think what is missing on this page is how the context influences whether a source is primary or not.

Let's say you are writing about some ancient person. You use the Encyclopædia Britannica to verify the article. This is ok and not a primary source.

However, for the article Encyclopædia Britannica itself, the very same Encyclopædia Britannica is now a primary source.

Similar more subtle example would be a section "topic xy in popular science" where you write about how a specific topic is received in popular science. Popular science literature itself is now the primary source, as this is the subject of interest in your section. Optimally, you would need another source summarizing different popular science literature, then you can use this summary to verify your section. Of course this would be overkill for some articles, but technically this would be the correct practice to truly get unbiased information. --TheRandomIP (talk) 13:26, 11 September 2020 (UTC)

TheRandomIP, you're correct that context matters, and it's already in the page. See Wikipedia:Identifying and using primary sources#All sources are primary for something. WhatamIdoing (talk) 22:52, 14 September 2020 (UTC)

In an article about a science, between "the first publication of any idea or experimental result" and "Narrative reviews, systematic reviews and meta-analyses" there exists a large range of RS publications about follow-on results, replication, interpretation, and so forth which this criterion-description ignores. IMO, because such publications refer to and inherently comment upon prior research, these too can be cited as secondary literature so long as the editor is not creating a new synthesis (i.e. doing the work of a review or meta-analysis). Shouldn't the advice here recognize this?
Bn (talk) 14:47, 24 February 2021 (UTC)

@Bn, you are correct that this happens, and that this is a type of secondary material, but in the past WP:MEDRS authors felt that this material was often incomplete in a somewhat biased way (I only mention the prior research that is relevant to my hypothesis), and that it might be too confusing for most editors. That is why I didn't include it here. WhatamIdoing (talk) 22:15, 24 February 2021 (UTC)

Ah, yes. I well understand the issue with medicine and in a different way with topics that are politically disputatious, but unless stated otherwise this 'explanatory supplement' applies to all topics. Best to make the paragraph more inclusive for the general case, and caution editors about this intermediate literature for topics where it might be confusing for e.g. either of these reasons (and perhaps others).
Bn (talk) 16:15, 10 March 2021 (UTC)

This page already says that it's possible for a source to be a mix of primary and secondary material. I'm not sure that we really need to expand upon that. It doesn't seem to be a common source of practical problems for editors. WhatamIdoing (talk) 03:52, 14 March 2021 (UTC)

I do understand that often enough primary sources, even close to an actual event, can be wrong. But often enough, the subject is an actual documents, especially in government documents and standards. I would rather read the actual words from the Declaration of Independence, instead of someone else telling me what it says. Though also, sometimes explanations of the context are also important. Also for things like government standards documents, the actual wording, even if wrong, is usually more important. But as noted in the article, there are many cases, even for scientific journal papers, where the primary source is wrong. Gah4 (talk) 00:32, 21 July 2021 (UTC)

There is no category of source that is always error-free. When the goal is quoting a document, then the original source is authoritative for what was said, but not for whether it said something true. WhatamIdoing (talk) 20:59, 26 February 2022 (UTC)

In general, library reference landing pages targeted at students are not of high quality, as they are often copied and pasted from other sites and massively simplified for an undergraduate audience. So, as can be expected when relying on what are essentially freshman handouts, the ones cited here have serious issues.

One page cited, from Yale University, is (as the essay states) self-contradictory, listing newspapers/magazines as both primary and secondary sources. The library page is also rather slapdash and poorly written and does not do much, by itself, to support this essay's claim that there is a clean line delineating primary vs. secondary newspaper articles. A more clear picture can be found in Yale's Primary Source collection -- indicating, if nothing else, how Yale interprets its own guidelines:

"Like books, serials can function both as primary sources and secondary sources depending on how one approaches them. Age is an important factor in determining whether a serial publication is most useful as a primary or a secondary source. For instance, an article on slavery in a recent issue of the Journal of Southern History should be read as a secondary source, as a scholar’s attempt to interpret primary source materials such as ledgers, diaries, or government documents in order to write an account of the past. An article on slavery published in the Journal of Southern History in 1935, however, can be read not only as a secondary source on slavery but also—and perhaps more appropriately—as a primary source that reveals how scholars in the 1930s interpreted slavery."

