Wikipedia talk:No original research/Archive 61

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 55

←

Archive 59

RfC: Should "news articles" be added to WP:PRIMARY?

The following discussion is an archived record of a request for comment. Please do not modify it. No further edits should be made to this discussion. A summary of the conclusions reached follows.

The answer is no, unanimously, to both questions. Kraxler (talk) 03:48, 20 July 2015 (UTC)

Question: should WP:PRIMARY be updated to define news sources as primary, aligning more closely with the definition used by historians?

Background

Our policy limits how primary sources can be used, particularly regarding living people. Exactly what constitutes a primary vs secondary source is subject to some debate; for example, the current version of our policy defines secondary sources as those that contain an author's interpretation, analysis, or evaluation of the facts, evidence, concepts, and ideas taken from primary sources. Primary sources are defined as personal accounts such as witness statements. Historians, in contrast, define contemporary news sources as primary - discussed, for example, here.

At Talk:Dennis Hastert (permalink), the latter definition is being applied. Because of the substantial impact to our sourcing practices if news sources are now considered primary at Wikipedia, our new practice should be updated in WP:NOR. Proposed revised wording of WP:PRIMARY follows:

Before/after wording

Current wording:

Primary sources are original materials that are close to an event, and are often accounts written by people who are directly involved. They offer an insider's view of an event, a period of history, a work of art, a political decision, and so on. Primary sources may or may not be independent or third-party sources. An account of a traffic accident written by a witness is a primary source of information about the accident; similarly, a scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment. Historical documents such as diaries are primary sources.

Proposed wording:

Primary sources are original materials that are close to or published in the same historical context as an event, and are often accounts written by people who are directly involved. They offer an insider's view of an event, a period of history, a work of art, a political decision, and so on. Primary sources may or may not be independent or third-party sources. An account of a traffic accident written by a witness is a primary source of information about the accident; similarly, a scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment. Historical documents such as diaries and contemporary sources such as newspaper articles and news reports are primary sources.

VQuakr (talk) 15:27, 18 June 2015 (UTC)

Discussion

Oppose & Speedy Close If news articles (from reputable pubs like the NY Times, Telegraph, BBC News, etc) from the time period can not be used as secondary sources, then a massive amount of content and many entire articles which have been live on wikipedia for years without any issues would need deleting. I'll go further and say that this RFC should be speedily closed since there's pretty much zero chance of this becoming policy - the issue here seems to be that an admin has closed an RFC in a way that is wholly inconsistent with existing wikipedia guidelines and policies, not that there is any actual need for a policy change. The best solution is to leave the policy as-is and overturn the erroneous RFC close. Fyddlestix (talk) 15:48, 18 June 2015 (UTC)
- I think that if you've actually built whole articles on century-old newspaper articles (especially of the "current events" or "society column" types), then you've done something seriously wrong, even if the publishers were reputable back then. WhatamIdoing (talk) 08:27, 20 June 2015 (UTC)
Oppose, because the proposer repeatedly fails to understand what's going on. Some news articles are secondary sources, while others are primary. In this context, these news stories are primary sources, because they're news about the event that's going on (try taking a historiography class and telling your professor that these are secondary!), but the idea that all news articles are primary sources is a basic failure to get the point. Nyttend (talk) 15:57, 18 June 2015 (UTC)

@Nyttend: probably "continues to fail to understand" would be more accurate than "repeatedly." I thought I had addressed this with the "contemporary" phrasing in the proposed addition. Is there an alternate formulation of the proposed policy revision that would help me (and future readers) understand without the need to supplement our policy with external sources such as GSU? VQuakr (talk) 16:03, 18 June 2015 (UTC)

@Nyttend: No one questions that historians would define news articles as primary sources. The question is whether wikipedia defines them as such (and, more importantly, whether it wholesale rules out their use in BLP articles like Dennis Hastert). Assuming that we're talking about reputable news sources (ie, newspapers or record), I have never seen anyone suggest that the latter is true. In fact, sources like this (ie, news articles "about events that are going on") are used in wikipedia articles constantly (like, hundreds of times, every day) without anyone having an issue with it. Fyddlestix (talk) 16:14, 18 June 2015 (UTC)

@Nyttend: You are incorrect. Newspapers from the time of the event are not primary sources from WP's perspective. If you want to change the rules to forbid them, you can start a RFC here with that wording specifically, but I can't see it having a serious chance of passing. --Aquillion (talk) 01:14, 19 June 2015 (UTC)

Nyttend was referring to (among other things) the Wikipedia article Primary source, not to the Wikipedia policy WP:Primary, as I understand.Anythingyouwant (talk) 01:22, 19 June 2015 (UTC)

Nyttend is correctly describing the policy as written on this page. I'm not quite sure why this is such a hot button with this group though: there is no prohibition on using primary sources in an article (not even a BLP). The overall article should be largely based on secondaries, but primaries are acceptable, including primary news reports in reputable newspapers. WhatamIdoing (talk) 08:39, 20 June 2015 (UTC)

Oppose defining news sources as primary, because many news sources are not. Incidentally, I do not see how opposing here will affect the legitimacy of the Hastert RFC close.Anythingyouwant (talk) 17:12, 18 June 2015 (UTC)
Oppose & speedy close - Newspaper and media sources from reputable media organizations are not primary sources from WP's perspective. If we change the policy as suggested, thousands of articles will become no longer compliant. - Cwobeel (talk) 19:08, 18 June 2015 (UTC)
Oppose Although I can see where it comes from. Press agency releases that just list some facts are probably close to primary reports, while a thorough journalistic article that analyses interprets and discusses multiple angles is definitely a secondary source. Arnoutf (talk) 19:40, 18 June 2015 (UTC)
Comment There is an obvious consensus against the specific phrasing I proposed. Rather than this being a slam-the-door, yes-no vote I think it would be more effective to discuss whether specifically addressing news sources in the policy would be worthwhile. We have done something similar with the extensive discussion of medical literature in WP:MEDRS. News is probably simpler but such sources are very, very commonly cited on Wikipedia. To formulate a more explicit problem statement: The present wording of our policy does not differentiate between news sources that review other sources, and current-event sources that lack historical perspective. VQuakr (talk) 00:56, 19 June 2015 (UTC)
Strong oppose. Newspapers from the time of the event are not primary by Wikipedia's standards. --Aquillion (talk) 01:04, 19 June 2015 (UTC)
Oppose & speedy close. I agree with Aquillion, Fyddlestix, and Cwobeel, for the reasons stated above. Neutrality^talk 03:51, 19 June 2015 (UTC)
Oppose and speedy close Unless someone comes along and wants to open an RFC to propose to change wikipedia's policy on this (and I would oppose that change as well) then we should not have this RFC just because of one article RFC closure. This would be unworkable for wikipedia without huge changes in what we cover, with practically any current events impossible to cover. Davewild (talk) 06:52, 19 June 2015 (UTC)
I recommend that the people here who are making claims about "Wikipedia's standards" actually read this policy, paying special attention to the third footnote. This policy has listed news articles as a type of source that is (typically) primary for years. WP:PRIMARYNEWS deals with the nuances better than this policy (Anythingyouwant is correct that there are different types of "news" sources, and that not all are considered primary), but this policy has called news sources "primary" for a very long time – and it is "Wikipedia's standard" for what constitutes a primary source on wiki. WhatamIdoing (talk) 08:35, 20 June 2015 (UTC)
Oppose as an incredible simplification of the issue. Grognard Extraordinaire Chess (talk) Ping when replying 00:12, 25 June 2015 (UTC)
Oppose. Newspapers are mostly secondary sources, despite what the RfC question says (which is not neutral, because it begs the question). A primary source is written by someone very close to the event, usually someone involved or an eyewitness. For example, a newspaper report by a journalist who witnessed 9/11 is a primary source for material about 9/11. But most newspaper reports are not written by involved people or eyewitnesses, and therefore most of the newspaper articles used as sources on Wikipedia are secondary sources.
A confusing factor, and perhaps this is what the OP means when referring to historians, is that contemporaneous newspaper articles about an event a long time ago – articles written at the time of the event – might be regarded by historians as primary sources, because the authors were close to the event relative to us. So contemporaneous newspaper articles about World War I might be primary sources about the war, because they tell us how the war was reported at the time. But for us, alive in 2015, an article by a journalist who watched 9/11 on television is a secondary source because the author wasn't close to the event relative to us. Sarah ^(talk) 02:44, 25 June 2015 (UTC)
- I agree with you that decades-old newspaper articles are (almost always) primary sources, either for what people thought was happening or for how they interpreted it. But a bit of physical distance isn't sufficient. Secondary sources require some sort of transformative thought. Merely being a thousand miles away, but writing about what you see happening through the miracle of telecommunications, is not the kind of mental separation that's necessary. WP:Secondary does not mean independent, and primary doesn't require being personally involved. Also, a good deal of them actually are written by journalists who are also eyewitnesses. I just flipped through the local news for the biggest newspaper in my area. I found a report of a local government meeting (the reporter was an eyewitness to the meeting), some kid talked to a politician (eyewitness), took pictures at a sports event (eyewitness), a photogenic political protest (eyewitness), student sports events (eyewitness), more government meetings (ditto), a drug bust (wasn't clear), routine court news... I don't see any reason to think that any of those are secondary sources, although they're all WP:INDY sources (with no COI). WhatamIdoing (talk) 03:08, 25 June 2015 (UTC)
  - @User:Aquillion I agree with SlimVirgin (that makes a change Sarah!), the only thing I question is whether 9/11 is still current (there may be people reading this who weren't even born then). The question of when an event slips into history is difficult to pin down, but the 30 years rule for government papers is probably as good a point as any. I have contributed to a number of English Civil War articles some of which mention contemporary newspaper articles. There is no way that such newspaper articles can be considered anything but primary (see for example The Moderate Intellingencer). On the other side of that equation there is a Wikipedia page Wikipedia:Press coverage 2015 and those articles are either reliable or unreliable secondary source, but one day they will be primary sources. Those in the page Wikipedia:Press coverage 2001 are moving that way, because they tell you more about what people thought about Wikipedia in 2001 than they do about what Wikipedia is now (eg see the [MIT Technology Review a linked on the Press coverage 2001 page). An interesting area for discussion in the future will be what happens when a Wikipedia articles grow old. For example I helped to write the article Eritrean–Ethiopian War (1998–2000). It was started in 2005 and the core of the article was completed by the end of 2006. There were next to no books on the subject, so it was written using contemporary news sources. In 2005–2006 they were less than 8 yeas old. Now 10 years on more and more of the sources are only available via the Wayback machine. So far that is not too much of a problem, but in 20 years time, unless those sources are replaced with later ones, then the article that was initially supported with secondary sources will become supported with primary ones without one citation being altered! -- PBS (talk) 14:13, 6 July 2015 (UTC)
    - I disagree that all of this year's media output is secondary. That's just plain wrong, and you are unlikely to find a reliable source that actually claims this. Eyewitness news reporting, including writing an article about your personal experience of editing Wikipedia, is a primary source, full stop. So are routine articles like "There was an edit-a-thon last night". This purported "press coverage" in Wikipedia:Press coverage 2015 is nothing more than a tutorial on how to type Special:Random with an explanation that the author believes you'll have more fun reading Wikipedia if you look at random articles.
      However, I'd like to expand on an interesting point that you make: The classification depends upon how you use it, not just on its inherent qualities. All sources are primary for something. A meta-analysis of original scientific reports is always, inherently, a secondary source, no matter how old/biased/unreliable it is. It takes original material (the original reports) and transforms that original material into a new thing (the results of the meta-analysis) – that is, even according to the most stringent and exclusive definitions, a secondary source. However, there are two important caveats:
      1. That meta-analysis is also a primary source for things like the authors' names, affiliations, the journal, the number of words contained in the abstract, or whatever else isn't part of the "transformed" thought.
      2. When that meta-analysis is many decades old, or if it comes from a disreputable publisher, then you shouldn't use it like a secondary source. You should only use it like a primary source, for "what people thought about (the subject of the meta-analysis) in" previous decades (or not at all, per WP:DUE), not directly for the subject itself. WhatamIdoing (talk) 17:51, 6 July 2015 (UTC)

WhatamIdoing I think that you are making a difference between primary and secondary sources that may be a case of style rather than substance. Take for example Brian Hanrahan's report from the Falkland's war that included the famous line "I counted them all out, and I counted them all back." The BBC states that this was done as "a clever ruse to get round reporting restrictions, so that he could say that all British Harrier jets had returned safely." If instead a BBC news editor back in London had edited the piece and simply reported that "All the British Harriers returned safely" (and used stock footage of a Sea Harrier) it would no longer have been a personal observation so presumably it would then be a secondary source. If the BBC had instead had written a contemporary piece and the report that "John Doe [not a BBC staff member] stated last night 'I counted them all out, and I counted them all back'.", then it is no different from a book quoting a primary source 17 years later (page 245 Rolling Thunder: Jet Combat from World War II to the Gulf War by Ivan Rendall (1999)). -- PBS (talk) 18:08, 7 July 2015 (UTC)

No, I'm making a difference based on what reliable scholarly sources say. You might want to read WP:LINKSINACHAIN. Merely quoting a primary source does not make your document become a secondary source. WhatamIdoing (talk) 01:58, 8 July 2015 (UTC)

I am familiar with links in a chain, there is a practical example over date of the knighthood of Sir Oliver Cromwell. However I think that your conclusion is wrong, and also that if a change of presentation alters something from a primary source to a secondary one, then the line between the two is more fluid than you are allowing for. As always the devil is in the detail. -- PBS (talk) 13:25, 12 July 2015 (UTC)

The line is fluid, and that's one of the problems with people going about deleting good content with the anti-WP:PRESERVE excuse that "But it's primary, so it's WP:NOTGOODSOURCE!" Whether a source is primary or secondary (or tertiary) depends on both its inherent qualities and how you use it. And whether it's technically reliable for a given piece of material is almost entirely unrelated to its classification under historiography. "Secondary" is not some fancy spelling for "reliable". WhatamIdoing (talk) 22:19, 12 July 2015 (UTC)

Oppose Big oversimplification, it is impossible to find all news stories as primary. The change would turn the policy on its head. AlbinoFerret 00:15, 9 July 2015 (UTC)
Oppose. WP is already burdened by ~~a lot~~ miles of red tape, no more of that crap please. The Yeti 14:15, 9 July 2015 (UTC)

Secondary RFC

Question: should WP:PRIMARY be updated to define all news sources from soon after an event as primary, without exception?

The background is covered above, but Nyttend has criticized the wording of the RFC above, so I feel we should have one on their actual interpretation. As I understand it, they believe that all news sources from soon after an event are automatically considered WP:PRIMARY sources (they haven't specified what they consider the cutoff, but they feel that there is a cutoff clear enough and strong enough to close an RFC against overwhelming consensus by citing WP:CONLIMITED based on their personal assessment of particular news sources as too recent and therefore primary.) While I agree that sources from soon after an event should sometimes be used with caution, I think it's clear that both the wording of the policy and the way it has been applied throughout history does not agree with the interpretation that they are WP:PRIMARY, so we need an RFC to settle this question specifically. --Aquillion (talk) 01:30, 19 June 2015 (UTC)

Discussion

Strong oppose this change as well. While we always have to be careful of WP:RECENTISM, sometimes topical events happen to BLPs which require immediate coverage from reliable sources; sometimes it will require saying something potentially negative under BLP. Those sources should be approached with care, definitely, but Nyttend's proposal to define all contemporary news sources as WP:PRIMARY is a hamhanded approach would make it impossible to write complete articles about many recent subjects, and I feel it's important to shoot it down decisively. (Consider: Nyttend used this argument to say that we cannot use the term 'sexual abuse', but in fact simply accusing someone of abuse based purely on primary sources would be unacceptable -- and according to their logic, no primary sources on the recent abuse scandal exist; therefore we would not be able to cover it at all until, I suppose, someone writes a book on it.) --Aquillion (talk) 01:30, 19 June 2015 (UTC)

You might be better off starting another RFC at some point. I'm not aware that a sub-RFC like this is workable.Anythingyouwant (talk) 01:32, 19 June 2015 (UTC)

It's not unusual to add an additional question to an RFC if someone objects to the wording of the first one, suggests an alternative resolution, or the like. --Aquillion (talk) 01:39, 19 June 2015 (UTC)

I object in this particular instance because it makes things much more complicated, and I don't think I would be entitled either to add another RFC question at this point. I don't object to starting another RFC once this one is closed.Anythingyouwant (talk) 01:46, 19 June 2015 (UTC)

WP:DROPTHESTICK - This has gone long enough already. No more RFCs on this aspect as we have very clear and long standing content policies. We have an RFC that was poorly closed and that will likely either re-opened or overturned. - Cwobeel (talk) 03:09, 19 June 2015 (UTC)

You guys drop the stick. Just let the admins decide on the first RFC. Why have a second, and then essentially a third, at a policy page to advance your position in a content dispute? Oy vey.Anythingyouwant (talk) 03:12, 19 June 2015 (UTC)

Strong oppose - Nyttend needs to acknowledge his mistake, and move on to avoid further drama. Let's go back to editing please. - Cwobeel (talk) 03:06, 19 June 2015 (UTC)
Oppose and speedy close Unless someone comes along and wants to open an RFC to propose to change wikipedia's policy on this (and I would oppose that change as well) then we should not have this RFC just because of one article RFC closure. This would be unworkable for wikipedia without huge changes in what we cover, with practically any current events impossible to cover. Davewild (talk) 06:52, 19 June 2015 (UTC)
Oppose and speedy close per Davewild. This only came up because a single RFC close misinterpreted existing policy. There's no need to reconsider the actual policy (for which there is a broad and durable consensus), and no need to let this tempest-in-a-teacup drag on. Fyddlestix (talk) 15:51, 19 June 2015 (UTC)
Comment - The basic problem seems to be that historians have a different definition of primary and secondary sources than Wikipedia does, because Wikipedia does have articles both on news and on history. If Wikipedia were to adopt the historical definition, we would be impoverishing our ability to cover news. We need to use common sense as to when newspaper journalism is a valid secondary source (for news) and when historiography is required. Robert McClenon (talk) 17:41, 19 June 2015 (UTC)
Oppose and speedy close - agree with above. Additionally, this change would lead to meandering and meaningless (and likely endless) discussions on what "soon after" means. Neutrality^talk 02:54, 20 June 2015 (UTC)
Suggestion: We should consider linking to WP:PRIMARYNEWS (or maybe just mentioning it in the /FAQ on this page, rather than putting it in the policy). The short answer is that most newspaper-type news articles are primary sources, but it's more complicated than that. Linking people out to a {{supplement}} is going to give us more room to explain (that section's running around 900 words at the moment) and therefore increase the odds that editors will find an accurate answer. WhatamIdoing (talk) 08:22, 20 June 2015 (UTC)
Strong Oppose as blanket declarations shouldn't be applied here. Grognard Extraordinaire Chess (talk) Ping when replying 00:12, 25 June 2015 (UTC)
Oppose. No, because not all news sources from soon after an event are primary sources. Most aren't until quite a bit of time has passed. (See my post above). Sarah ^(talk) 02:48, 25 June 2015 (UTC)
Oppose I think Robert McClenon's suggestion of adopting the historian's definition is unworkable (like changing which side of the road we drive), however, 'We need to use common sense as to when newspaper journalism is a valid secondary source (for news) and when historiography is required' is spot on. Is that thought, or some improved variant in guidelines? Pincrete (talk) 20:10, 5 July 2015 (UTC)
Comment - I didn't suggest adopting the historian's definition. I opposed it because it would impoverish our ability to cover news, and we do cover news. I was trying to point out that Wikipedia and historians have different definitions of primary and secondary, and we have to remember the distinction. Robert McClenon (talk) 19:43, 7 July 2015 (UTC)

Why does it matter?

