Wikipedia talk:No original research

Add topic
Active discussions
  (Redirected from Wikipedia talk:NOR)

Publisher website links and WP:PRIMARYEdit

– 15:34, 10 May 2021‎ (UTC)


It appears that the author of this article thinks that there is no distinction between not-for-profit and for-profit higher education in the United States, even going to far as to claim Title IV funding is available to for-profit schools, when it is in fact expressly denied, and that for-profit schools receive federal research funding and have endowments, all without proof, and simply citing lists of non-profit endowments, for example. I have removed a bunch of material as a result, and I would request that this article be reviewed for return of said material and original research, as it appears to be a focus topic for the author. (talk) 01:05, 6 June 2022 (UTC)

Lengthy summaries of court decisionsEdit

This happened frequently with older articles on case but has come up again recently such as a case decided today Carson v. Makin. Editors are building out long sections which purportedly summarize the decision but to an excess degree that I worry there is effectively original research going by way if deciding what are the most important points of the decision to summarize (this can happen with any lengthy passage that rests on one single source). To wit, there are news articles that highlight key statements and points from the opinions today, but would not at all support a lengthy section. May law articles in the future would give the analysis to this level to support.

I feel it is necessary to address this to a bit more degree here, and if generalizing it further is possible (re using single primary sources to support a lengthy summary) perhaps here, thus is related to having some skill of the art as an editor to "use" the source thus way, which may involve facets that an editor not skilled in the art would not see. And that such summaries of primary sources should avoid use of "skill of art" knowledge. (I can see this also applying to some our math and science article that are deep textbook dives into derivations, which really shouldn't be at WP.) --Masem (t) 18:32, 21 June 2022 (UTC)

IMO direct summary by an editor of the actual decision should never be done. Such will usually have errors and things that the court really didn't say and represent heavy OR / interpretation of primary sources. But for US Supreme Court decisions there is generally already a good summary available. The released document contains the actual decision but is also preceded by a Syllabus (summary) which is not the official decision but instead a summary of it written by the Reporter of Decisions. So an immensely expert, NPOV credible third party summary is already available. North8000 (talk) 19:05, 21 June 2022 (UTC)
I would still be in question of being overly detailed just from the synopsis, since even there there is some skill of the art to work through the most salient points. At least with SCOTUS, nearly every major decision has mainstream coverage (NYTimes, Scotusblog at worse) that I find is nearly the best way to pull the most salient points and interesting quotes. Masem (t) 19:10, 21 June 2022 (UTC)
Descriptions should come only from secondary or tertiary sources (possibly together with cautious use of direct quotations to illustrate something one of these sources says). Court documents of any kind (pleadings, opinions, etc.) are full of technicalities and terms of art which we should not be interpreting for ourselves. EEng 20:18, 21 June 2022 (UTC)

P/S/T sources; P/S/T sourcing; or P/S/T source materials, etc.Edit

Pinging editors previously engaged in discussion on (primary) sources in archive: Crossroads Huggums537 Jc3s5h North8000 Paul Siebert Rjensen SMcCandlish WhatamIdoing in case of their potential interest here.

It's confusing. Editors can view "sources" as being synonymous with, for instance, publications. After a period away I know I steered/was steered/became compliant toward this kind of view.

The NOR section currently titled Primary, secondary and tertiary sources begins with the sentence Wikipedia articles should be based on reliable, published secondary sources and, to a lesser extent, on tertiary sources and primary sources.

I think that a proportion of Wikipedia editors may stop reading at this point while a proportion more may be influenced by views of the kind of editors that do. Editors I experience can seem to view, for instance, a "secondary source" as a kind of monolith solely composed of material defined by some sort of prescribed sourcing as "secondary". There's also the potential that other editor, while understanding the issues involved, may utilise the lack of clarity for their own ends. There's a clarity problem.

I'm not proposing any definite wording suggestion but think that problem addressing wording might read something like: "Wikipedia articles should be based on reliable, published secondary source material and, to a lesser extent, on tertiary source materials and primary source materials. A problem addressing section title might refer to something like "Primary, secondary and tertiary sourcing".

I'd suggest that various content within the section could be reworked and, for what it's worth, here's my provisionally bold attempt.