This is saying two things: one, that primary vs. secondary as it applies to books and media articles is less a matter of what the sources are than how they are used; and two, that the distinction has to do with context and cultural factors, as anything written in the 1930s about slavery -- even a scholarly source that is doing some interpretation -- is going to be inextricable from a pre-Civil Rights Era perspective, and perhaps less useful for what it says about slavery than what its existence says about those scholars.

Meanwhile, the supposed James Cook University material does not, in fact, originate from James Cook University. If you actually read the "secondary sources" section, you will notice that it is cut off after "More generally, secondary sources...". The oldest version on the Internet Archive is cut off in the same place, suggesting that it was copied (poorly) from somewhere else. Googling the text turned up a lot of sketchy term-paper sites, but I believe I have found the origin: a textbook/pamphlet from Lovely Professional University. (The place where the James Cook page cuts off is right before some bullet points, which checks out if someone is hastily copy-pasting.) Skimming through this pamphlet, it seems of somewhat low quality; the sources it cites are this University of Illinois guide, which basically says what the Yale guide does ("it is important to consider by whom, how and for what purpose it was produced"), and something called "buzzle.com," which does not seem particularly reliable and is part user-generated. The irony of a Wikipedia essay about usable sources depending upon a plagiarized source is left to the reader.

The other university library pages' text has also been heavily copied-and-pasted across other university libraries, but it's hard to tell which was the original since most are are unsigned. Only the last page given (University of Michigan) lists its author, a graduate student. Basically, the whole thing is a mess, but at the very least the plagiarized ones need to go. Gnomingstuff (talk) 02:37, 28 March 2023 (UTC)

The problem to be solved is editors sincerely and genuinely believing that everything in a newspaper is a secondary source. This may happen because another editor told them that, because they started editing back when WP:PSTS said that anything "secondhand" was secondary (and a you reading a newspaper article written by a journalist who interviewed an eyewitness might seem like secondhand content), or because they didn't realize that secondary isn't just wikijargon for "reliable source".

What's on the page now does not violate our Wikipedia:Plagiarism guideline (because we gave credit to our source, even if the source might not have), and your analysis may be incorrect anyway. For example, you assume that the universities unfairly copied a hypothetical original document without giving proper credit, but it could be properly licensed text from a content service.

The distinction between what a source "is" and how an editor "should use" it is real, but perhaps too complex for this page. We're still working on the basics, like whether a newspaper article on the big game last weekend is a primary or secondary source. WhatamIdoing (talk) 00:38, 15 April 2023 (UTC)

An example text needs to be written about using Video games as a primary source. They are a good source to use for reference on writing a wiki article about it. The most one has to worry about with video games is what scenes are altered for different regions and if any parts have been changed so that the game can fit onto different consoles. 2600:1014:B077:2FA5:E183:CC33:E983:9A53 (talk) 03:38, 15 July 2023 (UTC)

Would you like to suggest a particular text? WhatamIdoing (talk) 20:29, 15 July 2023 (UTC)

I would suggest if the future written text will mention altered scenes then the writer should read the wiki article about video game censorship as some games are altered so a game will be able to release in a certain country. About the video games being changed to fit onto different consoles, that would also fit for games that are played on computers versus their video game console counterpart version. An example of a game looking different across a console generation is "Ghostbusters: The Video Game" 2009 release. The game was released for the PS2 and PS3, if you go look at the game you will see that the characters are styled differently for their respective console and there is even some gameplay differences as well. 2600:1014:B077:2FA5:E183:CC33:E983:9A53 (talk) 22:46, 15 July 2023 (UTC)

I wonder if this information might be more usefully placed in Wikipedia:WikiProject Video games/Sources or Wikipedia:Manual of Style/Video games. WhatamIdoing (talk) 23:42, 19 July 2023 (UTC)

What you are suggesting would be original research. Video games often tell a story, so they can be used a primary sources insofar as the plot, but explaining differences in censorship and graphics is not something the game does. Journalists do that. TarkusAB^talk/contrib 00:02, 20 July 2023 (UTC)