Once again, PSTS classifications seem to be the tail wagging the dog here. We seem to be missing the forest for the trees... so let me cut through the bullshit and ask the important question: Does it really matter (in terms of "original research") whether a news source is classified as PRIMARY or SECONDARY? I don't think it does.
Original research does not stem from whether a source is primary or secondary... original research stems from what we do with the source. Are we merely passing on what the source says in our own words? That is not original research. Are we engaging in analysis or interpretation of the source, or drawing our own conclusions based upon it? That is original research. NOR really is that simple. All the navel gazing about defining what is and is not a Primary or Secondary source ... all the argument about whether a source is Primary or Secondary is unnecessary. In both cases... it's not OR if we stick to the source ... it is OR if we go beyond the source. Blueboar (talk) 21:14, 25 June 2015 (UTC)

"Original research" is not part of the normal English language, it is special term defined in the original research policy. This can be proven by writing an article that isn't terribly insightful but defies the original research policy, adding it to Wikipedia, and observe it get speedily deleted as "original research". Then submit it to a scholarly journal and wait about 15 nanoseconds to receive a rejection notice because the article is too derivative.

Having established that "original research" is a Wikipedia term of art, we then see that the original research policy says "Do not base an entire article on primary sources." So either newspapers are potentially secondary sources, or all articles on current events must be deleted. Jc3s5h (talk) 21:37, 25 June 2015 (UTC)

Some pieces that run in newspapers certainly are secondary (just not the majority, or even close to it, judging from the paper I'm looking at).

I think that Blueboar is on his way to a proposal that PSTS be put into a separate page. WhatamIdoing (talk) 23:10, 25 June 2015 (UTC)

Been there... done that. Sure, I continue to have serious concerns about how PSTS continues to be misunderstood and how it has lost its original purpose (it's original purpose was to explain that Wikipedia is a Tertiary source, and should not be turned into a Primary source - which is what happens when you add OR), but I fully understand that there is no consensus to remove it from the policy. That said... I still don't get why people get so worked up over defining sources. The policy explicitly states that we are allowed to use Primary sources. So it does not really matter if we classify news source as Primary or Secondary... they can be used as long as we use them appropriately. Blueboar (talk) 00:25, 26 June 2015 (UTC)

I couldn't agree more. The core problem seems to be that an unwillingness on the part of some of the community to accept the importance of editorial judgement, implicit in determining whether sources have been used "appropriately". Peter coxhead (talk) 07:26, 26 June 2015 (UTC)

I think that part of the problem is people not believing that we can use primary sources. I looked up the dispute that spilled over to the above sections, and the closing statement said that the information couldn't be included because the sources were primary and PSTS banned the use of primary sources. Now that the community has sort of figured out that WP:Secondary does not mean independent (which took about three years of steady work), maybe we should try to work on "you can WP:USEPRIMARY sources" as the theme. WhatamIdoing (talk) 00:11, 30 June 2015 (UTC)

Primary sources are often the best sources to use. You just have to be careful because it's easy to get things wrong. This policy has never said or implied that they can't be used at all, so I've always wondered where that idea came from. Sarah ^(talk) 01:59, 30 June 2015 (UTC)

I think it may be a cumulative effect... there is a heavy emphasis on using secondary sources throughout PolicySpace (multiple policies and guidelines stress the need for secondary sources) and this adds up... eventually editors start to think they should never use primary sources. Blueboar (talk) 13:59, 30 June 2015 (UTC)

The problem — or one of the problems — is the failure to distinguish between original research and primary sources, which may not involve any research at all. — Robert Greer (talk) 00:48, 1 July 2015 (UTC)

WP:BLPPRIMARY certainly discourages use of primary sources in BLP's, even if it falls just short of outright banning them. VQuakr (talk) 00:56, 1 July 2015 (UTC)

This thread is a revelation to me, I've been told so often that primary sources are verboten! The trouble with that essential quality 'judgement' (as to when to use them), is that it is inherently undefinable. Which is a problem for a project dependent on guidelines.(my 2d.) Pincrete (talk) 20:26, 5 July 2015 (UTC)

It's only a problem if instead of "guidelines" you want "rules", and you don't accept that the really important feature of Wikipedia is consensus among editors based on the reasons behind the guidelines. There are some rightly rigid rules, such as no original research. For historical topics, it's almost always the case that using what historians call primary sources involves OR. In other topic areas (e.g. recent scientific discoveries), using journal articles – which some editors call "primary sources", unhelpfully in my view – usually doesn't involve OR, but does raise notability questions (an accurate report in Wikipedia based on a single scientific journal article isn't original research, but will often not be notable). What's important is that editors are free to discuss their use of sources openly and to reach informed decisions, and not have the discussion closed down by slogans masquerading as Wikipedia policies. Peter coxhead (talk) 10:47, 6 July 2015 (UTC)

As the policy says, Primary sources can be used (without slipping into OR) ... you simply have to understand when and how it is appropriate to use them, and when it is not. Blueboar (talk) 11:09, 6 July 2015 (UTC)

Sarah wrote above, "Primary sources are often the best sources to use." This is undeniably true. What do people think about adding those words to the policy? We might say "sometimes" rather than "often", but the effect might be positive. (We could give direct quotations as an example: it's usually better to cite the original than to play the telephone game with quotations.) WhatamIdoing (talk) 18:04, 6 July 2015 (UTC)

I think a lot of this goes back to notability and those endless debates at AfD. A secondary source carries more weight as evidence of notability than a primary source does, but once the topic has been established as a notable one, the primary source is often more valuable. For example, if someone discovers an important exoplanet and your sources are (1) a scientific paper and (2) the BBC's announcement of the discovery based on the scientific paper, then any conflict between those sources should be resolved in favour of the astronomer rather than the journalist who semi-understood his paper.—S Marshall T/C 11:13, 7 July 2015 (UTC)
- I think you're right about a lot of the 'of course it's secondary!' stuff being related to notability. WhatamIdoing (talk) 03:53, 8 July 2015 (UTC)
WhatamIdoing, I'd support adding something along those lines. Sarah ^(talk) 18:51, 7 July 2015 (UTC)

In the case of articles on history topics, primary sources are often not the best sources to use. For a example contemporary sources about war are often very biased and as well known "The first casualty of War is Truth". Or take the case of the breaking of codes any strategic history of any theatre that the Western Allies fought is severely flaw if ULTRA and MAGIC is not factored in. As one historian has commented "When [Monty] put Rommel's picture up in his caravan he wanted to be seen to be almost reading his opponent's mind. In fact he was reading his mail" (Andrew Roberts 2009). One of the structural problems about this particular policy (as it is about some guidelines) is that it is written as one shoe fits all. This causes problems for different project areas, because advise that is useful in one is just confusing in another. -- PBS (talk) 19:22, 7 July 2015 (UTC)

PBS, I wonder if you could live with a statement like, "Primary sources are sometimes the best sources to use, such as for supporting direct quotations or describing recent events." I could live with a "but not for proper history" clause, if that might help. Or substituting "most authoritative", rather than "best". That would address S Marshall's issue about the journalist vs the scientist. WhatamIdoing (talk) 03:53, 8 July 2015 (UTC)

Again... it all depends on what you are trying to support with the primary source. A WWI era poster depicting a German soldier with a baby impaled on a bayonet would be perfectly appropriate for supporting a statement that "During WWI, some propaganda posters depicted German soldiers as savage murderers"... I would say it is excellent source for that statement... precicely because it is a primary source. It's not a direct quote... but a direct description. Blueboar (talk) 01:30, 9 July 2015 (UTC)

I agree. Once the extent of such propaganda has been established by reliable secondary sources, primary sources are the best way of providing examples. Peter coxhead (talk) 07:22, 9 July 2015 (UTC)

Blueboar doesn't your example involved an explicit breach of WP:PSTS? Surly, as Peter coxhead points out, one would need a secondary source that explicitly make the statement that "[some Allied WWI] propaganda posters depicted German soldiers as savage murderers"? So while such poster may be appropriate as a eye candy the statement would need a secondary source to support it. If a contemporary primary source was to be used it would need to be in the form of a statement eg "[we are commissioning posters] depicting German soldiers as savage murderers to ...". -- PBS (talk) 10:33, 11 July 2015 (UTC)

Nope... the policy states: "A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge." In my example, there are no evaluative or analytical claims being made, no conclusion being drawn... it is simply a descriptive statement of fact about what the posters depict... a statement that anyone viewing the poster can verify. No original research involved. It's no different than citing a basic plot summary of a book or movie to the book or movie itself. Blueboar (talk) 12:42, 11 July 2015 (UTC)

In the example, without the support of a secondary source, I think you have analysed the content, "depicted German soldiers as savage murderers" how do you know that the depiction is of savage murder? To know if the killing were mean to represent savagery implies a knowledge of what the people of the time considered savage. For example Cromwell in his description of the massacre at Droghada describes his ordering the torching of a church tower in which some royalists had retreated, to our sensibilities such actions are savage (but Cromwell and the members of Parliament in London (to whom it was addressed and read), did not), but he notes that one of the victims said "God damn me, God confound me: I burn, I burn." to Cromwell that was shocking, because it was taking the Lords name in vain. To know if the killings in the poster represent murder involves a knowledge of the laws of war, which is an historical legal matter (and also a knowledge of what a contemporary audience would think they represented), and so I do not think it is a straightforward description of the content. -- PBS (talk) 13:12, 12 July 2015 (UTC)

The main value of the secondary source here is in determining whether mentioning this at all is WP:DUE. One could equally well pick out a propaganda poster that used the color orange, and write that some posters used a lot of orange. But why would anyone care, unless the use of that color were in some way significant? WhatamIdoing (talk) 22:25, 12 July 2015 (UTC)

Some of you might be amused to learn that in a discussion on another page, an editor with 35,000 edits has just said to me: Since it is a secondary source, it's clearly reliable.—S Marshall T/C 21:54, 13 July 2015 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. No further edits should be made to this discussion.

Television template encouraging OR?

I wasn't certain what to think about this issue, so I thought I'd bring it here for some input! My apologies in advance if my question is a bit nebulous! Basically, I feel in the context of TV/film, there is a tendency to privilege primary sources when Wikipedia generally privileges secondary sources.

We begin with an edit dispute over Humans (TV series), an 8-part TV mini-series starting this Sunday: see Talk:Humans (TV series) for details. It seemed to me and other editors that the infobox should say there will be 8 episodes, as reported in numerous reliable secondary sources. However, other editors feel differently. It appears that common practice with this template (Template:Infobox television) is that the number of episodes is incremented only as they air: see Template talk:Infobox television#num_episodes where discussion moved to. Likewise, the template says the infobox can only include the first-aired date after the episode has aired, not even a few days before when the date has been reported in numerous reliable secondary sources.

Now, there are a number of issues here around, for example, which number is more interesting and useful to the reader, the number of episodes there will be or the number that have been broadcast to date. I may be quite wrong about my feelings over those! But we don't need to get into them here. What strikes me is a more fundamental issue about Wikipedia.

As far as I understand, common practice among some TV-interested editors has become to watch a show and then increment the episode number (and first aired date) in the infobox having done so. I feel the infobox should be a summary of the article content (WP:IBX) and should reflect what secondary sources say. It seems to me very odd and un-Wikipedia-like to privilege this direct reportage of Wikipedians watching a show over what is reported in secondary sources. Someone watching an episode and incrementing the episode number is, it seems to me, at best using a primary source and, at worst, tantamount to original research. The bias in Wikipedia should always be to secondary sources. High profile content in an infobox should be for well-sourced, encyclopaedic content. If someone wants to tally each episode of a TV series in real time, I feel they can do that on a blog/Twitter/show-specific wiki/&c. (I can see WP:BALL is relevant here. I understand events some time in the future are a different matter, but I am talking here in the context of a show starting in 4 days that has been described by numerous reliable sources.)

I have seen this tendency to privilege primary sources and waiting for a show to air in related contexts. At Talk:Jenna_Coleman#New_Name and Talk:Doctor_Who/Archive_24#Jenna_.28Louise.29, you can see an old discussion over the name an actor uses where some editors were arguing we should wait until the first episode with the new name in the credits is broadcast. Likewise, there have been arguments over whether The Matrix should refer to the Wachowski Brothers (as they were credited in the primary source) or the Wachowski siblings (noting Lana Wachowski's current gender identity; see Wikipedia:Gender identity). Thinking about it, I'm more concerned about those examples and that's probably why I'm bothered by the Humans (TV series) example...!

So, what do others think? Is there some existing guidance on primary/secondary sources when dealing with such matters? Should an understandable desire to keep Wikipedia bang up-to-date be encouraging reference to primary sources? Do we wait to see what the credits say, or do we report what others have already said? Am I worrying over nothing? Any thoughts welcome. Bondegezou (talk) 18:06, 10 June 2015 (UTC)

WP:V / WP:CRYSTAL / WP:NOTADVERT all seem to be relevant. If we use the sources to reference all of the episodes before they actually air, we would need to couch it "Scheduled for 8 episodes" not as "8 episodes" -- TRPoD aka The Red Pen of Doom 18:43, 10 June 2015 (UTC)

I concur. Way too many TV show infoboxes state as fact things that are conjectural/projected/planned/announced with regard to the forthcoming or still-ongoing season/series. Happens with films a lot, too. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 00:10, 23 July 2015 (UTC)

Proposal on "rather than OR" wording

I see that one of my edits, in which I removed italics for emphasis, was reverted with the comment that the point being made was important because some people on the talk page "weren't getting it". My point was that overuse of italics for emphasis is generally regarded as poor writing style. Also they make the policy sound like an extension of a talk page argument. Ideally the policy should be based on consensus, not a king-of-the-mountain style struggle to wrest control of the policy from the other side of an argument.

Another thing that has gone missing is that this passage used to say that interpretive claims must be cited from secondary sources "rather than" being based on the editor's OR. That is something that everyone can agree about. After all, the name of the policy is "No original research". When presented with a choice between secondary sources and OR, everyone will choose the secondary sources. But currently the choice is gone and we have the assertion that all interpretive claims must be cited from secondary sources. Period. But what about tertiary sources? I realize that SMcCandlish has just written an essay arguing against the use of tertiary sources for such claims. But this seems to me to be a minority viewpoint, or at least one with less support that what we used to have. I propose that we put the "rather than OR" back in. – Margin1522 (talk) 10:44, 4 June 2015 (UTC)

Emphasis: I think we all agree that over-use is poor style; I'd been of a mind to remove the emphasis from the word "may" there because it seemed to serve no purpose. In the case of "only if", it should remain emphasized, because the rule is crucial (it's one of the most important content policy rules we have), yet frequently missed. There is no "other side of an argument" from which "control" must be "wrested"; I don't understand the point of imagining such a hyperbolic scenario. The policy is emphatic there because new editors (and some not-so new ones) often get this backwards. This is especially true of university and grad school students, and professional academics, all of whom are very used to seeking primary not secondary sources; see related other thread on this page that brings up this very point of confusion. Academic and encyclopedic writing have radically different purposes and bases.

"Rather than": We'd have to dig around in edit history and archives, with regard to the "rather than" bit. The problem with just shoehorning it back in is that the policy has been clarified in multiple ways in the interim. Analytic/evaluative/interpretive/synthetic claims require reliable secondary sources period, rather than everything else of any kind at all, including primary and tertiary sources, not just rather than an editor's own OR. (Another way of looking at it is that it's a form of OR to reinterpret primary sources' novel claims, or tertiary sources' uncritical regurgitation of data, as supporting an analysis or interpretation not found in reliable secondary sources, but it's much harder to explain this to editors than simply saying such claims require secondary sources, and moving on. Our policies aren't supposed to be essay-like.) So, if we want some "rather than" phrase, we need to come up with adequate wording that conveys that concisely. To me, it seems it would be redundant. Maybe it's okay for WP policies to be a little redundant, to get an important point across in multiple ways (see thread above about about adding back in something here that is kind of a restatement of "Wikipedia is not a publisher of original thought"). But that's basically just an argument for adding a different form of emphasis that some would regard as even worse writing than simply italicizing two words.