Within the current NOR text that begins "A secondary source provides..." a fair use of an example presents that: "A book by a military historian about the Second World War might be a secondary source about the war, but where it includes details of the author's own war experiences, it would be a primary source for those experiences." (I'll summarise this as the book including secondary source material with personal primary source material included). In more common practice "A book ... about ..." a topic might include photographs and various accounts related to that topic, (as potentially primary source material within secondary source material). News articles, if they provide any of the "author's own thinking" often take this to the extreme presenting predominantly primary source material with anything down to a potential mere border or tagging on of sometimes questionable secondary source opinion. (As an aside, this all also got to thinking on a potential juxtaposition between the secondary source requirement for the "author's own thinking" and the ruling against opinion). But back at the example quoted, of the "book by a military historian", if this book made reference to, say, another person's war experiences, that account may itself have included that "author's own thinking" on, for instance, personal perceived follies/successes in pervious wars, (secondary source material within primary source material). Closer to home, Wikipedia calls itself "Wikipedia" (as primary within tertiary with a whole encyclopaedia's worth of PST references placed into a phenomenally voluminous content along the way).

It's the material (or something like that) that counts.

GregKaye 09:32, 27 June 2022 (UTC)

This is complicated, including that I think you are dealing with at least two different areas (primary vs. secondary and trying to separate out opinion in that context) but I think that you are on to an important and useful point and direction there. One challenge is that secondary inherently includes the author's opinion, going the full range from choice of words in an unbiased / objective summary e.g "most" in "most of the people on the plane were Americans" to those which incorporate highly biased / distorted / spun summaries or choices of words and decisions on what to include/exclude. North8000 (talk) 12:03, 27 June 2022 (UTC)
Secondary does not inherently include the author's opinion. A Meta-analysis is a secondary source. It is not an opinion.
I am not sure why you believe that a statement like "most of the people on the plane were Americans". This is either true or false; it is a statement of fact. It is not a statement that depends upon your viewpoint. People cannot reasonably claim "According to Alice's viewpoint, most of them were Americans, but Bob holds the equally valid viewpoint that only a few people were Americans and most of them were French." Instead, the reasonable claim would sound like "Alice says they're mostly Americans, Bob says they're mostly not Americans, and either Alice or Bob is factually wrong."
I contrast this with actual opinions: "Alice says the coffee tastes good, but Bob holds the equally valid viewpoint that the coffee does not taste good". Whether or not the coffee tastes good is an opinion because it is actually subjective, meaning that the answer depends on who (=the subject) is speaking. WhatamIdoing (talk) 16:04, 27 June 2022 (UTC)
@WhatamIdoing: This is going to be so abstract that it will sound backwards, but not only do I agree with the point that you are making, but my post was actually to promote the line of thought that you just expressed. Basically to make the distinction between objective factual summary that isn't a reach and biased opinion. But one can derail that quest by claiming that even the most straightforward summaries involve author subjective choices. E.G. in ""most of the people on the plane were Americans" can you exclude using "most" for a 49% plurality? Or might 51% be not enough to use the term? And is "American" just US citizens vs residents. To avoid that derailing, you need to be able to say "yeah, but it has no signs of bias or highly subjective stuff, so we call that objective, and so objective accuracy does exist. And lastly I noted some of the less recognized forms of bias as flags that it isn't such. North8000 (talk) 17:50, 27 June 2022 (UTC)
Part of this sounds like a situation in which multiple valid definitions of this word exist, and the statement is only true for some of those definitions, so a POV pusher tries to pretend that the statement is wrong for all definitions. Consider "Females can get pregnant", and someone rejecting the statement because that's not how it works in seahorses, or because some females are sterile, or currently too old or too young to produce offspring. The statement isn't wrong, but you do need to know what the statement means, which is somewhat different from "every single female organism can get pregnant right this minute".
Similarly: It is true that "most of the people on the plane were Americans" if even a slightly majority of the people on the plane were US citizens and/or US residents and/or US nationals. If there were 51% US residents, including 10% that were not US citizens, then it is still true that "most" were Americans. Depending upon DUE and BALASP considerations, one might wish to be more precise ("Most were US residents" instead of "Most were Americans"), but the more generic wording isn't wrong just because you could misunderstand it.
I don't think we should accept claims that this is "subjective" or "an opinion" at all. If someone says "Their count of Americans on board is biased subjective opinion because the One True™ Definition of this word with multiple definitions is mine", then maybe we should consider whether that editor is has CIR or NOTHERE problems. It's true that definitions can be hugely important (consider counting "patients with the common cold" vs "people with the common cold": since most people don't seek professional medical care for the common cold, these two counts produce wildly different numbers), but Wikipedia editors shouldn't be deciding that the sources are wrong because we wish that they counted the other group. WhatamIdoing (talk) 23:55, 28 June 2022 (UTC)
When it comes deciding if a passage in a secondary source is primary because it reflects the author's own thinking, I think we need to distinguish cases where the author's thinking is based on the secondary sources she read, or is based on her own experience. If the author wrote a book about a war after spending a couple of years going through a defense department archive of documents related to the war, and quoted a passage from a report by a 1st lieutenant about conditions at a fire base, and described the report as typical of hundreds of other reports, the quote is secondary because it is based on multiple sources that the author read. If, later in the book, the author described an incident in a communications center in Hawaii that she took part in, that would be primary. Jc3s5h (talk) 12:55, 27 June 2022 (UTC)
I agree with Jc3s5h.
I think the problem @GregKaye is asking about is whether we should be implying "secondary source document" rather than saying "secondary source material". I agree with his identification of the problem and accept his solution. Given our still-shaky grasp of the difference between secondary and independent and the potential for abuse by POV pushers, I'm a little leery of encouraging editors to decide that this sentence or that sentence is primary/secondary, but it's worth a try.
On the question of "author's own thinking", I can see that the language is misleading. We want the kind of thinking that results in a transformation of primary sources – comparing, contrasting, analyzing, etc. We don't want the other kind of "thinking". "I think I'll go fix breakfast in a minute" might be this "author's own thinking", but it's not a secondary source. WhatamIdoing (talk) 16:10, 27 June 2022 (UTC)
We need to be indicating "secondary source material". There is no such thing as a "secondary source document/publication" except when one just incidentally happens to contain nothing but secondary material; there isn't a publication type that is categorically secondary from start to finish. "Editors can view "sources" as being synonymous with, for instance, publications." That's an error. For example, any given newspaper contains various forms of primary-source material, like advice columns, editorials, op-eds, advertising, etc. Only certain kinds of writing within the publication are secondary (and even then might actually contain bits of primary; to use the above example of someone writing an investigative journalism piece after doing a bunch DoD document research, if they close out the piece with a policy-change recommedation, that last bit (an opinion) is primary.  — SMcCandlish ¢ 😼  21:00, 27 June 2022 (UTC)
One of the hardest things on wiki has been to convince editors that most of the news in the newspaper is WP:PRIMARYNEWS content. Almost nothing in a traditional newspaper is secondary source material; in the modern age, some analytical work has become more common. You are actually more likely to find some secondary material in the opinion section than in the news. "Paul Politician gave a speech at the city council meeting last night" is pure primary. "Carol Challenger and Paul Politician are evenly matched candidates. They agree on property taxes, the importance of local agriculture, zoning, and land-use issues, but Challenger is stronger on climate change and Politician is stronger on education" is secondary material (because it's a compare-and-contrast analysis).
But since GNG requires "secondary" sources, and news articles is all editors have for some subjects, we insist that news articles must be secondary sources, because I must have my article, and all else – including the definitions of independent, secondary, and significant coverage – will bow to the goal of keeping my article. WhatamIdoing (talk) 00:10, 29 June 2022 (UTC)