Seconding Tarkus. These differences also strike me as exceedingly minor and would not ordinarily warrant mention in a generalist encyclopedia article, even if true. Axem Titanium (talk) 16:36, 20 July 2023 (UTC)

I no longer remember what the source for the statement about legal studies not using the concept of tertiary sources, and I'm occasionally asked about it, so here are a couple of websites that may be useful for the curious:

• https://sia.libguides.com/c.php?g=521408#s-lg-box-11009371 – provides "Primary and Secondary Sources in Law", and has tertiary for other fields, but not for this one. Includes "Law reference books" as an example of a secondary source (e.g., a legal dictionary), which includes what would normally be considered tertiary sources.
• https://shsulibraryguides.org/c.php?g=699719&p=4963009 – names "legal encyclopedias, legal periodicals (law reviews), legal dictionaries, treatises, and digests" as examples of secondary sources in the field of legal studies.
• https://guides.loc.gov/law-secondary-resources – names six types of secondary sources, many of which would be considered tertiary in any other discipline:
  1. Definitions (dictionaries with legal terminology and terms defined by law)
  2. Legal Encyclopedias (multi-volume works about many different topics)
  3. Books & Binders (single- or multi-volume works about one particular topic)
  4. Articles (in-depth works about highly specific topics)
  5. Restatements (rules and examples from common law)
  6. Legal Directories (listings of people and organizations)

WhatamIdoing (talk) 00:33, 26 September 2024 (UTC)

WP:PRIMARYNOTBAD states 'Sometimes, a primary source is even the best possible source', goes against other policies such as OR and NPOV. Secondary sources are required to establish importance and justify inclusion. Traumnovelle (talk) 23:08, 14 February 2025 (UTC)

Given a choice between:

• a secondary source that accidentally misquotes the original, and
• quoting directly and accurately from the original source

which would you choose? If you pick the latter, then you have to agree that "sometimes" a primary source is the best possible source.

BTW, WP:NPOV barely mentions primary/secondary sources at all, and WP:PSTS (the only section in NOR to mention the distinction at all) does not require secondary sources for everything. We should not have whole articles WP:Based upon a primary source, but you are allowed to cite them. WhatamIdoing (talk) 23:28, 14 February 2025 (UTC)

Why can't you use the secondary source to establish weight and the primary source to correct a mistake? Although the primary source could have a typographical error and minor changes to a quote shouldn't be considered incorrect. I only ever see this policy linked to to justify bad use of primary sources. A primary source cannot establish importance/notability. Traumnovelle (talk) 00:54, 15 February 2025 (UTC)

Sure, primary sources are not appropriate for establishing importance/notability. However, they are often great sources for verifying the details once that importance/notability has been established by independent secondary sources. Our best articles use both. The hard part is understanding HOW and WHEN to appropriately use each. Blueboar (talk) 02:08, 15 February 2025 (UTC)

And this guideline is frequently cited in situations where secondary sources have not been provided. Traumnovelle (talk) 20:16, 15 February 2025 (UTC)

Um… this isn’t a guideline. It’s an explanatory essay. Blueboar (talk) 21:19, 15 February 2025 (UTC)

Well, I guess that changes a lot. Apologies for ignoring the banner at the top. Traumnovelle (talk) 21:22, 15 February 2025 (UTC)

WP:JUSTANESSAY is not a strong argument. To the extent that this page provides true and accurate information, the label at the top is irrelevant.

If you are in a dispute, and someone says 'Sometimes, a primary source is even the best possible source', I suggest that your reply will be more effective it it sounds more like "Yes, well, that might be true in theory, but this isn't one of those times" than like "That's 'just' a widely supported essay". WhatamIdoing (talk) 21:35, 15 February 2025 (UTC)

I won't reference specific disputes, but every time I have seen it used it is to justify inclusion of content I would characterise as either UNDUE or OR. Traumnovelle (talk) 21:37, 15 February 2025 (UTC)

If someone is incorrectly pointing to this essay to justify adding UNDUE or OR material, I doubt they have actually read the essay. This essay attempts to explain how and when to use various sources correctly. Blueboar (talk) 22:12, 15 February 2025 (UTC)

You're correct, sorry this has been a bit of a waste of time. I don't have any actual suggestions on how to improve the wording of the section to avoid bad interpretations. Traumnovelle (talk) 22:18, 15 February 2025 (UTC)

The last sentence of the An article about a person section says:

Many other primary sources, including birth certificates, the Social Security Death Index, and court documents, are usually not acceptable primary sources, because it is impossible for the viewer to know whether the person listed on the document is the notable subject rather than another person who happens to have the same name.