Tertiary sources: The idea that tertiary sources can be used for these kinds of claims (an idea that never seems to come with clear examples) is the minority viewpoint, and the policy already indicates in three distinct places that secondary sources are actually required. The very fact that you and one other editor commenting earlier (who appears to have come around, since he made a copyedit that clarified the matter even better) had been somehow interpreting "only if" as if it's not there, or doesn't mean what it says, is strong evidence in favor of the need to emphasize that "only if" phrase. See the other discussions, but the short version is: If an otherwise tertiary source is making such a claim then it is, for that claim, not a tertiary source, by definition. Tertiary sources just collect and re-present data from secondary and sometimes primary sources, without significant analysis/interpretation; secondary sources are those that engage in analysis/evaluation/interpretation/synthesis. The problem here is in misconceptualizing primary/secondary/tertiary. They are not innate or necessarily immutable qualities of a work, but descriptions of how they relate to a particular piece of information as it is being used in a particular context. There's an implicit "for" that comes after "tertiary source" (or primary or secondary). A source can be all three of these at once for different claims even in the same article. As one type of example, a topical encyclopedia written by a subject-matter expert is often all three kinds of source, depending on the details in question. A piece of editorial journalism, as another example, may also mix all three types (e.g. tertiary source for various gun-fatality statistics in a sidebar, secondary for analysis of pro- and anti-gun-ownership stances and their motivations, primary for the conclusion reached by the writer about how well a piece of legislation will address a gun-control issue). PS: the WP:TERTIARYUSE essay doesn't "argue against" their use for analytic claims; the policy says not to use them that way. The essay explains the policy and advises on how to apply it (and it's an in-progress draft, so it may not do so inadequately at this point). PPS: You said yourself earlier on this page: 'Anything that is controversial or involves interpretation or judgment has to come from a secondary source. That's what we should be focusing on in this policy.' So, it's not clear to me where any real disagreement lies if we get the terminology straight. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 20:49, 4 June 2015 (UTC)

Here is an evaluative statement from the 1911 Encyclopedia Britannica (the classical example of a tertiary source), reproduced without change in the first version of our article on Anaximander:

Tradition, probably correct in its general estimate, represents him as a successful student of astronomy and geography, and as one of the pioneers of exact science among the Greeks.

The problem with an absolute statement like "only if secondary" is that it only takes one counterexample to prove the statement false. This example shows that it has never been true, since we have been including evaluative statements from tertiary sources since the very start of Wikipedia.

Nor do I think it's helpful to split hairs over "secondary statements" embedded in "tertiary sources", or engage in similar kinds of legerdemain. The only thing this policy needs to do is explain to editors that they can't say "probably correct" in their own voice. – Margin1522 (talk) 08:50, 5 June 2015 (UTC)

Aren't we much more concerned about it being said in Wikipedia's voice? It's usually easy to see when people insert their own random opinions, but much harder to detect when a reasonable-seeming evaluative statement is synthesized from a pile of sources but doesn't actually appear in any secondary ones. It can take hours of source verification to detect. A statement like the Anaximander one might be followed with 3, or 5, or 7 citations, all of which show that tradition says that about him, but none of which suggest the tradition is "probably correct" (or maybe one does, but it's a primary source). Cases of incautious use of material like Britannica 1911 has led to some inclusion of synthesis untraceable to a secondary source (if you want to view it that way) doesn't "disprove" that we have a rule that requires reliable secondary sources for it. The fact that a majority of people drive a little over the speed limit doesn't disprove that there's a law against it, it just demonstrates lax enforcement and a human penchant for minor transgression. A lot of Wikipedians probably don't even realize they're transgressing when they take the average view of the sources they've bothered to look at, and if those are reasonably consistent, say something like "The majority of sources agree that ..."; we need to keep saying this isn't the way to do it, and maybe say it more clearly. The essayists in many of us don't like writing of the form "Many sources, including X, Y, and Z, agree that ..., while others, such as N, conclude instead that ..."; but it's often necessary here, when the views in the RSes conflict and there's not a WP:UNDUE problem in reporting the disagreement. I don't see a way around that. PS: The "old sources become primary" theory, toward which WP seems to lean, would at some point (perhaps already) treat a century-old encyclopedia with laxer editorial standards than ours as a primary source anyway, no longer tertiary. It makes me wonder how much Britannica 1911 material we still have around, not yet re-worked and re-sourced. There aren't many topics, even figures of classical history, about which absolutely no scholarship has advanced in over 100 years. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 19:43, 8 June 2015 (UTC)

I have to say that I see no benefit whatsoever in redefining the EB1911 as a primary or secondary source as convenient, depending on whether we want to keep the passage in question or throw it out. Of course research into the presocratic philosophers has advanced in the past 100 years. Recent sources may be better. For that reason, we may want to say that the EB is no longer a "reliable" source. In that case, I would suggest taking that question up in the appropriate forum. The goal of this policy is is not to offer a watertight definition of "reliable", or to close every "loophole" that might allow editors to include content that someone doesn't like in the encyclopedia. This policy is supposed to be about original research. There is no sense in which citing the 1911 Britannica for a statement like the "probably correct" one above is original research, and this policy should never be used to prevent someone from doing that. – Margin1522 (talk) 02:37, 11 June 2015 (UTC)

It has nothing to do with whether we want to keep it or throw it out, only what kind of claim is being made. Well, there is one other factor too: Age. By the "100 years" rule observed by most serious researchers, Britannica 1911 has to be treated as a primary source, because it is too close the events about which it was writing (on then-modern topics), and too much new research has been done in the intervening century-plus. And that 100-year cutoff (often much shorter, more like 20 years) actually is extended to secondary sources, not tertiary ones; those actually date much more quickly. The 1911 edition has been regarded as primary for several generations now. It's tertiary only in tone and approach, not for any reliability analysis. Cf. #RfC: Should "news articles" be added to WP:PRIMARY?. Unanimously against treating all news sources as primary "the way historians do"; as various commenters pointed out, it's older news sources that have to be treated as primary. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 00:14, 23 July 2015 (UTC)

"an analytic, evaluative, interpretive, or synthetic claim"

Scientific papers are primary sources and they usually do contain analytic, evaluative, interpretive or synthetic claims. For example, our article on HIP 56948 (a solar twin) is based largely on papers by Jorge Melendez and others published in 2007 and 2012; there are journalistic sources but all they do is parrot the primary source in dumbed-down language. Professor Melendez and his colleagues gathered the data, evaluated the results and interpreted them for us. And that's as it should be when your source is a full professor at a major university who's the world's leading authority on a narrowly-focused subject.

I can't square this with the bit of WP:NOR that says "Articles may include an analytic, evaluative, interpretive, or synthetic claim only if the claim has been published in a cited reliable secondary source." Actually I think that articles can and should repeat the primary source's evaluative claims ---- provided the primary source is by a scholar or academic who's acknowledged as a leader in their field.—S Marshall T/C 00:13, 11 July 2015 (UTC)

FYI... this has been discussed before. The problem is that the OR policy was mostly written by editors familiar with the humanities... and in the humanities a paper such as you mention would be considered secondary ... the paper is analyzing and interpreting original data... and that date would be the primary material. the data is the equivalent of original documents used in a history paper... primary material which is then analyzed and interpreted in the secondary academic paper.

That said... remember that we are explicitly allowed to describe a primary source as long as we don't interpret or analyze it ourselves, or draw our own conclusions from it. So it does not really matter whether we call Professor Melendez's paper primary or secondary. As long as we don't analyze it ourselves, it's not OR to state what a source says. This is why I keep raising on the concept of "appropriateness"... it does not matter if a source is primary or secondary if we use it appropriately. Blueboar (talk) 01:22, 11 July 2015 (UTC)

Discussed before but not fixed, I see. How about "An analytic, evaluative, interpretive, or synthetic claim not found in the primary source itself"?—S Marshall T/C 13:44, 11 July 2015 (UTC)

See our article District of Columbia v. Heller. The article cites the US Supreme Court decision in the case, as well as secondary sources. The decision itself has many characteristics of a secondary source; the Supreme Court never conducts experiments or visits the scene of incidents, nor does it take testimony. It reads the decisions of lower courts, and considers written briefs and oral testimony from the parties' lawyers about how to interpret the law and lower court decisions. Then it decides. It is the fact that the Supreme Court makes a binding decision that makes it a primary source; otherwise it would be a secondary source. And of course the decision is chock full of evaluations. I like S Marshall's formulation at 13:44 UT. Jc3s5h (talk) 14:03, 11 July 2015 (UTC)

I, too, think that you make a valid point and approve of your suggestion, S Marshall. –Prototime (talk · contribs) 21:26, 11 July 2015 (UTC)

Why it's become so confused for Wikipedians: It's about the content not type or name of publication

Wikipedia's own definition of primary and secondary (and tertiary) do not change with the topic. It's simply not possible, but I keep seeing people repeat it as if it were true. WP has a standard for how to determine source reliability level. It is WP:RS. Various academic, journalistic, etc., fields, and individual organizations, have their own standards. Their standards do not overturn each other. They don't overturn ours when it comes to how we evaluate sources. Ours doesn't overturn theirs in their home court, either. There is no primary vs. secondary difference between a humanities journal article or a physics journal article for Wikipedia. If you ask a humanties academic, they'll classify them both the same as each other for their purposes, and a hard-sciences professor will do likewise, and when you compare the two academics' assessments of both works you find that the two individuals categorized them differently. If you asked a panel of experts in their two disciplines to repeat this process, you'd get the same result. BTW, see the RFC a bit above where the pretty substantial !voter turnout unanimously rejected the idea than external classification preferences from a particular discipline (all news sourcing as primary, in history and related fields) can affect how WP chooses to classify sources for our own internal purposes. It can't happen because our criteria are different. We're mostly asking "where did they get this claim from?", and also making sure that the publisher is reputable enough to bother with. In science, the prestige of the journal and its editorial selectivity is paramount; they want to see the most amazing and best-vetted primary sources, and don't care so much about secondary literature reviews; the groundbreaking papers are the "sexy" party. WP wants the reviews, and we don't trust "groundbreakingness" until secondary sources tell us we can.

Next, a source, as a object, is not immutably and always either a primary, or secondary (or tertiary) source, for all time, in all contexts, for all facts. Just the opposite. RS is poorly written with regard to this and it has led to years of confusion. We need to fix it. When we say "this is a secondary source" [or whatever - don't fixate on "secondary" here] this means "this is a secondary source, in the present, for the following claims, in this particular context". It cannot sanely mean anything else. Everyone in academia (and several other forms of publishing) understand this. A reputably-published newspaper article (not op-ed – we can't verify independence!) is a secondary source for the material it has WP:AEISed from various primary and other secondary sources and run through a quality editorial process. If the journalist closes with a personal conclusion of her own, it is a primary source for that view, which must be individually attributed to or directly quoted from that writer. Where it presents "factoids" that were not subject to any AEIS, but reguritated (e.g. the fact that scientific name of the domestic cat is either Felis catus or Felis silvestris catus depending upon the source consulted, something she got from an encyclopedia, her article is a tertiary source for that information. News journalism is a tertiary source for virtually all "hard science" material (including news); it is summarizing and glossing over many nuances in the actual secondary (and maybe some primary) peer-reviewed material, if it even consults any, rather than abstracting from other tertiary science sources (there are many other tertiary factors at work, too, but it would be lengthy to cover them right here).

A serial publication is not uniformly secondary, either. Movie reviews, columns (at least typical ones), op-eds, and much of the content of editorials, along with any other opinional material, in news sources is primary. Subjective opinion does not magically transmogrify from opinion into fact just because of who churned it out. But it's staggering how many Wikipedians really do believe that every single word published by the New York Times or whatever must be secondary sourcing, because of how terribly this guideline is written.

Same goes for journals. There's this counterfactual belief that secondary material in a paper is somehow transubstantiated into primary material, if elsewhere in the work it's also presenting primary material (the point of all that secondary research they did). Peer reviewed AEIS is peer-reviewed AEIS; if it's from a reputable journal (and not outdated), it's high-quality secondary sourcing, as long as the secondary and primary material are distinct and not commingled. Bracketing it with a proposition in the intro and a conclusion at the end that give the publisher's position on all that researched material does not "degrade" the entire work into "primaryness". Any new data, results, hypothesis, model, description of methodology, etc., that is not AEIS of other high-quality sources, but has come from the author(s) of the paper (is their "original research") is necessarily primary. (Exactly the same thing is true of a journalist injecting their own views into an investigative piece.) Often, the same generally high quality source is mixture of all three; any topic-specific encyclopedic compendium by subject matters experts, for example, will mostly be tertiary, but contain plenty of the other two categories. Obviously a publisher is not a source, and isn't "reliable" (they do a good job printing?) but reputable [or not]. And age of the source reduced what classification level a source can be held at; all sources eventually become treated as if primary, no matter how secondary or tertiary they were at one time. You can't cite a 1941 article on particle physics or the causes of a riot as a secondary source, but attribute and use with caution as primary.

This isn't as simplistic as it's been made out to be by WP:RS's poorly written bits at WP:PSTS, with consequent related inclarities in WP:V and WP:NOR. But it's easy to figure out if you just it off and start fresh. For the claim in question, no matter where it is reputably published and in what form: Did the writer come up with idea/data (or can we not be sure)? Primary. Did they work at producing a complete picture from primary and/or previously published secondary material, without a serious bias? Secondary. Are they just summarizing, collecting, or repeating previous material without AEIS? Tertiary. Is it clearly opinional, hypothesizing, reaching? Primary. Was it produced close in time to the events it is about? Primary. Is the source categorically obsolete, or factually or culturally surpassed? Primary. For the source classification analysis, it doesn't matter one bit who the authors were, who the publisher was, or what the topic is (unless these interact to produce a conflict of interest (primary); those are otherwise factors for the quality level analysis (low to high): the reputations of the publisher and writer, and their competence/expertise in the topic.

I've gone into some other factors at WT:V and WP:RS (with some overlap). — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 01:42, 23 July 2015 (UTC)

When to fact tag vs. just delete

FYI

– Pointer to relevant discussion elsewhere.

This topic has had a lot of churn on various pages lately, so I thought people here might be interested in this discussion: Template talk:Citation needed#When to remove unsourced info vs. when to add this tag?. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 05:23, 23 July 2015 (UTC)

OR making its way into mainstream thought?

Are there any examples of WP:OR being taken up by mainstream media, or even scientific literature?

e.g. H writes a paper on a new scientific discovery, but either do not name it, or give something obscure. User J writes an article about the discovery, but not knowing what to call it, invents a plausible name (OR). The popular press then picks up on the discovery, and uses the name from wikipedia as they don't have another name to use. H then writes more papers on the subject, eventually using the name from the popular press, and thus wikipedia. Martin451 21:07, 28 July 2015 (UTC)

Not sure about Wikipedia as a source but such things do happen. See e.g. the Iowa gambling task. Arnoutf (talk) 21:33, 28 July 2015 (UTC)

If this ever does happen (or if it has already happened) I would simply apply WP:Ignore all rules and WP:COMMON SENSE to deal with it. Obviously we would like to remove an Original Wikipedia invented name before the media picks up on it... but if the Wiki-originated name is being used extensively (especially if it is now being used by H himself)... it's kind of too late to put the genie back in the bottle. Blueboar (talk) 00:44, 29 July 2015 (UTC)

It may even be the case that Wikipedia editor J is a scientist who is writing a scientific paper using the new name, or has heard the new name mentioned on a symposium/conference from scientist X (or maybe even H themselves as they may have reconsidered the name). As scientific publishing generally takes several months from completing a first draft to publication, J's Wikipedia edits may seem to come earlier than publication in a reliable source; but may in fact be informed by an emerging name in the scientific community not yet available in a reliable source.

In any case it does not much matter where the scientific community picks up their inspiration in naming ideas - which are often influenced by popular culture (see e.g. 26734 Terryfarrell); but as Blueboar states above, we should remove self created names not yet picked up broadly. Arnoutf (talk) 07:34, 29 July 2015 (UTC)

Can we really state that x book doesn't contain Y statement with no source other than the book?

If we can't, I'd like something explicit in policy, as I'm being told it's basically ok as you can read the book. Thanks. Dougweller (talk) 15:19, 7 May 2015 (UTC)

Isn't this too specific/detailed an issue to include in such an high level policy. Arnoutf (talk) 17:42, 7 May 2015 (UTC)

No, it's not too specific. And yes, we can say that. Books (and TV shows, etc.) are their own sources, which anyone can verify, albeit not always without some difficulty (e.g. inter-library loan). Our sourcing rules require that sources be publicly available (you can't cite an unpublished manuscript, or something that was published in a small quantity and then all of them destroyed). Our facts have to be verifiable, but sources are often behind paywalls and otherwise not trivial to fact-check, and that's ok. The only quasi-problem with "X book does not contain Y statement" is that if someone wants to prove it beyond all shadow of doubt, they have to read the entire source book. In practical application, this is not really an issue, because books tend to be divided into chapters or other sections. It doesn't take much work at all to discover that a particular dictionary has no entry on a neologism like "sexting", that the Star Trek novel The Rings of Time does not end with Scotty's death, that W. H. Murray in Roby Roy MacGregor: His Life and Times does not say that Rob Roy grew up in Clan Campbell, and so on. Most such claims can be verified by reading a single section/chapter. If someone is attempting to game the system by frequently including dubious negative assertions of this sort in a bad-faith exercise intended to waste many hours of editors' time reading entire books to disprove the false assertions, this will become apparent quickly, and is a user behavior and disciplinary matter for WP:ANI, not a flaw in the WP:NOR guideline, which like our other rules assumes good faith about editorial approaches to it. All that said, it's obviously better if we indicate what a source does write rather than what it didn't. "Sources A and B say X, while source C indicates Y" is far more useful to the reader than "Sources A and B say X, while source C disagrees". It's just begs the question, "disagrees in what way(s)?" — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 09:20, 11 May 2015 (UTC)

I outline some exceptions to that last statement, in later discussion. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 05:31, 29 May 2015 (UTC)

For the claim that "x book doesn't contain Y statement", x book is a primary source. In the 4th paragraph of the section Primary, secondary and tertiary sources, which begins with "Policy: Unless", is the following,

"A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge."

An editor may be prohibited from stating that x book doesn't contain Y statement with no source other than the book, by adding to the above excerpt from policy the underlined text in the following.

"A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified in a straightforward way by any educated person with access to the primary source but without further, specialized knowledge."

--Bob K31416 (talk) 22:46, 12 May 2015 (UTC)

The addition of poorly defined words to a sentence that has problems already won't help the situation. Frankly I don't think a general rule can cover the complexities of this issue. Often it will be WP:SYNTH to write that a book doesn't contain something, but not always. It's easy to think of examples that would be perfectly fine and examples where it wouldn't be. Zero^talk 01:08, 13 May 2015 (UTC)

Would you care to give the details of your comment? --Bob K31416 (talk) 03:13, 13 May 2015 (UTC)

Imaging there is a book on woodworm in the USA. There is a list of cities where woodworm is a problem. The fact that the book omits a particular city from the list is citable (subject to WP:WEIGHT etc), but it could still be SYNTH depending on the circumstances. The fact that the book doesn't explain why most of the cities are in warm states is not citable, since it requires not only reading of the whole book but interpretation of the content. As for problems the sentence has already, think of our articles on quantum physics. Many of them rely on sources that need a higher degree in physics to understand. A literal enforcement of "without further, specialist knowledge" would essentially eliminate those articles to the great loss of the encyclopedia; same for advanced mathematics, etc. I tried to fix this problem several times, but getting consensus on changes to core sentences is very difficult. Zero^talk 03:30, 13 May 2015 (UTC)

Re "The fact that the book omits a particular city from the list is citable" — That would be allowable according to "that can be verified in a straightforward way” .