I wanted to leave the initial thread question open but my inclination would be to steer away from reference to P/S/T sources and toward reference to both P/S/T sourcing and P/S/T source materials, etc. A difficult question (though maybe for another time) is where the border lies between an opinion piece of a rookie journalists and the thoughtful analysis of an accomplished scholar. An expedient solution could be to say news opinion pieces are out and other content with "author's own thinking" can be evaluated. Otherwise some articles may be swamped with potentially fad or similar opinion. I'm sure it may also be in a newspaper's interest to get articles cited in Wikipedia and I'm doubtful that potentially developing dynamics would help us build encyclopaedic content.

Revisiting the in use example: "A book by a military historian about the Second World War might be a secondary source about the war, but where it includes details of the author's own war experiences, it would be a primary source for those experiences." I see at least four-six potential categories of information here:

  1. things the author recollects or had personally recorded about ---self
  2. things the author recollects or had personally recorded about other people or things
  3. (things the author recollects or had personally recorded about a group of people perhaps inc. things that included the author)
  4. things the author has later researched so as to learn more about their own ---time situation (and the rest mirrors the above)
  5. things the author has later researched so as to learn more about other people or things from the time
  6. (things the author has later researched so as to learn more about a group of people perhaps inc. things that included the author)

The "author's own thinking" could be applied to any of this, or it could all equally apply to events that occurred for another author just hours or days previously. And we describe it as the "author's own thinking", but is it though? How do we know it's not, for instance, that it isn't The "author's mate-from-down-the-pub's own thinking" or that it isn't plagarised in some other way. I suspect this stuff may get complicated when addressed in future time. But it may certainly be worth noting in relation to ways we present classification terminologies. Later if the issues I mentioned can be tackled, I confess to suspecting that we might need something like that expedient solution I mentioned earlier. GregKaye 22:43, 27 June 2022 (UTC)