But WP:BLPPRIMARY says:

Do not use trial transcripts and other court records, or other public documents, to support assertions about a living person. Do not use public records that include personal details, such as date of birth, home value, traffic citations, vehicle registrations, and home or business addresses.

Am I missing something about why the text here says "usually," when the policy says to never use public documents? FactOrOpinion (talk) 20:46, 17 April 2025 (UTC)

BLP is for living people only, you can use court documents for dead people (but you shouldn't typically). Traumnovelle (talk) 20:55, 17 April 2025 (UTC)

I assume that the editors who worked on this page understood that there are inevitably “occasional exceptions” to all our policy and guideline “rules” - sooner or later there is bound to be a situation we didn’t think about when we wrote the “rule”. It is better to avoid “never” and to instead say “usually don’t” instead. That way you don’t have to rewrite “the rule” when those occasional exceptions crop up. Blueboar (talk) 21:02, 17 April 2025 (UTC)

Blueboar, the problem is that then you have newer editors arguing that it only says "usually," so it's really OK to use this court document that they want to use as a source for a living person (which happened today, leading me here to read the text). Personally, I'd rather say that it's never acceptable for living persons (recognizing that there may be an IAR exception that gets OKed through consensus), and that they're usually not acceptable for people who have died.

Traumnovelle, thanks for pointing that out, I'd totally overlooked that. I think it would be good to distinguish between people who are/aren't living in this text. FactOrOpinion (talk) 21:39, 17 April 2025 (UTC)

Anyone who is arguing that an explanatory essay says "usually", and so that overrides a core content policy saying "Do not" has a losing argument, and they probably know it. WhatamIdoing (talk) 21:43, 17 April 2025 (UTC)

It's simpler than that.

• First it says that some primary sources are acceptable: "The person's autobiography, own website, or a page about the person on an employer's or publisher's website ... can normally be used..."
• Then it says that some other primarily sources are not acceptable: "Many other primary sources...are usually not acceptable...".

The "usually" applies to "Many other primary sources", not to the example of birth certificates. (Though, as Blueboar notes, there are inevitably "occasional exceptions", and you can use even a birth certificate if, e.g., it is published by the person in question.) WhatamIdoing (talk) 21:42, 17 April 2025 (UTC)

The BLPPRIMARY bit is about BLPs. The first part is not. Don’t think you would get a social security death index about a living person. PARAKANYAA (talk) 21:44, 17 April 2025 (UTC)

PARAKANYAA, yup, I wasn't thinking. The biographical content that I work with usually involves living persons, and the exchange that led me here was on the talk page of an article about a living person (Deportation of Kilmar Abrego Garcia), so I just had "living person" in mind. I decided to walk away from the exchange, as this other editor wants to use totally unacceptable court docs (posted to a Dept. of Homeland Security webpage as part of a smear campaign to try to justify their illegal act).

Thanks to everyone for the quick clarifications. FactOrOpinion (talk) 22:24, 17 April 2025 (UTC)

I came here to take a look, essentially, at WP:SECONDARYNOTGOOD and WP:PRIMARYNOTBAD, to resolve a niggling doubt that has been on my mind. I have seen cases where secondary source simplifies information from a primary source, which introduces some small inaccuracies. I've seen previous discussions, resulting from primary and secondary sources giving different statistics where the "winning" argument seemed to hold the position that secondary sources automatically took precedence over primary ones. And another recent discussion (on a tangentially related subject) I had once again brought to mind that doubt I had. Articles on tornadoes commonly directly cite damage surveys (usually conducted by a division of the affected country's government, such as the U.S. National Weather Service) for information such as the rating. One issue I've seen come up is that secondary sources will not uncommonly conflate ratings on the Fujita scale and Enhanced Fujita scale. In some cases this appears to be a simplification the secondary source makes to avoid confusing a general public audience. In other cases it may be an error made on the part of a news reported who is not familiar with the subject matter. I once even contacted the author of a local news article with a correction on this matter. Still, I am concerned about the possibility of someone trying to enforce a preference for secondary sources in such an instance. TornadoLGS (talk) 22:23, 13 November 2025 (UTC)