Re "The fact that the book doesn't explain why most of the cities are in warm states is not citable, since it requires not only reading of the whole book but interpretation of the content.” — That would not be allowable according to "that can be verified in a straightforward way” .

Re "As for problems the sentence has already, think of our articles on quantum physics. Many of them rely on sources that need a higher degree in physics to understand.” — But those are secondary sources, not primary sources, the latter being the type of source we are discussing. Recall that the sentence we are discussing begins with "A primary source".

--Bob K31416 (talk) 13:54, 13 May 2015 (UTC)

While I agree that it is not illegal to do this per the reasons SMcCandlish provided, I wonder if there are many legitimate cases in which it makes sense to do so without it ultimately being synthesis. While The Rings of Time may not end in Scotty's death, it doesn't make sense to include that statement by itself in the article. That book probably also doesn't include Cyndi Lauper or literally billions of other things, but we wouldn't include those omissions either. Now, if for some reason most Star Trek novels ended in Scotty's death and this one was notable in that it didn't, then that statement would require some sourcing as there are several assertions being made apart from the content of the one primary source. To give another example, an article could make two statements like "Dubai is a city. It is not included in Godot's Comprehensive Handbook of Important Cities of the World." Those two statements, while both true and neither requiring a secondary source, together imply that Dubai is not an important city. Regards, Orange Suede Sofa (talk) 03:49, 13 May 2015 (UTC)

The example I had in mind was the statement The Men Who Stare at Goats that the book, which is about the Stargate Project, doesn't actually name it: "This was the Stargate Project[1][2][3][4], which the book never mentions by name." How practically can that be verified? How many times would you have to read the book to be sure? I guess you could if it had an ebook version. The editor insists he can use the book as a source. Dougweller (talk) 14:31, 13 May 2015 (UTC)

One could argue that looking for the words "Stargate Project" in the book and concluding that it is not there, is analysis that reaches a conclusion not stated in a reliable source, and thus violates WP:NOR.
Regarding the part of policy that says, "A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge." — Because of the word "only", this is a necessary condition for using the book as a primary source for the statement, not a sufficient condition. So one could not argue from this that the statement should be included.

If you want to clarify this policy to exclude such a statement, then the change indicated by my first message would do that, i.e adding to the above excerpt from policy the underlined text in the following.

"A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified in a straightforward way by any educated person with access to the primary source but without further, specialized knowledge."

--Bob K31416 (talk) 17:22, 13 May 2015 (UTC)

One cannot reasonably 'argue that looking for the words "Stargate Project" in the book and concluding that it is not there, is analysis that reaches a conclusion not stated in a reliable source, and thus violates WP:NOW.'. To actually apply that logic would be precisely the sort of absurdist, ultra-literalist misinterpretation of policy, a form of WP:GAMING the system by distorting the spirit and clear meaning of WP:POLICY, that Zero0000 observes would also effectively disallow specialist sources to be used at all. Reading a book [or otherwise examining a source in the normal way, e.g. watching a film] is precisely "verifi[cation] in a straightforward way". There is no possible more straightforward way to verify what a book says or omits than by reading it. Of course, observing that a source does not cover something can be WP:UNDUE weight, and can involve WP:OR, per various examples above, like the Dubai one; but there are obvious cases where it does not. E.g., in an article on a classic film, it would be fine to include a sentence of the form, 'It was listed in the "Top 100 Films of All Time" published by the New York Times in 2013, the "Top 50 American Films" list of the Motion Picture Academy of America (as of 2015), the "Best Films of the 20th Century" by Entertainment magazine, ... [5 more such lists here], but did not make the American Film Institute's list of "50 Most Important American Films of the Last Century".' In fact, omitting that omission might well be a WP:NPOV problem, an example of cherry-picking critical coverage to artificially inflate the acclaim and importance of the subject. I'd almost be willing to bet money that we already have a number of film articles with exactly this kind of sentence/paragraph, with that sort of balance, and I'm 100% certain that I've seen song/album articles with similar constructions, noting that songs or albums charted in certain markets, but not others (i.e., they do not appear on certain Billboard and other charts, i.e. are not mentioned in specific sources).

As I observed earlier, but didn't spell out very clearly, the length/size of the source isn't relevant. While it's easier to see what is or is not in a list of 100 songs or films than whether a name appears in a long work like War and Peace, it's also easy to see what does and does not have an entry in a truly enormous work like the Encyclopaedia Brittanica. There is no principle in WP:NOR or WP:V (or other) policy, or the WP:RS guideline, that verification has to be "easy" or "convenient"; in fact, the opposite is true, and citing sources that are quite difficult to even obtain is still permitted, as is citing facts from academic journal articles that no one but a subject-matter expert is likely to be able to fully understand correctly (without engaging in applying an interpretation of the facts therein; that requires secondary sources).

Note that this sort of straightforward verification of what a source does or does not contain, and using this information in a neutral, non-synthesis manner, is entirely distinct from the fairly frequent error of doing some source research and concluding something like "No sources mention...", or "This idea is unknown to science", or "It cannot be verified that...". The problem with such patently OR negative assertions is that they presuppose that someone has verified every possible source that could imaginably exist. The fact that we can't do this can be frustrating at times (especially when dealing with pseudo-science and fringe material), but we just can't. We can cite an external authority stating that science can't prove something, or whatever (subject to WP:UNDUE, of course), but we can't come to this conclusion ourselves in Wikipedia's voice. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 20:13, 14 May 2015 (UTC)

Re "There is no possible more straightforward way to verify what a book says or omits than by reading it." — With regard to verifying that a book doesn't contain some material, yes there is. By having a reliable source that states that the book does not contain the material. For an editor to determine that the material is not in a book by reading the whole book, including all the footnotes, intro, forward, etc., or just checking the parts of the book where one thinks it can be, is subject to possible mistakes. Avoiding these kind of mistakes by Wikipedia editors is a reason for the policy No original research. --Bob K31416 (talk) 22:56, 14 May 2015 (UTC)

There's also this thing called a "search" tool. Get an eBook version (or Google Books version) and search for it. As Blueboar says, the problem here is WP:DUE, not NOR. WhatamIdoing (talk) 16:23, 24 May 2015 (UTC)

The book The Men Who Stare at Goats, apparently does not have an ebook version.[1] Even if there was an ebook version, you couldn't be sure that you or the editor who did the original research for the edit used the right keywords and sequence of keywords in the search to determine that the book didn't mention the Stargate Project because it could have been mentioned in the book by various other names, e.g. Project Stargate, Project Star Gate, Star Gate Project, or names that are synonymous but completely different. --Bob K31416 (talk) 23:13, 24 May 2015 (UTC)

Yep, and these are seriously lame semantic nitpicks. The odds that book A actually says something like "book B does not contain statement X" are pretty close to astronomically low. It's like responding to "I feel very safe with my new home security system" with "But what if your house was hit by a comet?" Aside from Bob K31416's later correct observation, the search thing is pointless argument. I didn't refer to the most expedient way to check book content, I said most straightforward. And besides, reading online book content pulled up by searching the e-book version is – guess what? – reading the book. I didn't say reading the paper edition of the book. JFC, people. If the best anyone could think of to try to refute, poorly, were these two bits of trivia, then I guess my detailed argument is pretty well-reasoned. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 05:20, 29 May 2015 (UTC)

I am not sure we are asking the right questions here... before we ask the question: Is it OR to cite Book X itself for the statement "Book X does not contain Y statement" ... I think we should be asking why we need to mention this fact in the first place - is it is APPROPRIATE to say it? That's more of an WP:UNDUE question than a WP:NOR question. Blueboar (talk) 16:13, 15 May 2015 (UTC)

It is in violation of spirit if not the expressed letter of both OR and UNDUE. Calling out that X isnt in the source is discussing X out of context of any source. " Similarly, do not combine different parts of one source to reach or imply a conclusion not explicitly stated by the source." when the source does not contain anything about X using it as a source to comment about the lack of X is utilizing a source to support a claim that is not explicit in that source. and if no sources have commented on the lack of X in source Y, for Wikipedia editors to do so on their own is clearly placing UNDUE emphasis on the fact that Y doesnt address X. -- TRPoD aka The Red Pen of Doom 22:19, 22 May 2015 (UTC)

I think this is true in most cases, but I can think of plausible exceptions. An article on the Crusades could list the main chronicles and indicate which of them cover the siege of Acre. This is not OR but just the normal editorial process of summarising the sources. (However it might be OR to draw conclusions from the list.) An article on a dead person suspected of murder might mention that that person's autobiography does not admit it. This is verifiable information that would interest readers. Zero^talk 02:15, 23 May 2015 (UTC)

It could be an UNDUE problem, but there are many cases where it would not be. Another trivially simple example: In describing the traits of an standardized animal breed, where the US fancier organization has one published standard and the UK one has a slightly different standard, that differs from the US one only in the omission of a couple of criteria among dozens, it's entirely reasonable to explicitly state that the breed standard WP is using in the article is the US one except where noted, and then note at each omitted trait that it's not in the UK standard. This is a real example; I did this just the other week, and I stand by it.

We already have a well-developed supplementary essay that covers this: WP:These are not original research#Works of fiction and non-fiction. Note carefully that it even includes an example addressing precisely the kind of case under discussion. I've been editing or 8 or 9 years or something with precisely the same interpretation as almost everything on that page, and only noticed the essay yesterday. I'd say it's pretty solid (well I did edit a few bits of it upon first read).

Anyway, yes, really pointed, POV-pushing abuse of the ability to mention that something isn't in a source would be problematic, but we have WP:NPOV policy for that, both generally and for undue weight in particular. An example, to direct contrast with one I gave earlier, might be trying to show that a positively reviewed album lauded by many critics wasn't all that great in your opinion, because Rolling Stone's "Top 10 Albums of the Year" didn't include it, and doing this not in the context of a paragraph mentioning the top-whatever lists, and other accolades, that it did receive, but just stuck in there as an isolated, doubt-casting negative. These kinds of things are really easy to detect, though. I.e., there is no problem here to address. If we really wanted to, it wouldn't be hard, though, to draft language distinguishing these kinds of cases, and we already have enough pro and con examples to work with. The simple act of citing what a source doesn't contain isn't problematic; it's a problem when it is done it in a non-neutral way, or a synthesizing or unverified assumption-making way ("this academic review article didn't cover paper X, so the theory in it must have been rejected by the scientific community...").

In closing, I agree that, in the original question of the thread, that The Men Who Stare At Goats is its own primary source for the statement that it doesn't name the Stargate Project. This sort of "negative citation" is totally routine. We have thousands of these, at least. But I also agree with Zero (as did Bob K, with an additional reason) that '"The fact that the book doesn't explain why most of the cities are in warm states is not citable, since it requires not only reading of the whole book but interpretation of the content.”' (Bob K added: 'That would not be allowable according to "that can be verified in a straightforward way”', which which I also concur). — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 05:20, 29 May 2015 (UTC)

Sorry, but I think your comments have been confused and I stand by my previous remarks. --Bob K31416 (talk) 13:23, 29 May 2015 (UTC)

Fallacy of proof by assertion. You demonstrate no such "confusion" on my part, and have not refuted any point I've made, in very step-wise logic. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 01:29, 3 June 2015 (UTC)

Reliable sources reliably published

We keep going from "reliable primary sources" to "primary sources that have been reliably published", and back again. I think we're confusing people here.

The probable interpretation for "reliable primary sources" is pretty straightfoward:

A primary source that meets WP:Reliable sources

The other phrase, "primary sources that have been reliably published", is more confusing. Does that mean "a primary source that is not self-published"? This would rule out all WP:SPS sources, definitely including everything on Twitter or a personal website. Does that mean "A primary source that has a reputable publisher"? That would rule out most WP:SPS sources and a good number of articles written by experts but published in weak publications (like corporate or advocacy websites, which do not have "reputable" publishers). Does that mean "A primary source that has a multi-person editorial staff"? That would rule out all SPS sources plus all tiny publications (probably anything with a single editor, and definitely any news outlet whose editor also writes articles).

I think we can figure out what to say, if we can agree on what we mean to say here. Here's what would help me: Give me a couple of examples of sources (real or hypothetical) that you want to include or exclude here. Once we've agreed on what we're trying to accomplish, we should be able to find words to express that. WhatamIdoing (talk) 16:53, 24 May 2015 (UTC)

It should definitely be "reliable primary sources". I've been meaning to fix this for some time, but keep forgetting. The "sources reliably published" thing is someone trying to avoid repetition, but incautiously, at the cost of radically changing the meaning. We don't need to avoid repetition at the cost of much of anything; our policy pages are not novels or journalism pieces, they're technical writing, and a page like this needs a high level of precision, even if it's not a joy to read. This "reliably published" confusion does in fact frequently lead to attempts at WP:GAMING the system, such as assertions that a source with known reliability problems is still fine, because the publishing company is generally reputable. I encountered this very argument today, spilling over into confusion about secondary sources. Someone asserted that two outlying sources must be authoritative (despite others having shown they were contradicted by all actually reliable sources on the topic), because they came from usually "reliable" publishers, and this somehow made them unassailable. Well, to hell with that nonsense. It's a logically bankrupt idea, no different from an assumption that because PBS has a reputation for producing quality documentaries, that everything they've ever aired is a reliable source. We don't even concede that everything ever written by a particular, usually reliable, author is necessarily a RS, nor even that everything in a generally RS is necessarily reliable, so it's logically impossible for us be believe that general positive regard for a publisher can magically bestow reliability on everything it publishes.

At any rate, it can't really mean "a publisher that is reliable" because that doesn't even make sense. A "reliable" publisher would be one that fulfilled its obligations, i.e. met deadlines, delivered books without missing pages, etc. There is no such thing as a publisher that is intrinsically reliable in the source sense. Source reliability is about content, authorship, independence, and editorial control (i.e. fact checking). Besides, WP has no real means of or criteria for assessing whether a publisher is "reputable", which isn't even what the wording says (NOR doesn't contain that word), though we go about this informally to an extent, sometimes. We mostly just care that the publisher is not (except where SPS is okay) the author (directly or through a self-published book outfit), and is not institutionally tied directly to the subject (house organ), or institutionally grossly biased (e.g. a fundamentalist religious organization publishing a "science" book that contradicts actual science). WP:RS has this covered well enough. RS does not contain "reliably published", nor do WP:V or WP:NPOV. NPOV does not mention publishers. V, at WP:SOURCE, suggests that publishers "can affect reliability", but does not elaborate in any way. V says the same thing and elaborates, but only by stating "Reliable sources may be published materials with a reliable publication process, authors who are regarded as authoritative in relation to the subject, or both." But "reliable publication process" is a confused and essentially meaningless phrase in our context. The plain-English interpretation is that as long as their printing presses work properly and they manage to print books with the covers on straight, that this somehow increases reliability. Nuts. It's hard to be absolutely certain what RS is trying to convey there, but it's probably something like "editorial process with reliable fact checking". I guess we'll have to post over at WT:RS and see what the regulars there (to the extent they aren't the same as the ones here) say. Anyway, to come full circle, NOR should say "reliable primary sources", even if RS wants to make some point about publisher reputation lending some reliability points. How to determine reliability is a RS matter; what kinds of sources to use (i.e., reliable ones) and how to avoid OR is an NOR matter. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 04:28, 29 May 2015 (UTC)

Hi User:SMcCandlish. No source is reliable per se, but the reliability always depends on the context where you want to use the source. A source might be reliable for one purpose, but unreliable for another. But I disagree with your interpretation, though. For example, are you saying that "Oxford University Press" isn't reliable? Jayaguru-Shishya (talk) 12:15, 29 May 2015 (UTC)

@Jayaguru-Shishya: I never said any source is reliable per se, and of course reliability is context-dependent. I guess you simply didn't notice that I said: "We don't even concede that everything ever written by a particular, usually reliable, author is necessarily a RS, nor even that everything in a generally RS is necessarily reliable". As in our discussion at another thread on this talk page, you do not appear to be carefully reading what I wrote, and don't seem to understand my interpretation, so I'm skeptical that you actually disagree with it in any practical sense. Re: 'are you saying that "Oxford University Press" isn't reliable?' – That's not actually a meaningful question.

I'll explain in more detail why "reliable publisher" is a meaningless phase on Wikipedia, and a compound logic error. We have an internal definition of what a reliable source is. WP:RS is a whole guideline defining it and the process by which we evaluate source reliability. One aspect of that evaluation is whether the source is from a reputable publisher, i.e. one that is well-established and has a well-regarded reputation. (All three of those italicized terms are used to describe reputable publishers in WP:RS, while "reliable" never is.) We have no concept on WP that the publisher itself is "reliable". The wording confusion going on here, changing the idea that all reliable sources come from reputable publishers, to the nonsensical construction that all reliable sources come from reliable publishers, is exactly the same copyediting error as changing the construction "all tailless cat varieties arise from insular environments" to "all tailless cat varieties arise from tailless environments". The only difference is that "reputable publisher" isn't as laughably bizarre a construction as "tailless environments", so we don't notice the problem as quickly. The problem in this policy text has now been noticed, so we're going to correct it. The copyediting error (which arose probably just because both words are of the "r...le" form, and relate to trustworthiness, thus are easily confused) results in circular reasoning in both cases, compounded by the factual error of mislabeling the publisher/environment half of the equation with a quality that pertains to things that came from it (this is a variant of the fallacy of composition).