I think that the end result should be to do our best at assessing the expertise and objectivity of the source with respect to the text/item which cited it. Obviously not a no-brainer to determine, but even setting that as the standard/goal would solve a zillion problems in Wikipedia. North8000 (talk) 00:27, 28 June 2022 (UTC)
I think the stuff about opinions could fill a page by itself. It's far more complicated than we admit here. WhatamIdoing (talk) 00:12, 29 June 2022 (UTC)

side bar re PSTEdit

  • This might be a good time to remind editors about how the PST section came to be. The policy originally warned editors not to add OR to articles, because doing so turned WIKIPEDIA into a primary source.
It was then thought that (to better explain what we were talking about) we needed to define what a “primary source” was. That, in turn, led to us defining what secondary and tertiary sources were. However, along the way we somehow lost that original crucial point: that WIKIPEDIA should not be primary.
PST has grown into a distracting side discussion about the nature of our sources, when it should be about the nature of WIKIPEDIA.
I think any discussion of PST needs to return this original crucial point… because NOR isn’t really about the sources we use, but what WIKIPEDIA is. Rant over. Blueboar (talk) 21:54, 27 June 2022 (UTC)
Rant/grunt or whatever away. I read this when reviewing archive refs to primary and had in mind to quote you anyway. Awesome stuff. GregKaye 22:46, 27 June 2022 (UTC)
I have been thinking about this problem. Could we WP:SPLIT the PSTS section out of here, to its own page? Then this page could have a {{Main}} link to PSTS, and say the parts that are actually relevant to NOR, namely "Wikipedia is not a primary source, so don't stuff your own original research into articles". WhatamIdoing (talk) 00:13, 29 June 2022 (UTC)

Defining primary ...Edit

Perhaps something like this would work:
Primary sources are composed of direct or firsthand evidence in regard to a topic. Often primary source materials are created at the time when the events or conditions occurred, but they can also include content in autobiographies, memoirs, and oral histories recorded later. Reproductions of primary source materials retain their primary status and different citations used in Wikipedia will link to references presenting varying extents of primary source content.
I think that this fits well with the WP:Identifying and using primary sources, Yale... citation GregKaye 10:05, 28 June 2022 (UTC)

Defining secondary ...Edit

Secondary sources describe, interpret, analyze, evaluate, comment or discuss primary source materials that they will often also quote. They are works which are one or more steps removed from the event or information they refer to, being written after the fact with the benefit of hindsight but do not need to be independent.
This draws from the WP:Identifying and using primary sources, James Cook University... citation GregKaye 10:35, 28 June 2022 (UTC)


Working on from OP 3rd para, perhaps:
Wikipedia articles should ideally be based on reliable, published secondary sources and, to a lesser extent, on tertiary sources and primary sources. The establishment of the topic's notability will also typically be achieved through reference in secondary or tertiary sources which are also helpful for the avoidance of novel interpretations of primary source content. All analyses and interpretive or synthetic claims about primary sources must be referenced to a secondary or tertiary source and must not be an original analysis of the primary-source material by Wikipedia editors.

Appropriate sourcing can be a complicated issue, and these are general rules. Deciding whether the use of primary, secondary, or tertiary source material is appropriate in any given instance is a matter of good editorial judgment and common sense, and should be discussed on article talk pages. A cited source may be considered primary for one statement but secondary for a different one. Even a given source can contain both primary and secondary source material for one particular statement. For the purposes of this policy, primary, secondary and tertiary sources are defined as follows:

I added the word "ideally" towards the beginning because, if secondary sources are written with hindsight, there may be none about when some articles are first written. In the second paragraph I think that text might be cut (or edited back) from "A cited source..." as the issues mentioned may be otherwise dealt with something like in the proposed texts. GregKaye 11:21, 28 June 2022 (UTC)

GregKaye 11:21, 28 June 2022 (UTC)

  • Question… I am unclear as to the relevance of the proposed changes to the underlying concept of this policy page (ie don’t add your own original research). Would it be something that would be better placed on some other policy/guideline page?

Two main things that I think would be useful here:

  1. I'd like to clarify from the beginning that a section of a content can be a P/S/T source and that the entire content does not necessarily get labelled as primary, secondary or tertiary.
  2. I'd like to clarify from an early stage that a secondary source is not just a content that quotes a primary source. I think this is a common misunderstanding.

It's also all something that I'm trying to get my head around. To now I've been under the misunderstanding that Wikipedia sourcing was against citation of contemporary opinion which WP:RSOPINION indicates is not the case and retracted one short comment I'd previously made. GregKaye 16:49, 28 June 2022 (UTC)

All valid questions and concerns… BUT… how do they relate to “don’t add YOUR OWN original research”? I think you may have gotten distracted by focusing on the nature of the source, when this policy is (or should be) focused on the nature of what WE do with the source (ie what we write based on our sources). NOR is a content policy, not a sourcing policy. Blueboar (talk) 17:05, 28 June 2022 (UTC)