It's possible that your biggest obstacle might be this bold edit to NPOV by @Bon courage, but I don't think even that is an insurmountable barrier. Articles should be WP:Based upon sources that are independent of the subject. For weather, that's pretty easy: you're never going to encounter a source written by a tornado, after all. Articles should also be WP:Based upon sources that are secondary (e.g., a comparison of several tornadoes or an analysis of one tornado [e.g., articles asking why this small tornado was disproportionately destructive?]).

But:

• Secondary sources can be bad sources overall and/or in some particular point. Use Wikipedia:Editorial discretion to cite sources you believe are better. (You can't control other editors, but you never have to add a bad source yourself.)
• Not every single source in an article has to be secondary.
• Whenever possible, you should cite sources for what they intend to say. That sometimes means replacing outdated terms, fixing typos, etc. If there is a dispute, then it's easiest if you're dealing with a sentence where the mistake is very clear, e.g., "According to this linked source, this tornado was an EF2 on the Fujita scale". Then you can tell editors, "Oh, it looks like they have a simple error/typo. 'EF2' always means the Enhanced scale, and look – the linked source they're citing specifies the Enhanced scale. So we should specify the Enhanced Fujita scale in this sentence, because that's obviously what they meant."

I will admit that this last item is not very popular with some of our more rule-bound editors, but it's the right thing to do. If you get stuck in a dispute, though, it may be faster to find a new source than to argue about it. Alternatively, fairly often the alleged secondary sources turn out to be WP:PRIMARYNEWS, and can be discarded, or at least put on an equal footing with the better source, anyway. WhatamIdoing (talk) 01:55, 14 November 2025 (UTC)

News reports are primary sources not secondary, it'd be fine to cite the weather service itself over a news report here. Traumnovelle (talk) 02:09, 14 November 2025 (UTC)

Yeah I was thinking in cases where sources switch the F and EF ratings. For instance, the 1999 Bridge Creek–Moore tornado was rated F5, but some secondary sources refer to it as EF5, even though the EF scale is not applied to pre-2007 US tornadoes. Some do this simplify to avoid confusing audiences between two scales. Others might do it by mistake, not understanding that these are separate, albeit similar scales. It can get more confusing since some countries still use the old scale, which can become unclear with secondary reporting, as happened with a recent tornado in Brazil. With regard to tornado ratings at least, I would say news outlets are secondary as there usually are published statements from the agency that did the survey. In some instances, might it work to say “these are common errors from secondary sources on [topic]”? TornadoLGS (talk) 16:44, 15 November 2025 (UTC)

News sources from the time of the event are primary. Traumnovelle (talk) 20:10, 15 November 2025 (UTC)

Even in regard to information they are relaying from another published source such as an NWS statement? TornadoLGS (talk) 21:41, 15 November 2025 (UTC)

Yes, if they are relaying information from elsewhere it is a primary source, secondary sources interpret and analyse primary sources and provide their own conclusion based on them. Traumnovelle (talk) 22:06, 15 November 2025 (UTC)

Yes, that's right. See WP:LINKSINACHAIN. WhatamIdoing (talk) 22:32, 15 November 2025 (UTC)

More precisely, news reports that contain no analytical content (e.g., compare/contrast) are primary. That's most news sources, but it's not all of them. WhatamIdoing (talk) 22:18, 15 November 2025 (UTC)

I see. And similar for documentaries? I see similar switching of scales there. In that case it seems best to cite the NWS directly since it avoids telephone type errors. TornadoLGS (talk) 22:58, 15 November 2025 (UTC)