WP absolutely, positively does not treat all output from a reputable publishing company as reliable sources. (An obvious proof of this is that an obsolete source from a reputable publisher is not a reliable source, even though the publisher is reputable; Q.E.D.) We do tend to presume they're more likely to be reliable than material published by less well-regarded publishers, but that's not the same thing. Publisher reputability is just a factor in determining source reliability, not proof of it. Secondary source reliability is principally a matter of the expertise and research competence of the authors/editors, with various other factors weighed, like age of source, biases, target audience, and depth of the work, among others, including public perception of the publisher's reputability. That last point is important. Publisher reputability is something external to Wikipedia, while (as noted earlier) source reliability is something we determine internally for WP's own purposes by WP's own criteria. It's logically impossible for these two things to be equivalent. Reliability (a WP concept about a specific source when used for specific facts in a specific article context, as you noted yourself) cannot magically transfer itself out into the world and adhere to an entire publishing company and everything it produces. PS: The exact phrase "reliably published" is even more confused and meaningless than "reliable publisher"; it's simply irrational gibberish, and it has to go. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 02:04, 30 May 2015 (UTC)
P.S.: Here's some further evidence that some editors just get "reliable" from "reliable sources" stuck in their heads and misuse it. I just encountered, without looking for any such examples, ""Domesticated rats are physiologically and psychologically different from their wild relatives, and, when acquired from reliable sources (such as a breeder), they pose no more of a health risk than other common pets.", in the article Fancy rat. I sanitized this (in the sense of sanity, not cleanliness) to "...acquired from a reputable breeder or pet shop". — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 18:23, 31 May 2015 (UTC)

Specific proposal

Per the above discussion, clarify and normalize the following:

"Unless restricted by another policy, primary sources that have been reliably published may be used in Wikipedia"

to read:

"Unless restricted by another policy, primary sources from reputable publishers may be used in Wikipedia"

so that the wording agrees with WP:Reliable sources.

Reputability of a publisher (how well-regarded it is) is directly addressed at WP:Reliable sources, but the terms "reliable"/"reliability" are never used there to refer to a publisher, and as detailed above, cannot logically apply. In short form: Sources may be reliable (a determination Wikipedia makes internally, with a whole guideline full of criteria); a publisher may be reputable, a quality that WP:RS indicates we observe about the off-WP world's regard for the publisher. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 01:57, 3 June 2015 (UTC)

Support but suggest as a perhaps preferable alternative that this and any similar text might read "Unless restricted by another policy, reputable primary sources may be used in Wikipedia"
This is pedantry on the basis that many primary sources may be reputable but not many will be publishers. Ping: SMcCandlish. GregKaye 15:44, 8 June 2015 (UTC)
- That would be a "little big" change in the meaning in two ways. It shifts the external criterion (well-regardedness, a.k.a. reputability, a.k.a. respected editorial control) from the publisher to the source itself, and removes the internal criterion (reliability, determined by our WP:RS rubric) entirely. It would thus permit well-regarded self-published sources, whether we thought they were reliable or not. The reason we want reputable publishers for primary sources is that there's a huge difference between a peer-reviewed journal article presenting new research on MAO inhibitors, and some guy's movie review blog even if it has a zillion followers. >;-)
Does this have the effect of deprecating self-published primary sources? A corporate or even a personal website might be a very appropriate and reliable source for some facts (e.g., name of the CEO or products sold for a business, the birthdate for a BLP subject for a personal website [assuming it's the BLP's own personal website]) but a self-published source cannot really be said to have a "reputable publisher". WhatamIdoing (talk) 08:05, 20 June 2015 (UTC)
Oppose This would exclude legitimate WP:ABOUTSELF sources that are primary and not reputable publishers. As I said months ago, there's no reason to try to encapsulate so much of WP:V when the intent is just to say that you can use primary sources. WP:V and WP:IRS are right there to look at for all the details. This sentence just needs to say a lot less: "Primary sources are not excluded from use on Wikipedia". Rhoark (talk) 01:26, 1 July 2015 (UTC)
- It couldn't have that effect, since WP:V explicitly permits such sources for a very limited purpose, at WP:ABOUTSELF. A WP:SPS isn't "reliably published" any more than it is "reputably published". Remember that this is the NOR policy, not the V policy; a policy on original research doesn't prevent, e.g., direct quotation or attributed paraphrasal of what a subject says about themselves, if we're not using it to make any controversial or WP:AEIS claims. Meanwhile, policy already prevents us using (in ways not permitted by WP:V) any self-published material, and this is extented to "publishing mills" and user-generated content (i.e., not reputably published). So, this literally is nothing but a copyedit to stop saying using "reliable" to define itself in the circular reasoning fallacy, and to consistently use "reputable" publisher not the meaningless "reliable" publisher, across all these related pages, so we get our terminology consistent. It's important because there really are editors convinced that a publisher is reliable, i.e. that anything released by a "reliable" publisher is automatically a reliable source, no matter what it is. I've encountered this view half a dozen times on WP in the last month. In reality, reputability of the publisher is only one factor in source reliability, but our wording error here is confusing editors into not grasping this. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 05:38, 16 August 2015 (UTC)

The "about primary sources" clause

I fully agree with the first of the two sentences SMC quotes [in the previous thread]:

'All interpretive claims, analyses, or synthetic claims about primary sources must be referenced to a secondary source [...]'

The phrase I put in bold is important. When talking about a primary source, I agree that we do need a secondary source to support what we say (although I would expand it to also allow summaries of analysis and evaluative claims contained in and cited to tertiary sources)

I am less sure I support the second sentence:

'Articles may make an analytic or evaluative claim only if that has been published by a reliable secondary source.'

This sentence could be interpreted as banning all mention of an analytic or evaluative claim that is contained in a non-secondary source. For example: if (for example) an old chronicle (such as that written by the Venerable Bede) includes an analytic or evaluative claim about the Saxons, someone might argue that we could not mention it because the chronicle is a primary source. I feel we should be able to mention it... we simply need to attribute it when we do. Blueboar (talk) 01:41, 30 May 2015 (UTC)

@Blueboar: :You didn't dispute my reasoning in the previous thrad, despite being the other party in the interpretational discussion I linked to, and have only responded to the "about primary sources" side point I brought up for later discussion. I'll take that as an indication the disagreement is resolved, unless you say otherwise.

In responding to your proposals, I'll number my responses here for easy reference. I apologize for the length, but this stuff is complicated to analyze, and it must be done very precisely.

Detailed analysis

Inserting "about primary sources" into the second sentence would be a major policy change, needing it's own proposal. The first sentence is from an interpretational, explanatory passage in the policy, but the second is the actual policy-defining sentence on this point, clearly labeled with an unmistakable "Policy" marker. (By contrast, normalizing the two passages to both use "referenced to" wording is not a policy change at all, since WP:V policy already requires citations; it's simply a copyedit.) The problem with narrowing the actual policy to only require secondary sourcing for an analytical/evaluative/interpretive/synthetic claim about what a primary source says, is that for WP purposes there is no difference whatsoever between an editor making such a claim in a article about what primary sources say, vs. doing that about what secondary or tertiary sources say.
Simple, real-world example: I see that one source says that a game called B.B.C. Company pool was popular in North America and the UK in early 20th century. I see another source say that modern eight-ball played throughout the world is similar. Whether either or both sources for those two sets of facts I combined in a novel way were primary, secondary, or tertiary, it's precisely the same original research for me to analyze/interpret/evaluate/synthesize this into "modern eight-ball is derived from B.B.C. Company pool" (even though the inference might be true, and it actually turns out to be). That synthetic connection can't come from a primary source, either: Some guy somewhere theorizing that the games may be linked, without any editorial overview and fact-checking by a reputable publisher to see if he did research or is just smoking dope and making stuff up, isn't a reliable source. If an ostensibly tertiary source makes this claim (e.g. Michael Ian Shamos's New Illustrated Encyclopedia of Billiards, which does so), then for that analytic claim it is in fact a secondary source, by definition. Primary sources make claims based on direct experience/experiment/opinion. Secondary sources analyze primary-source material and synthesize it, often making claims in the process based on vetting primary and often other secondary sources, and these secondary claims are subject to editorial review. Tertiary sources just compile and regurgitate previously published data, usually from secondary sources, without making any new claims. (An example of the tertiary source material in such a case would be the information that the two games have existed, where and when, and summaries of what their known rules are, in that same work by Shamos). If a tertiary source seems to be making a synthetic or original claim then it is, in that case, a secondary or primary source, respectively for that claim. Or the claim is not actually (or is no longer considered) synthetic or novel; e.g. as scientific and other ideas are tested and become widely accepted, they begin to be treated as facts, not hypotheses, by tertiary sources, including Wikipedia itself.
You parenthetically suggested changing the policy to allow a tertiary source for an analytic/synthetic/etc. claim, if all the tertiary source does is summarize the claim from secondary sources. That would also be a major policy change and a distinct proposal we'd need to think carefully about. I have considered it some, and I'd have to oppose it, because tertiary sources are less reliable in several ways. They are more prone to error, present a more superficial and less nuanced and detailed view (this is inherent in the very nature of summary), often contain obsolete information before they are even published, frequently blur distinctions between subtly different secondary-source claims, are often indiscriminate in what they include, while others are too exclusive (biased), may gather data from both primary and secondary sources directly, etc. Many of these are definite problems with book abstracts, for example. Or, put another way, we can't use a neutral, summarizing book review as a source for the book's contents. Some of the problems pertain more to database-like tertiary sources, though. As I detailed in our other discussion, the criteria by which we can evaluate tertiary and secondary sources are different. What makes a tertiary source well-respected is largely how good its producers are considered at compiling information well, not their subject-matter expertise and analytic research skills. But it's those second qualities, not the first, that help us evaluate the reliability of an analytic claim. A citation to a tertiary, summary version of the claim "insulates" it from WP's reliable secondary source analysis, even if the claim allegedly ultimately originated in secondary sources. If we try to do it, the amount of work required to figure out (if we can at all) what sources the tertiary one got this claim from, and whether they are secondary or primary, and whether they're reliable, is often going to be at least as much as what it would it take to find and cite a secondary source to begin with. Especially since, because obsolescence is a frequent issue with tertiary sources, we often need secondary sources anyway. Tertiary sources are best used for citations (often temporary, until secondary sources are found), for non-controversial, non-analytic facts. Anyway, over-reliance on tertiary sources is a major cause of low article quality on Wikipedia. It's a different kind of problem from inappropriate use of primary ones, but a serious one nonetheless. (The proximal cause of this is failure of WP:RS and WP:NOR to more clearly distinguish between secondary and tertiary source material, which I hope to address soon-ish.)
If a mostly-tertiary source by a subject-matter expert includes an analytic claim previously published in some secondary sources, the expert is acting as a secondary source here, using their expertise to concur with the analysis. In the pool example above, if Shamos had gotten the idea of the connection between the two pool games from some previously published sources, which is likely, he's still a secondary source on the topic, because he's using his expertise as a billiards historian to evaluate a hypothesis he's considering including in his encyclopedia, which is no different from evaluating hypotheses he found in primary sources. Just repeating recorded factual details about two games (the tertiary material) isn't that kind of evaluative process, but one of much more basic checking against his inclusion criteria. (This illustrates why most topical encyclopedias by subject-matter experts are both secondary and tertiary sources, for different information.)
The Bede example cannot legitimately arise. Pre-modern sources are all primary, and we do not cite them directly, but cite secondary sources that interpret them. Any modern, published edition of Bede with analytical commentary, which is probably all of them, will suffice. If you cited it directly, you would literally be citing as the source a medieval manuscript in the Bodelian Library, which you had not read yourself, and that is against WP:SAYWHEREYOUGOTIT. Bede's manuscript is also unpublished, so it fails WP:V. You still need to attribute it, because it's a primary source being quoted or paraphrased in a secondary one. Attribution and citation are completely separate concepts. E.g. if Joe Bloggs quotes primary researcher Jane Smith in his (secondary) review of astrophysics theories, it's wrong to quote the statement and attribute it to Bloggs; we attribute it to Smith, and cite it in Bloggs, just as you'd attribute Bede and cite a modern book publishing and analyzing his manuscript materials. So, yes, of course you can mention what Bede said about the Saxons, if it's encyclopedically relevant in the context.

Conclusions: A) Tertiary sources are not equated with secondary ones, including for analytic claims, for good reasons. B1) The addition of "about primary sources" to the interpreting sentence, absent from the policy-defining sentence, is an erroneous interpolation, and should be removed. It was probably added, in good faith, because when Wikipedians make this kind of synthetic error, it's most obvious (and probably most frequent) when they're working with primary sources. We certainly detect it more often, because we're more suspicious of primary source citations. B2) It would probably be reasonable to reword the interpretational sentence to say "especially about primary sources". I'd be in favor of that clarification, since it would interpret the policy correctly, while also still conveying the "use caution with primary sources" message. D) You can attribute Bede, where what he said is actually relevant, but not directly cite him (his work is manuscript material, not published). You have to cite a modern publication.

— SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 04:58, 30 May 2015 (UTC)

Perhaps we are concerned about different things... the key point I am trying to make is that if a source... any source (be it Primary, Secondary or Tertiary)... actually contains an analysis or evaluative claim, it is not Original Research for us to mention that analysis or evaluation in an article (for the simple reason that it is not original to us - we are taking it from the source). This is why I think the language of the second sentence you quote is flawed. It is actually not a WP:OR violation for an article to include an analysis or evaluation that has been made by a source (especially if you attribute it)... what we can't do per WP:OR is create and include our own analysis or evaluation based on what a source says. THAT is Original Research. However, this distinction between reporting on analysis or evaluative claim contained in a source, and creating our own analysis or evaluation based on sources holds true no matter what "type" of source we are using (creating our own analysis or evaluation based on secondary sources is just as much original research as creating one using primary or tertiary sources}. The issue isn't the nature of the source, but what we do with it. Blueboar (talk) 13:41, 30 May 2015 (UTC)

That simply isn't correct, though, and the policy says this twice in slightly differently-worded sentences. Many primary sources contain such claims, but we cannot use them as sources for such claims; in primary sources, they are hypotheses, unevaluated by other subject-matter experts in secondary sources. Yes, it is true that we can't create and include our analysis/evaluation/interpretation/synthesis based on what a source says; but this is not the only thing we cannot do. As a concrete example, many primary sources (e.g. opinional book reviews, and papers in academic literary and arts journals), have made the analytic claim that J.R.R. Tolkien's The Lord of the Rings is an allegory about England and its allies vs. Germany in the two World Wars. Reliable secondary sources pretty uniformly disagree with this assessment. If it were a brand new novel claim, unexamined in secondary sources, it would be WP:UNDUE weight to insert this untested, POV-pushing, questionably plausible interpretive claim into a WP article, and a form of original research (namely Wikipedians taking it on themselves to analyze whether this idea has merit as a literary hypothesis). Since it has been examined and been found lacking, in reliable secondary sources, and has been the source of plenty of reliably sourceable debate, it's perfectly fine to include mention of it as a hypothesis, and how well (actually poorly) that hypothesis has been received (in secondary sources). Doing so would be neither original research nor undue weight. [If you're curious, in actuality most secondary sources agree that Tolkien worked in a few allegorical elements, but that the work as a whole is not an allegory, unlike C.S. Lewis's contemporary Narnia series. The two authors frequently met to share and critique each others' new work, and Lewis's reliance on direct allegory was Tolkien's principle criticism of it.]

PS: Just because it's fun to use Tolkien for examples, here's an even simpler one. Tolkien himself (primary source) said that TLotR is not a trilogy, but a single work published in multiple volumes. Secondary sources nearly universally agree that this is true. There are many primary and tertiary sources that refer to it as a trilogy, but is wrong for WP to reanalyze on its own and agree with the primary and tertiary sources, and re-label it a trilogy.

If you think these examples do not get at the kind of scenario you have in mind, then please give us some examples that illustrate why you think the policy must be changed in such a major way. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 18:10, 31 May 2015 (UTC)

PS: Actually, the policy makes the point three times, since it also says "A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge." An analytic/evaluative/interpretive/synthetic claim in a primary source is not a "fact", but a hypothesis. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 19:02, 31 May 2015 (UTC)

Re: Many primary sources contain such claims, but we cannot use them as sources for such claims. Yes... we can... although we are limited in what we can say about them. If primary source X contains claim Y, we can state as a descriptive fact that primary source X contains claim Y (citing primary source X itself). That is NOT original research.

What the policy bans is "any interpretation of primary source material" without a secondary source... and I fully agree with that. But what I am talking about is not an interpretation of the primary source material... I am talking about a simple description of the source's content. Hell, if necessary you can quote the primary source directly. The same is true for any source (primary, secondary or tertiary). We may not be able to assert that "TLotR is not a trilogy" based on what Tolkien said... but we can mention the fact that he said it (and cite where he did so). Blueboar (talk) 19:30, 31 May 2015 (UTC)

As I said in semi-related discussion at WT:TERTIARYUSE, I think we keep talking past each other. Yes, of course we can state as a descriptive fact that primary source X contains claim Y; that's not what's at issue here. That is not using a primary source to cite any analytic/evaluative/interpretational/synthetic claim we're making in an article, it's quotation (or perhaps paraphrasal, depending on how it's written), with attribution. To get back to the gist, earlier you said: 'I am less sure I support the second sentence:', and then quoted it: 'Articles may make an analytic or evaluative claim only if that has been published by a reliable secondary source.'. You continued: 'This sentence could be interpreted as banning all mention of an analytic or evaluative claim that is contained in a non-secondary source.' (boldface added). But that's clearly not true, by your own "state as a descriptive fact that primary source X contains claim Y" example, which we both agree is a valid approach. So, I'm at a loss for what we could actually be disagreeing about in principle, or more to the point, where anything about the policy wording of that point should change. People clearly aren't interpreting it that way, since surely there are at least hundreds of thousands of statements and attributive citations of this sort in WP articles. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 02:07, 3 June 2015 (UTC)

This is a bit TLDR for me right now, but User:Blueboar, I think a distinction between making a claim and mentioning one addresses your problem. In your example, the Venerable Bede says something about the Saxons. The Wikipedia article could "mention" the Venerable Bede's claim that the Saxons are a bunch of loutish ruffians; it should not "make" that claim itself. A primary source is reliable for the fact that the source says something; it is not (necessarily) reliable for a claim that the alleged fact is true. WhatamIdoing (talk) 08:11, 20 June 2015 (UTC)

Blueboar wrote "actually contains an analysis or evaluative claim, it is not Original Research for us to mention that analysis or evaluation in an article..." SMcCandlish wrote "That simply isn't correct, though, and the policy says this twice in slightly differently-worded sentences. Many primary sources contain such claims, but we cannot use them as sources for such claims", but we can and do. To use an example that has been used since at least 2007 in the archives of this talk page: It is used for statements such as Wellington's comment on the Battle of Waterloo that it was "the nearest-run thing you ever saw in your life" -- he was after all was one of the two leading experts on the battle -- and despite the 200 years of debate (they say more has been published on the battle than there are days since the battle), no naysayer has come forward to contradict it.

There is more to the quote but it is not misleading to take just the one sentence. The first sentence tends to be omitted because it does not say much and it is confusing to modern generation "Its been a damned nice thing." and Wikipedia does not mention the third sentence in the Battle of Waterloo article because it is not so neutral: "By God! I don't think it would have done if I had not been there." and expresses a personal speculative opinion (although almost certainly true).

As to a matter of presentation, as Wellington's opinion is a POV, as WhatamIdoing suggests such opinions need in-text attribution although all such POVs ought carry in-text attribution whether the source is primary, secondary or tertiary. If the information was factual, for example the taking of the weather measurements at Greenwich at noon on the day, I do not think it would be necessary to quote it and use in-text attribution, eg "The weather in London at midday on 18 June 1815 was ... (citation)".

-- PBS (talk) 12:47, 11 July 2015 (UTC)

Thats not relevant in any way. WP (and whoever) is not advancing the claim that the Napoleonic Wars were a close call, and simply citing Wellington for this fact and moving on. They may quote or paraphrase Wellington and attribute this view to him, and use his statement to illustrate and bolster the point, but the close-call nature of the conflict is verifiable with in-depth research in many secondary sources. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 05:43, 16 August 2015 (UTC)

Primary source – example of scientific paper

In the section Primary, secondary and tertiary sources, an example of a primary source is given as, "a scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment". Is this referring to just the raw data from the experiment and its data reduction, or to the whole paper including the author's interpretation and conclusions regarding the data? --Bob K31416 (talk) 06:42, 20 July 2015 (UTC)

It appears that the example doesn't refer to the conclusions in the scientific paper because that would be of no purpose for this policy NOR, which prohibits using conclusions by editors, rather than conclusions by reliable sources. Perhaps this should be clarified in the example? --Bob K31416 (talk) 21:22, 23 July 2015 (UTC)

It needs to be clarified in more than just that example. Repeating a conclusion that is explicitly reached in a source (be that source primary, secondary or tertiary) is NEVER original research... because the conclusion did not originate with Wikipedia... it comes from a source.

Original research occurs when we take what a source says, and reach our own conclusions from that... conclusions NOT explicitly stated in the source. And, while I support our cautionary statement about using primary sources with care (because it is so common to see editors creating their own OR conclusions based on primary sources)... we should also note that it is just as wrong to create your own OR conclusion based on secondary and tertiary sources. The flaw is creating a conclusion, the type of source used to do this is actually irrelevant. Blueboar (talk) 23:32, 23 July 2015 (UTC)

The problem is that there is an unclear example of a primary source which needs clarifying. Here's an improvement which simply changes "outcome" to "data".

"a scientific paper documenting a new experiment conducted by the author is a primary source on the data of that experiment".

Blueboar, Would this change be acceptable to you?

--Bob K31416 (talk) 02:51, 24 July 2015 (UTC)

It includes the authors' interpretations and conclusions, of course. Some material in journal papers is secondary, sometimes, when it's presentation and analysis of previously published material. A typical paper might be structured like this:

Analysis of why it's primary

1. Abstract
2. Introduction - what the research goals are, methodology is, etc.
3. Background - what the prior research is (may be several pages, a tightly-focused literature review of the state of the field on the questions central to the paper)
4. All the details of the work (experiments, whatever)
5. Analysis of the results
6. Conclusion

All of that is primary, except #3. The question to ask is "Is this particular fact coming straight from the minds of the author(s) based on their own work, or is it from synthesis of previously published work, before it reaches the peer review committee?" The authors summarized their own work. They described their own goals and methodology, then all the details of their work. They did the analysis of their own work, and came to their own conclusions. That's all primary by definition; it's single-party material. Their review of the literature in #3 is secondary, being analysis/evaluation/interpretation/synthesis of prior sources, like a regular literature review in the same journal (though the quality level of the latter might be a little better), or an journalism piece (not op-ed!) in a major newspaper, or a history book from a reputable publisher, etc. In all those cases, the underlying facts are coming from someone external to the writer, and the writer is making contextual sense of them in relation to other facts from other external sources.

Another way of looking at it is: Each brain (or set of them - a research team counts as one) that a claim of fact passes through is a sanity check, and a secondary source has had at least two of them, that's why it's secondary. Because WP:RS requires reputably published sources, i.e. those that have been through an editorial review process (another set of brains) and isn't just self-published, a reliable secondary source requires at least three [sets of] brains saying it's not nonsense. A reliable primary source, which we can use "with caution" for something things, has had two (the authors and the peer review committee, or the op-ed writer and the newspaper's editorial board). An unreliable primary source, like a celebrity's Twitter post, has come straight from the original brain, and we usually can't use it except for WP:ABOUTSELF.

PS: When a journal or some other source is the one producing an abstract of something, not its authors, the abstract is a tertiary source; book reviews in science journals, and in some other publications, tend to be of this character, being summarizing rather than judging. Those in newspapers and such tend to be opinional, so are primary sources (the value judgements are coming straight from the mind of the reviewer).

Oppose, therefore, the proposed change. 'a scientific paper documenting a new experiment conducted by the author is a primary source on the data of that experiment' is terribly misleading and suggests, completely incorrectly, that the conclusions a paper's author draws about their own research is a secondary source. There is no difference at all, for WP:CORE purposes, between the data from the author's experiment, the description of the methodology used to gather it, the conclusions the drew from it, or even their abstract of the whole paper. It's all one author saying "what I tell you in this paper is true", and one level of sanity checking, the peer review committee, saying "this doesn't seem to be nonsense". Many short journal papers that do not include any secondary WP:AEIS material are primary from their first word to their last (other that the references list, which is tertiary). — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 06:43, 24 July 2015 (UTC)

SMcCandlish, Your comments seem to neglect the purposes of this policy No Original Research, which is to prohibit editors from using their own conclusions or information. It doesn't further the purposes of this policy to define the conclusions in the example scientific paper as a primary source. It furthers the purposes of this policy to define the data of the paper as a primary source because editors need to exercise caution that they don't make their own conclusions from the data. --Bob K31416 (talk) 08:01, 24 July 2015 (UTC)

@Bob K31416: It doesn't "un-further" the purposes either. It's just as much a case of OR (of a different kind) to take two unconnected studies that, in independent conclusions about different kind of research, suggest "cat hair causes cancer", and as a Wikipedian conclude that this has been proven, and repeat it as fact in WP's voice citing those two studies. This kind of synthesis is rampant in WP articles. It would be legitimate to just state this claim and cite a source, if it's a quality secondary one, like a literature review. But if it's just cited to the two original studies, not representing research that has been reproduced and accepted, and repeated in various secondary sources, it needs to be directly attributed: "According to a 2014 study ... In 2015, a new study ...".

Do we need a specific example of only taking raw data from a paper and coming to one's own conclusion about it? If so, how do we construct that example to not imply that the other primary parts of the paper aren't primary? They still are. The conclusions of the paper are a higher quality primary source that the raw data, but it doesn't magically transmutate into a secondary source. A secondary source is one that engages in analysis/evaluation/interpretation/synthesis of other primary and sometimes secondary sources, not of raw data. The text you proposed doesn't look like an example of editors doing OR by taking data from a table in a journal paper; it looks like a statement of policy that raw data, and by implication only raw data, in journal articles is primary sourcing, and by further implication that the rest of them are secondary source material. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 09:54, 24 July 2015 (UTC)

I'll just give you some feedback here that your message seems incoherent to me and leave it at that. FWIW, I think you're acting in good faith. --Bob K31416 (talk) 22:16, 24 July 2015 (UTC)

Unless other editors respond, it looks like this suggestion is going to be smothered by a wall of irrelevant information from one editor. --Bob K31416 (talk) 08:38, 24 July 2015 (UTC)

I collapse-boxed the bulk of it. I hope my followup makes it clear why it's not irrelevant. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 09:54, 24 July 2015 (UTC)

I entirely agree with you, Bob, as to the meaning of "primary" in the context of a scientific paper. The issue with scientific papers is not that they are "primary sources" in the historian's sense (the raw data is the nearest equivalent), but whether they are notable. A report in a single scientific paper is not a suitable source in an encyclopedia. However, if notability is established via other sources, whether multiple journal articles or secondary summaries, there's no reason not to use journal articles for specific information. Otherwise you have the absurd situation that it's ok to paraphrase an already paraphrased report of scientific research, but not to return to the original papers. In practice, if you look at science articles in Wikipedia, editors rightly make wide use of journal papers and will continue to do so. Peter coxhead (talk) 08:42, 24 July 2015 (UTC)

Please see my response just above to Bob K31416. I'm saying exactly what you are with "A report in a single scientific paper is not a suitable source in an encyclopedia". I'm not using the historians sense; I'm using Wikipedia's sense. In the historians sense, "primary", etc., tend to be labels for entire classes of works, not of specific content (e.g. all news media is "primary" to many historians, even the kind we'd call secondary). I agree with you, more or less, on the notability point, except we should probably call it reputability. (I'd be cautious of applying the term "notability" because it's satisfied, on WP, with coverage regardless what the reason for the coverage is; e.g., a fraudulent paper that generates controversy may be very notable for the controversy, as its reputability falls through the floor). A WP problem is that the average editor doesn't know how personally reputable any particular researcher is. Professional in science are looking for journal articles; that's where the meat is. But WP isn't; it's looking for "has this hypothesis tricked down into the science book publishing market, or at least appeared in a literature review?" I certainly agree that it's okay to use reputable journal articles and I hold that the "use with caution" rule about primary sources is more and more clearly satisfied the more reputable the authors, the more reputable the journal, and the more frequently the work has been cited in professional literature.

I want to be clear that I'm a defender of using journal papers in articles here, and I cite them all the time, just not using Wikipedia's voice to present their claims; attribution is important. There are long-term, hard-core Wikipedians who believe wrongly that nothing in a journal paper (as distinct from a literature review) can ever possibly be secondary source material. If you still think my interpretation is overly strict, you should hear those folks. The workaround for a case where one of them says you can't use a paper for the claim at hand, and wants to revert you, would be to cite the literature review or whatever, then add the citation to the real paper along with it. If there's no lit. rev. or other secondary source, they might make WP:UNDUE claims, and other might support them, especially if it's new research. But usually carefully attributing the claim mollies people. Anyway, maybe the concerns I've been raising on this thread weren't resonating for you because of the article space you've been mostly working in or something, and you're not seeing people synth'ing up PoV nonsense from unconnected research. One case I remember was a "drinking alcohol is actually good for you" thing I could give journal paper source-abuse details about, but it would add another couple of sentences. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 09:54, 24 July 2015 (UTC)

If we disagree (and we may indeed not), then it may be about the value of using the labels "primary", "secondary" and "tertiary" in the context of scientific articles. As far as I can see, they have become harmful to Wikipedia, because they are constantly misused. What matters is the kind of material a source is being used to support, and why. If I want to support an explanation of why the author of a taxon name gave it that name, the most reliable and reputable source is the original paper. As I know to my cost, authors of compendia of scientific names and their origins and meanings copy endlessly from one another and errors get propagated. If I want to know what order the APG III system puts a particular family in, the most reliable and reputable source is the original APG III paper. But in each case wider sources are required to construct the articles in which the piece of information appears. A taxon name has to have some solid degree of acceptance in the biological community, appearing in textbooks or major biological databases. A system like APG III has to be shown to be widely used by botanists before being used in Wikipedia, and we need to know whether later revisions have been accepted. In scientific articles, routine use of "primary" = bad, "secondary" and "tertiary" = good is just plain wrong. Peter coxhead (talk) 10:13, 24 July 2015 (UTC)

I do agree that use of the terms has been confusing here, very much so, because the hard sciences, social sciences, humanities, law, history, etc., uses of these terms are all different. I do believe this is fixable by having a sentence at WP:PSTS saying as much, and then explaining that a) WP has its own definitions of these terms; b) they do not vary by subject area; c) they are descriptive of particular content in a particular context, not of entire classes of publications: the kind of material a source is being used to support, and why is key; and d) they are: [definitions here, 1, 2, 3]. In the course of explaining how to apply "the kind of material a source is being used to support, and why", we'll have to categorize, and we already have these categories, so why not just fix our terribly-worded approach to the matter? If after a year or so (it would take that long for the clarification to propagate at least, see if it's sufficient, and debates with newly arrived people go something like "I know you want to prevent this citation as a 'primary source' in historians' terms, but this is a Wikipedia secondary source", or "no, this is not a secondary source under Wikipedia's definition, it's tertiary (a category that doesn't exist in your field); as such you can't use it for this kind of claim". If there was no improvement, then see about introducing new terminology. We already have so much jargon, I think that's a last resort. People from editorial tradition A tend to be smart enough to figure out what people in tradition B mean when it's explained to them, and to understand that there are multiple traditions N. They shouldn't have any trouble grokking a tradition W. They haven't been because we're not writing it clearly. I have no opinion on APG III, and am only addressing the issue in the broadest cross-WP sense. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 23:48, 25 July 2015 (UTC)

WP:PSTS has always been the "tail that wags the dog" in this policy. Whether a source is Primary or Secondary or Tertiary is often irrelevant to the issue of whether the source is being misused to support Original Research. All three types of sources can be used correctly, and all three can be abused (used incorrectly) in ways that constitute Original Research.

What is frustrating to me is that we spend so much time and energy trying to define what constitutes a Primary, Secondary or Tertiary source, that we keep keep loosing our focus on the point of this policy... defining what constitutes Original Research. That's what this policy needs to focus on. Blueboar (talk) 12:07, 24 July 2015 (UTC)

Actually I think that what constitutes OR is well understood and, in my experience, almost never an issue – those who add OR to an article generally know perfectly well what they are doing, they just feel it's justified, usually because what they've added is "true", not because it's not OR. But I absolutely agree that the time and energy spent on trying to define types of source is wasteful and frustrating. Peter coxhead (talk) 12:45, 24 July 2015 (UTC)

Yes, we all already get that point. No one in this discussion doesn't know that making up your own theories and posting them on WP is verboten no matter what sources you use and what their classification, and few non-noobs seem to fail to understand it; it's mostly problematic PoV-warriors who are trying to do that. That does not solve the problem I, among many others, are seeing all the way across WP:CORE, and issue that exists independently of Blueboar's concern. It is the idea, directly forbidden by this policy but very badly confused by many (see thread above and recent alteration to the policy that has to be reverted) that publication of primary source material (as defined by WP) in a mostly secondary (ditto) publication magically makes the primary source material into secondary and thus an acceptable place to get WP:AEIS material from. Blueboar personally doesn't care about the PSTS distinction. We get that. Others do, because misconceptualizing it is directly leading to WP:GAMING and serious source abuse. That is, there are [at least] two different OR cases to cover with regard to primary sources, and its a serious mistake to cover why the one is wrong in wording that directly encourages more of the other. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 23:48, 25 July 2015 (UTC)

What is and is not OR

To illustrate why I think it is a distraction to define Primary, Secondary and Tertiary sources in this policy... let's examine what constitutes Original Research in each case:

Primary sources
- Original Research - Analyzing, evaluating, interpreting, or synthesizing material found in the source (or multiple sources) yourself to state (or imply) a conclusion not directly stated in the source.
- Not Original Research - Reporting on the analysis, evaluations, interpretations and conclusions directly stated within the source. (However, this should always be done with in text attribution to the source)
Secondary sources
- Original Research - Analyzing, evaluating, interpreting, or synthesizing material found in the source (or multiple sources) yourself to state (or imply) a conclusion not directly stated in the source.
- Not Original Research - Reporting on the analysis, evaluations, interpretations and conclusions directly stated within the source. (Best done with attribution, but conclusions can be stated in Wikipedia's voice if multiple sources conduct the the same or very similar analysis, evaluations, and interpretations to reach the same conclusions)
Tertiary sources
- Original Research - Analyzing, evaluating, interpreting, or synthesizing material found in the source (or multiple sources) yourself to state (or imply) a conclusion not directly stated in the source.
- Not Original Research - Reporting on the analysis, evaluations, interpretations and conclusions directly stated within the source (Any conclusions can usually be stated in Wikipedia's voice).

Have I left anything out? If not, then definition of what constitutes Original Research and not is exactly the same for all three "types" of sources. The only difference is whether the conclusions being reported are stated in Wikipedia's voice (as fact) or reported with attribution (as opinion). But that is not really an OR issue ... it's more a NPOV / DUE WEIGHT issue.
Now for the tricky part... I think we would all agree that it is far more likely that people will misuse primary sources in ways which constitute OR (whether they are doing so intentionally or unintentionally)... so some sort of cautionary warning is appropriate. The question is whether we need to define what constitutes a primary source in order to give that cautionary warning. I'm not sure that we do. Especially if doing so distracts the reader from understanding what constitutes Original Research (as I think our current definitions do). Blueboar (talk) 13:41, 24 July 2015 (UTC)

I agree, but also think there's no chance of a consensus for this view. It would be nice to be proved wrong! Peter coxhead (talk) 14:57, 24 July 2015 (UTC)

I agree with both of you. Sadly also that implementation will be difficult. Problem is that it is much easier to use /proscribe a black or white heuristic (primary source = bad, secondary source = good) than to weigh the sources on their own value; in the relevant context. Arnoutf (talk) 15:56, 24 July 2015 (UTC)

The implementation difficulty will be in fixing WP:PSTS at WP:RS, which is where it needs to be done. I agree that WP:NOR does not need to define "primary source"; etc. This material could be merged to PSTS where it belongs, and then NOR just address these (pointing to PSTS for definitions). That would go a long way to fixing the problems with both pages, though the exact wording in the extant definitions will need some help at some point.

I strongly agree with Blueboar that "it is far more likely that people will misuse primary sources in ways which constitute OR (whether they are doing so intentionally or unintentionally)... so some sort of cautionary warning is appropriate." What's triggered my concern in this thread is the flip-side of that: We also should not construct any language that implies that the only way to abuse primary sources for OR is with raw data! Some of the reasoning in the top subsection of this thread is faulty. This statement: "Repeating a conclusion that is explicitly reached in a source (be that source primary, secondary or tertiary) is NEVER original research... because the conclusion did not originate with Wikipedia... it comes from a source." is not correct. One of the most frequent NOR noticeboard dispute types is repeating (and properly citing and attributing) a claim from a paper that is unrelated to the material cited before and after it, but which, with that juxtaposition, pointedly leads the reader to a conclusion the editor believes is true but for which there is no source whatsoever. I think that's the clearest way to phrase what I've been getting at all along.

The obvious way out of this is: A scientific paper documenting a new experiment conducted by the author is a primary source on the data and outcome of that experiment. That will address both forms of OR we're talking about there. Then let's focus on getting these P/S/T definition stuff into the right page. Then on making it make sense to everyone, but clarifying that these are WP's definition, and we are not importing definitions from history, law, or any other field, much less randomly applying different definitions on-the-fly depending on what field someone things can claim scope over the topic (this view is more common than most of us thought, but it's utterly unworkable). — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 00:34, 26 July 2015 (UTC)

SMC I think we actually agree. Note the key words I put in bold... I stated that: "Repeating a conclusion that is explicitly reached in a source (be that source primary, secondary or tertiary) is NEVER original research".

You then state... "repeating (and properly citing and attributing) a claim from a paper that is unrelated to the material cited before and after it, but which, with that juxtaposition, pointedly leads the reader to a conclusion ...that juxtaposition, pointedly leads the reader to a conclusion the editor believes is true but for which there is no source whatsoever."

My response is... our two statements are not in disagreement... the synthetic conclusion that the juxtaposition points the reader to is Original Research because it is NOT explicitly reached in a source. The OR is not in repeating the claim... The OR consists of repeating the claim in conjunction with other material so that it reaches an implied conclusion that is not explicitly reached in a source. Blueboar (talk) 14:09, 26 July 2015 (UTC)

To put this another way... if we have Conclusion A (explicitly stated by a source), and "Conclusion B (explicitly stated by a source) and Conclusion C (not explicitly stated by a source)... it is not Original Research to repeat Conclusion A, nor is it OR to repeat Conclusion B... what is Original Research to state (or imply) Conclusion C. Why? Because C is not stated in any source. State A or B separately (ie not in conjunction) and they are fine as far as WP:NOR is concerned. Of course, one or the other will often become irrelevant or tivial when you separate them, but that is not an OR issue.

by the way... One very good way to highlight and explain synthesis to someone who does not get it is to simply switch A and B around. This often completely changes the synthetic Conclusion... this can help someone understand that A and B were being used improperly. - A similar test can be used to highlight POV pushing... if an article states A + B, and switching them to B + A changes the POV... you know that the article is not being neutral. Blueboar (talk) 15:44, 26 July 2015 (UTC)

Good "trick" at the end there. I'll remember that one. I can think of a specific place to use it already. Anyway, I agree that our views on what is OR are not in conflict. But many editors have great difficulty with this stuff. All I want is ... is a primary source on the data and outcome of that experiment so it doesn't mislead people into thinking that only the data in the paper is a primary source. The conclusion is, too, so it can be used "with caution" and directly attributed; but we mustn't imply that it can be used without attribution for the claims the conclusion makes, because it is not actually a secondary source. We don't need to explain that it can't, here, because this is the NOR page, not the V or RS pages. It's simply intolerable to have NOR appear to contradict RS. You and I might know that it it wouldn't really conflict, if you know your policy really well, but all the evidence before our eyes across Wikipedia is that it would be widely misinterpreted/gamed. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 02:18, 27 July 2015 (UTC)

From the beginning of WP:NOR

"Wikipedia articles must not contain original research. The phrase "original research" (OR) is used on Wikipedia to refer to material—such as facts, allegations, and ideas—for which no reliable, published sources exist.[1] This includes any analysis or synthesis of published material that serves to reach or imply a conclusion not stated by the sources."

If a conclusion appears in a reliable source that also has new data from an experiment, that conclusion is not original research according to this policy. Furthermore, I see no reason why that conclusion is any more susceptible to being used for synth than any conclusion in any other reliable source. --Bob K31416 (talk) 14:30, 27 July 2015 (UTC)

That's not the point at issue. And it's more susceptible because it's a primary source, presenting something brand new to the world (at least in part). WP trusts reputably published secondary sources because they are not doing this, but are the output of a multi-layered editorial "brain filtering" process that tells us that the real world is giving this claim some credence. I covered in this in detail in material that you denigrated as "irrelevant". Maybe you should have read it instead. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 21:21, 27 July 2015 (UTC)

Hmmm... could it be that we are not quite in sync as to what the actual issue under discussion is? Bob's comment certainly goes directly to the issue that I am concerned about. Blueboar (talk) 21:34, 27 July 2015 (UTC)

Retry: It's not the point at issue because the segment in question isn't defining what is or isn't OR, but what is or isn't primary (in relation to OR, theoretically, but worded as a broadly defining statement that will be WP:GAMEd as one). A primary-sourced conclusion is more susceptible to OR than one found in secondary sources (the kind from which we can repeat a WP:AEIS claim) because it's a primary source, presenting something brand new to the world (at least in part). WP trusts reputably published secondary sources because they are not doing this, but are the output of a multi-layered editorial "brain filtering" process that tells us that the real world is giving this claim some credence. I covered in this in detail in material that you denigrated as "irrelevant". Primary sources are highly prone to contextualizing/interpretive original research, especially via assumptions of acceptance of claims/conclusions (i.e. of veracity), by misinterpretation of precisely worded conclusions during recasting in a Wikipedian's own words (a don't-plagiarize process we're supposed to engage in but which is more "dangerous" when working with primary sources), correlation–causation errors, mistaking topically similar conclusions as correlative at all, etc., etc., etc.

Let's turn this around: Why is it seemingly very important to you to prevent WP:NOR from saying that anything but the raw data in a journal paper is primary sourcing, when there's no consensus that it is anything but? No one objects to adding the specificity that the data is primary, in addition to the conclusion (I agree wholeheartedly!). No one's even saying everything in a paper is primary; many papers contain a large amount of secondary material, where they engage in small-scale literature reviews on a particular matter; I also already covered this. [Actually I shouldn't say "no one" here; I know of two editors who should know better who keep trying to tell people that this secondary material is primary.] If this passage were leading directly into an example of abuse of raw data to engage in one kind of OR, it would be understandable to limit that wording to "data" (and then follow it up with a "conclusion" version and an illustration of the other kind of OR I'm talking about). But there are no such examples. And it's been proposed here to relocate the bulk of this material to WP:PSTS, anyway, which might just moot the matter entirely. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 21:47, 27 July 2015 (UTC)

1) It seems like the issue you are concerned with doesn't have to do with original research but rather with the reliability of sources. Am I understanding you correctly?

2) The definitions of the three types of sources are introduced with the following statement at the end of the second paragraph of WP:PSTS],

"For the purposes of this policy, primary, secondary and tertiary sources are defined as follows:[2]"

--Bob K31416 (talk) 23:39, 27 July 2015 (UTC)

1) Nope. 2) So? Does not address the issue I raised: The change you want to make is directly misleading. There is no rationale for removing the one word and substituting the other instead of just adding the word you want to be covered there. I'm highly skeptical that re-re-re-re-re-re-re-stating this another dozen different ways is going to be productive. Please see WP:NOTGETTINGIT. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 00:22, 28 July 2015 (UTC)

1)The reason I thought your issue had to do with reliability was your comment, "WP trusts reputably published secondary sources because they are not doing this, but are the output of a multi-layered editorial 'brain filtering' process that tells us that the real world is giving this claim some credence."

2) Regarding your comment "A primary-sourced conclusion is more susceptible to OR than one found in secondary sources (the kind from which we can repeat a WP:AEIS claim) because it's a primary source, presenting something brand new to the world (at least in part)." – When you say "primary-sourced conclusion", are you referring to a conclusion made by a Wikipedia editor, or a conclusion that appears in a reliable source that a Wikipedia editor is using in an article without going beyond that conclusion? --Bob K31416 (talk) 00:28, 28 July 2015 (UTC)

1) Ok ... I don't see where the inference comes from, though. 2) In that particular construction, I'm referring to the conclusion reached by the authors of a journal paper, which NOR defined unquestionably as a primary source (but we have used "conclusion" in some other ways in this discussion). Whether it's reliable enough to cite (WP:RS) at all (WP:V) doesn't relate strongly to how we can cite it, for what (WP:NOR at WP:PSTS).

I'll re-ask my above question, in different terms, since staying "on-track" seems to be a goal (and I agree with it): Why do you want WP:NOR to state that data specifically, and by implication only data, in a journal paper is primary source material? How could this be jusified, when the policy clearly says otherwise, and we know for certain that many editors are already confused, despite the clarity of the policy's definitions, into thinking that a paper's conclusion somehow isn't primary [for WP purposes]? Why make this confusion worse, with a substitution of "data" for "conclusion", instead of helping clarify by adding "data" to "conclusion"? We even know that the confusion comes from differing external definitions of "primary". What end could possibly be served by perpetuating the incorrect belief that WP uses external, randomly changing field-by-field, definitions of "primary", when the policy concretely defines what that term means for WP editing purposes? — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 00:57, 28 July 2015 (UTC)

But Bob's contention (correct me if I have this wrong, Bob) is that the conclusion shouldn't be defined as Primary. That this definition is inaccurate. Bob's proposal is to change the definition (so it no longer defines the conclusion as Primary... only the Data). SMC, Your argument above seems circular... you seem to be saying: you can't change the definition contained in the policy because the policy contains that definition. I suspect (and again correct me if I have this wrong) that the real reason for your objection is that you disagree with Bob and think the conclusion is primary. To be honest, I don't know which of you is correct... so... to help me make up my mind, let me ask you both to lay out why you take the stance you do. Why is a conclusion primary?... or ... why isn't it primary? Blueboar (talk) 01:31, 28 July 2015 (UTC)

You're understanding me correctly that the conclusion shouldn't be defined as primary. My reason is that the definition is for purposes of this policy No original research, as stated in the introduction to the definition, and that using any conclusion from any reliable source, without going beyond that conclusion, is not original research. And there's no reason to believe that new conclusions in a scientific paper are any more prone to being used for synth than are any conclusions in any other reliable source. --Bob K31416 (talk) 12:20, 28 July 2015 (UTC)

OK... you seem to be saying that if Primary = OR then the flip side must also be true: Not-OR = not-Primary. My reaction is that your argument is flawed from the beginning, because Primary doesn't = OR. A conclusion can be Primary without any OR involved. And OR can occur without the sources being primary.

That said... what I really wanted to know is: are conclusions in scientific papers considered primary or secondary in the real world (outside of Wikipedia)? Blueboar (talk) 12:59, 28 July 2015 (UTC)

I didn't say that primary=OR, or mean to imply that. But that's OK, sometimes editors are misunderstood. I have the same position as this policy regarding primary sources, i.e. "primary sources that have been reputably published may be used in Wikipedia; but only with care, because it is easy to misuse them". For example, the data in an experiment is prone to different interpretations and Wikipedia editors might offer their own conclusions about the data. Whereas the conclusions in a paper about an experiment, aren't any more prone to Wikipedia editors using them to make their own conclusions than are the conclusions in any other paper.

Re "are conclusions in scientific papers considered primary or secondary in the real world" – It seems to vary. Also note that the definition of primary source outside of Wikipedia is not for the purpose of discouraging the use of primary sources for original research. --Bob K31416 (talk) 14:24, 28 July 2015 (UTC)

(edit conflict)I can't find any evidence they're considered anything but primary (in the sciences, anyway; I forget how humanities approaches this question). But science places high value on primary sources and low on secondary ones, because their needs are radically different. They're looking for the steaming fresh ideas that need testing and integration; the secondary sources are self-education material to catch up on and maybe argue about (literature reviews etc.) or popularization stuff for the masses (the monographs published by Oxford U. Pr., etc.). WP is the opposite of this. We don't want bleeding edge material that may be shortly disproven and the interpretation of which is uncertain, while we do want secondary sources to show us that other people (subject to editorial processes that reality-check them) have considered what the primary source research behind the claims(s) are saying, and done some WP:AEIS on them to fit them into the big picture (we're forbidden from this OR ourselves).

That said, the "here's the previously-published research analysis we're working from" background material in many otherwise (i.e., as to its data and conclusions) primary papers in journal is secondary material and tens of thousands, at least, of our articles cite it this way. When it's "According to Jackson (2012), the stamen of F. blueboaricus averages 0.79 times the length of that of F. smccandlishi" ..." that's secondary material, just a like journalist writing "According to Commerce Department figures, textile imports from Turkey last year...". The conclusion, like the data, in a paper is a primary source because it's coming straight from the mouths of the advocates of its veracity. The peer review committee (the editorial process at a journal) doesn't tell the authors what data the authors collected or what conclusions the authors should draw. Literature reviews are generally secondary because all they're doing is analysis of previously published work (though I've seen a few inject brand new ideas here and there, clearly labeled as the idea of the author of the review).

The high value placed on both primary research sources, and on top-level policy statements and press releases of science organizations (which are primary as to the position taken and tertiary as to the data) is confusing a lot of people into equating "high value" (i.e. reputable authors) with "secondary"; see WT:MEDRS for a big thread illustrating this cognitive dissonance. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 06:19, 16 August 2015 (UTC)

Quick aside... time to give some history. If you look back to the development of the PSTS section... At one point, the policy included a statement that said (essentially): Wikipedia is a Tertiary source... adding Original Research turns it into a Primary source, so don't add OR. At the time, it was felt that it would be helpful to explain the terms used in that statement. The original intent was not to discourage the use of primary sources, or to say "using primary sources causes OR"... but to explain that adding Original Research turns Wikipedia into a primary source. I know that's quite different from what we say now ... a lot of instruction creep has occurred since those early days... I just wanted to remind everyone why we included the definitions in the first place. Blueboar (talk) 15:53, 28 July 2015 (UTC)

Yet the policy has evolved to do so, and it's served use very, very well, vastly improving the quality of our articles. WP:AEIS in particular has thwarted an almost unimaginable amount of WP:SYNTH. Primary sources we're already discouraged at WP:V, but this policy corrected other abuses of primary sources that WP:V did not, especially those from reputable publishers that were unfiltered personal opinion/reporting/claims. It's important because journals will publishing things that seems plausible based on peer review and are thought provoking, but liable to be controversial and often proven wrong in short order. Newspapers publish editorials often for no reason other than "from someone famous and likely to arouse controversy and sell newspapers". And so on. WP:AEIS policy severs the mistaken mental link that leads from "this publisher is generally reputable" into the fallacy of "therefore this must be a reliable source just because of who the publisher was". It's absolutely crucial. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 06:19, 16 August 2015 (UTC)

Status of proposed change

The proposed change in the subject sentence of Primary, secondary and tertiary sources was to change "outcome" to "data", so that the sentence becomes,

"a scientific paper documenting a new experiment conducted by the author is a primary source on the data of that experiment".

So far the responses suggest,

Support from Bob K31416, ~~Blueboar~~, and Peter coxhead.

Oppose from SMcCandlish.

If an editor's intentions have been incorrectly interpreted in the above, that editor is invited to give their correct position. --Bob K31416 (talk) 22:32, 25 July 2015 (UTC)

P.S. The other editor in the above discussion, Arnoutf, and any other editors are invited to give their positions on the proposal. --Bob K31416 (talk) 23:15, 25 July 2015 (UTC)

Weak support – Only weak support. The data is definitely primary, but also the discussion section often contains primary arguments of the authors (especially in social sciences. Making that explicit would create a lot of confusion so I think the proposal is an improvement overall). Arnoutf (talk) 10:04, 26 July 2015 (UTC)
- @Arnoutf: I'm curious why you think the conclusion reached by a paper in a journal is not also primary, when the WP:PSTS clearly says that it is, and WP:RS#Scholarship reiterates this in summary form. As discussed elsewhere in here (cf. comments by Peter Coxhead), there's obviously been perennial confusion going on between different external definitions of primary vs. secondary source in different fields, but WP has its own definitions, and they're quite detailed. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 00:37, 28 July 2015 (UTC)

- - If your read my rationale for weak support you could see I fully agree that conclusions and discussions of scientific papers are primary insofar that they reflect on the data collected by the same author. If we do, however, rigorously apply this all empirical papers are primary (and so would be all journalism based on interviews), and to take it even more into the extreme, even the conclusion/discussion of review papers would be primary as that would give the ideas of the authors. So to take this argument into the extreme would nullify the difference between primary and secondary sources. The world is not a simple black and white. Arnoutf (talk) 08:32, 28 July 2015 (UTC)

I have crossed myself off of the list of those supporting. I would say I am still undecided. Blueboar (talk) 14:15, 26 July 2015 (UTC)

"Vote counting" by involved party in mid-discussion. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 00:05, 26 July 2015 (UTC)

Support from Bob K31416, ~~Blueboar~~, and Peter coxhead.

Oppose from SMcCandlish.

"Status of proposed change" is "under discussion still". Wikipedia is not a vote and it's inappropriate for an involved party in the discussion to try to "assess the consensus" in a steering way, when the discussion is still on-going. I also object to your special pleading above that amounts to 'help! my pet proposal is being derailed by someone whose objection I don't want to address; please come help me shut him up'. WP:Consensus is formed by actually taking the time to understand what people's objections are and try to resolve them; not by trying to make the objector go away, or drum up an entourage to drown out the objection. Looking above at the discussion, the concerns I've raised have yet to be addressed. Some participants want to make one particular kind of point about OR and primary sources, while I am pointing out that the proposed wording implies that a completely different kind of primary-source-based OR is not OR. Some appear to understand my point. Everyone agrees that WP:PSTS has wording problems (though this is naturally not the place to try to fix them). No one is objecting to the idea of the original proposal (using raw data from a paper to draw your own conclusions is OR), only the wording. So WP:DONTPANIC. The discussion is ongoing. We don't need to count anything yet, and no consensus is going to be reached by refusing to address concerns of some of the participants. You can't "WP:WIN" that way. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 00:05, 26 July 2015 (UTC)

Out of order. This section is out of order because the alleged proposal for a change in wording described at the beginning of the section either does not exist at all, or cannot readily be found. The concerns about an involved editor trying to summarize an unfinished discussion also apply. Jc3s5h (talk) 15:24, 27 July 2015 (UTC)

Here is the the full paragraph of WP:PSTS, where I have underlined the subject phrase.

Primary sources are original materials that are close to an event, and are often accounts written by people who are directly involved. They offer an insider's view of an event, a period of history, a work of art, a political decision, and so on. Primary sources may or may not be independent or third-party sources. An account of a traffic accident written by a witness is a primary source of information about the accident; similarly, a scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment. Historical documents such as diaries are primary sources.

The proposal is to change "outcome" to "data".

Regarding giving the status, there had been a lot of discussion digressing from the proposal and this was a way of getting it back on track. Also, it helped clarify editors' positions by requesting any corrections. For example, Blueboar clarified being undecided instead of supportive. --Bob K31416 (talk) 18:31, 27 July 2015 (UTC)

The title if this subsection, "Status of proposed change", implies there is already a proposal, and this section is giving the status of it. But in fact, the proposal is introduced in this subsection. This method of introducing the proposal fails to give interested editors reasonable notice that the discussion is occurring. Thus this discussion is invalid and cannot create a consensus to edit the policy. Jc3s5h (talk) 19:20, 27 July 2015 (UTC)

Jc3s5h... for the record... the proposal was actually introduced in Bob's comment at 02:51, 24 July 2015 (above ... it's the 4th paragraph of the original thread - Wikipedia talk:No original research#Primary source – example of scientific paper). OK... perhaps Bob could have marked it better, so others realized that he actually had made a proposal (But then again, do we really need to mark every proposal with the word "PROPOSAL" in flashing neon so no one can miss it?)... but my point is that it wasn't introduced in this sub-section. I don't think Bob was trying to improperly influence the outcome of the discussion by reporting on the "status" of that original discussion... I think he was trying to get us back on track... reminding us what his original proposal actually was, after we had gotten somewhat side-tracked away from it. Blueboar (talk) 20:37, 27 July 2015 (UTC)

I don't think it is fair to claim consensus for any change if the proposal is introduced after the discussion has lead editors to think the thread will not be of interest to them, causing the uninterested editors to ignore the thread in their watchlist. Jc3s5h (talk) 21:47, 27 July 2015 (UTC)

And "getting us back on track" doesn't require misleading vote-counting; even half those whose votes he counted disagree with his reckoning, anyway. We're also not "off-track"; at least one person in this overall discussion does not appear to have understood the quite central point I'm making, despite it having been explained in clear and detailed terms repeatedly. The deletion of "conclusion" and replacement with "data", instead of addition of "data", is misleading, would be WP:GAMEd without doubt, and is based on a misapprehension of what "primary source" means on (and for) WP and for this policy. (This misapprehension is reflected in the thread immediately above this overall one; I haven't even addressed that yet; it turns the meaning of that policy on it's head). I don't think there's a bad-faith attempt to improperly influence the outcome; It seems to just be heel-digging in response to the proposal being challenged in terms less easy to dismiss than they seemed to him at first. I have little direct interaction with the proponent and have no reason to assume bad faith about him. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 22:04, 27 July 2015 (UTC)

I think my previous comments about why I gave the status should be sufficient and I think the comments in the last two messages about giving the status are unreasonable criticism and divert the discussion from the substance of the proposal. --Bob K31416 (talk) 00:01, 28 July 2015 (UTC)

But this entire subsection, which you created, is a diversion of the discussion from the substance of the proposal. And I'm defending you on WP:AGF grounds in the wake of an uninvolved editors' criticism, even if also critical about the idea (expressed by someone else) that the discussion is "off-track", and of your judgement in launching the diversion this way to begin with. Given the immediately previous exchange (in edit history, not this subthread), it's clearly not off-track, because the central objection to the substitution-instead-of-addition edit you want to make has not been addressed. No amount of circular argument gets around that. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 00:29, 28 July 2015 (UTC)

Suggestion: Since we are already back on track in the thread above (ironically done by going off track in my "what is and is not OR" sub-thread)... may I suggest that we simply halt this sub-thread, and continue the "on track" discussion. Let's just pretend that this sub-thread never happened, and concentrate on trying to reach consensus. Blueboar (talk) 01:21, 28 July 2015 (UTC)

Comment - Having read through the discussion above, and thinking over the text a scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment; is there any reason why we should need to specify which sections of the experiment are considered primary?

Would it not be simpler to simply remove the underlined section; giving a scientific paper documenting a new experiment conducted by the author is a primary source on that experiment? NB: Please feel free to move this comment to the appropriate place in the various threads above. I could not work out where was best. - Ryk72 ^'c.s.n.s.' 11:05, 28 July 2015 (UTC)

First off, thanks for joining this discussion. (If anyone moves Ryk72's message, please move my response along with it and keep them together.)

In an experiment, data is taken. Then a paper is written about the experiment and the paper includes the data from the experiment and conclusions about the data. The point of contention is whether or not for the purposes of this policy of no original research, to include the conclusions of the paper as a primary source. Did you mean for the term "experiment" to include or exclude the conclusions of the paper? --Bob K31416 (talk) 12:02, 28 July 2015 (UTC)

Hi Bob K31416, My inclination is that the paper, as a whole & its component parts, are primary; but I am keen to see the answer to Blueboar's question, below - perhaps it is not the categorisation of primary or secondary that is the part of policy that wants changing. - Ryk72 ^'c.s.n.s.' 13:57, 28 July 2015 (UTC)

Bob... does it matter whether the conclusions are considered primary or secondary? If so, why does it matter? Blueboar (talk) 13:30, 28 July 2015 (UTC)

It matters because Wikipedia discourages the use of primary sources. So material using a source that is labeled primary is more prone to being deleted. --Bob K31416 (talk) 14:30, 28 July 2015 (UTC)

It seems that I've said all I have to say on the subject in this long discussion, so I'll leave now, still supporting the proposed change from "outcome" to "data" after reading all the arguments against it. --Bob K31416 (talk) 15:02, 28 July 2015 (UTC)

It sounds like your underlying concern isn't so much with the definition... but with the discouragement and inappropriate deletion. Is this an accurate assessment? Blueboar (talk) 15:11, 28 July 2015 (UTC)

MEDRS and primary/secondary

FYI

– Pointer to relevant discussion elsewhere

There's a fundamental issue at WP:MEDRS of confusing the concepts of "low- vs. high-quality source" (reputation of who is making the statement) with "primary vs. secondary" source (is there editorial control independent of who is making the statement?). I'm trying to clarify this at Wikipedia talk:Identifying reliable sources (medicine)#Proposed copyediting, including resolution of conflicts with NOR policy and not meeting with much success. Perhaps someone else can do a better job (or maybe I'm wrong, and WP:PSTS policy needs to change and say "except in medicine, where press releases are secondary sources"). — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 05:21, 16 August 2015 (UTC)

Sigh... yet another debate focused on "is X Primary or Secondary"? One of the problems with PSTS is that it is written as if there were a neat divide - where a source is either Primary or Secondary (or Tertiary, but for some reason Tertiary never gets argued about). PSTS neglects the fact that sources can contain a mix of both... a source can be Primary for statement X and Secondary for statement Y ... Another problem is that too many editors are still equating "Primary" = "Bad" (no matter how many times we point out that "Primary sources may be used, as long as they are used with caution"... our editors just don't seem to get it). We need to rewrite PSTS so debates no longer focus on "is X primary or secondary?" (which is often irrelevant) and instead focus on "Is X an appropriate source for statement Y?". Blueboar (talk) 14:44, 17 August 2015 (UTC)

@Blueboar: I join you in sighing. WP:PSTS is not well worded and based far too heavily on distinctions that work in history and allied subjects but not in science. Peter coxhead (talk) 14:58, 17 August 2015 (UTC)

Time to once again recount the history of PSTS... At one point, many years ago, this policy stated something along the lines of: "Wikipedia is supposed to be a Tertiary source, and adding Original Research turns Wikipedia into a Primary source." To help readers understand that statement, we decided to define the terms Primary and Tertiary source (which meant also defining what a secondary source is). In that original context, the PSTS definitions made some degree of sense. The definitions used in the humanities fit the context. Unfortunately, in subsequent edits we removed the bit about Wikipedia being a Tertiary source and the warning about not turning Wikipedia a Primary source... and without that context, the definitions have been problematic ever since.

The question is... how do we resolve this? My choice would be to return the context, rather than once again tweak the definitions... since it was the context statement that directly related to our policy on "No Original Research" (it was an explanation as to why adding OR is bad), not the definitions. Blueboar (talk) 22:21, 17 August 2015 (UTC)

I'm in the sighing-about-PSTS club too, and I signify my membership of this rarefied group by using lots of distracting emphasis in my talk page posts. ;-)
Seriously, Wikipedian policies and guidelines on primary and secondary sources have got to be sorted out. As written they work well in the humanities and particularly well when dealing with astrology, cryptozoology, conspiracy theories, homeopathy etc. However, they're crap and useless for mainstream science and mathematical articles. If you're dealing with a proper scientific topic, then the published paper from a professor at a mainstream university is a primary source and a journalistic article that reports the professor's findings is a secondary one. However, any contradiction between those sources should always be resolved in favour of the world-leading expert rather than the journalist who's semi-understood what the professor is saying. The primary source is more reliable than the secondary one. Getting this changed is quite important. It's also likely to be an uphill struggle of similar magnitude as the one about verifiability not truth, but I think it could be worth it.—S Marshall T/C 20:37, 19 August 2015 (UTC)

I suspect numerous problems with our PSTS definition is that while we define a secondary source as "one step removed" from the event, in practice a secondary source is one that is transformative (a point only captured in a footnote on the WP page); this latter is demonstrated by both WP:NOR and WP:N. We want to build WP articles on transformation that puts the primary source in context of some larger world view that we as WP editors cannot do without engaging in OR. It also strongly depends on the topic too, as the same source that may be secondary for one thing may be primary for an other, so its difficult to classify a source that works for all of WP the same way. --MASEM (t) 21:27, 19 August 2015 (UTC)

Well, I don't think we always want something transformative. The example article I always give is HIP 56948, which is almost entirely based on two papers, both with the same principal author (Professor Jorge Melendez, who is the world's leading expert on the star), published in 2007 and 2012. There are journalistic sources, but all they do is parrot what Professor Melendez says ---- there's no transformative interpretation there, and if I came across a secondary source that was transformative, I'd be suspicious of its accuracy. Because in proper scientific topics, all the interpretation should be done by scientists, not journalists.—S Marshall T/C 21:48, 19 August 2015 (UTC)

Transformative does not necessary mean that it is necessary more reliable, only that it is more applicable to the mission. In the case of HIP 56948 this New Scientist article is the transformative work, even if it mostly restating what the primary source of the published papers put out; it establishes why the identification of that star is important (possibly of extraterrestial life, etc.) They aren't transforming the science and data that Prof. Melendaz has performed, but instead establishing why this science is important, which may be patently obviously to anyone in astronomy but not to the population at large. I would not expect the NS article to be authoritative (though more authority than a more mainstream newspaper) and still defer to the original papers (primary sources) for the accuracy, but to place it within WP the NS article is critical as a key secondary sources (There might be others, I plunked that one to spot check). It comes back to that we want to place value on secondary sources to help distill and highlight good information to avoid WP:IINFO, but balance and fill necessary gaps with primary to be accurate and informative. --MASEM (t) 22:10, 19 August 2015 (UTC)

I'm really talking about something that satisfies four limbs. (1) A verifiable full professor of a major university, (2) Writing about a topic in the professor's main field, (3) Published in a scholarly journal, and (4) On a topic in pure science or mathematics. Assuming all four of those limbs were satisfied, Wikipedia should clearly place more reliance on what the professor says and less on what interpreting nonspecialists say about it, so any conflict should be resolved in favour of the scientist. The secondary sources on HIP 56948 are really there to address Wikipedia's notability obsession; to be quite honest the only reason I bothered to include them was to AfD-proof the article.—S Marshall T/C 22:17, 19 August 2015 (UTC)

Which is why we should remember that our goal here is to be a summary work, not simply documenting because we can even with the most reliable, most accurate, most direct source. It's why NOR stress balance on secondary vs primary; primary sources are not bad, by no means, and absolutely have their place for good article building, but it is the secondary that help us guide proper summation of a topic; otherwise, we all have our pet interests and could easily use solely primary sources, even of the quality of these example academic papers, to populate without heeding WP:NOT#IINFO. The "one step removed" definition (which even as stated is a rule of thumb and not meant as a hard definition) that we use as the main definition is what gets us into a lot of arguments when it comes to notability and NOR because of how editor take that advice to heart. See how many times the issues of newspaper articles being a primary or secondary source comes up. --MASEM (t) 23:20, 19 August 2015 (UTC)

I do remember that our goal is to summarise, I assure you. :-) Secondary sources are a good way to judge whether we should have an article on a given topic. If it's established that we should have an article, that's where we have to evaluate the available sources and decide which ones are the most reliable, and I think that's the context of this discussion. User:SMcCandlish was asking about which sources are the most reliable and I think the point we're coming to is that scientific articles should say what the scientists think, mathematical articles should say what the mathematicians think, and medical articles should say what the physicians think. The professionals are the most reliable sources even when they're primary, wouldn't you agree?—S Marshall T/C 23:35, 19 August 2015 (UTC)

There's nothing reliable at all about newly published research that hasn't been confirmed by others; this is why we don't want to rely on fresh-off-the-press papers in journals but on literature reviews (secondary sources) in the journals, and recent books from academic presses. It doesn't matter to WP that for a research scientist, at work, a literature review is "catch-up homework", light reading to skim for what's going on the field lately (to follow up on with reading original papers of interest), or that topical books in the field are "just popularization" or "intended for students", while he/she goes back to reading primary research papers. WP is not a laboratory. Someone the other week on this page said something like "why should I cite a literature review summarizing papers when I can just cite the papers?" Here's why: Plenty of research is more inconclusive than its presenters think it is, and plenty of it's downright disproven later, or goes nowhere, or remains controversial for a generation. The principal purpose of publication of papers in journals is intra-profession, inter-professional communication of research that needs further attention, not informing the general public of dependable facts. Literature (including systemic) reviews are closer to the later (though still intended for experts for the most part). For our pruposes, they tell subject-expert Wikipedians whether the papers published recently about the topic in question are passing sniff tests or in some cases seem to be off-kilter, and they allow our readers with any paper-researching experience where we're getting our information, in a way that doesn't require an advanced degree to assess for whether the WP article is doing its job properly. Cite the lit. revs.; add citations to the original papers too, if you want, but do cite those secondary lit. revs.

As for the "a press release from the FDA is a totally reliable source", since the discussion linked to above where I explain why it's not; it's a political primary source, of what the organization's stance is, based on both research being reviewed and political pressures. One organization is apt to take a different stance than another based on the socio-politics shaping which sources they decide to give weight to. It's not that we can't use organizational position statements on medical and other scientific issues, we just have to attribute them. "According to the AMA ...".

This is not difficult stuff to follow. I think the only reason anyone's having any difficulty with WP:PSTS and having a "not this secondary and primary sources stuff again" reaction, is they're refusing to acknowledge that WP has it's own clear definition of these terms as they apply here, but people keep trying to apply their own field's definitions (and weighting) on WP. Stop doing that, no more problem. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 00:29, 20 August 2015 (UTC)

Newly-published research is slightly different from what I was talking about; the example article HIP 56948 was based on sources several years old at the time. But I do think that recent research is one of the things that Wikipedia does well. Or at least, one of the things that Wikipedia ought to do well ---- our material can be updated almost in real time, so our content on science and mathematics can, and should, reflect the very latest research. I'd certainly question the wisdom of excluding recently-published papers in scientific topics! But in cases where we do take the decision to include bleeding edge research, then it should normally come with labels such as in-text attribution.—S Marshall T/C 00:42, 20 August 2015 (UTC)

Medical topics need to have special treatment, with the reasonable excuse that we don't want Wikipedia to kill anyone. Leaving medical topics aside, one of the great strengths of WIkipedia is its legions of qualified editors who are able to understand and summarise recent peer-reviewed research. Not only in science and mathematics, but in most topics. History for example: it would be a disaster if academic history journals were excluded since they are in fact the most reliable sources. Zero^talk 01:28, 20 August 2015 (UTC)

The only "qualification" necessary to edit Wikipedia is an internet connection, consequently, Wikipedia editors generally shouldn't be interpreting primary sources. Geogene (talk) 02:02, 20 August 2015 (UTC)

There's a different factor here at work that primary/secondary/tertiary or high vs low quality does not cover yet, as applied to "new" research, whether it be medicine or astrology or the like. I don't think we have a good measure of this in any sense on NOR/PSTS, but it is reflected in the idea of using peer-reviewed sources, or in certain fields, identifying the best and/or worst journals in terms of how rigorous that is. But how to define or capture this in language, I don't have an immediate answer to. But I do believe this is beyond just the difference between primary and secondary. And this is not just limited to sciences, it does apply to historical topics as well as breaking current events, with the idea that the farther out from the event - so that there's more separation from that piece - the more likely the source will have been able to address various mis-stated facts, reconfirmations of theories, and so on. Again, this is more than just primary vs secondary, but some other quality we yet to have identify. --MASEM (t) 02:03, 20 August 2015 (UTC)

Yes, when we treat primary sources as such: attribute them and use them with caution. No way, no how should WP be reporting "the latest research" on anything as accepted fact in WP's own voice. I don't know what it is about this that is making people's brains melt in certain topical areas on WP. "Aaaaiiiieee! He's saying this source I like is primary! My life is over!" FFS, people. It just means attribute what it's saying instead of repeating it as fact in WP's own voice, and give it WP:DUE weight. This is easy. PS: WP:PSTS does cover "new research": Primary sources are original materials that are close to an event, and are often accounts written by people who are directly involved. They offer an insider's view of an event, ... and so on. [...] A ... scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment. A lit. rev. is not a primary source. An FDA position statement usually has some tertiary (summarizing) material in it based on [one hopes] the same sources as a lit. rev. on the same topic, but is a primary source otherwise: Primary sources are original materials that are close to an event, and are often accounts written by people who are directly involved. They offer an insider's view of ... a political decision, and so on. Even position statement issued by such an organization is highly political, even for NGOs. They employ policy analysts for a reason (I've worked with some, e.g. at AAAS). — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 02:13, 20 August 2015 (UTC)

It's not quite as simple as Primary = Attribute and Secondary = Say it in Wikipedia's voice. Yes, that works in many cases... but there are lots of secondary sources that should be attributed (and there are at least a few primary sources that are so well established as being accurate that what they say can be presented in Wikipedia's voice). The question is... how does any of this relate to the concept of No Original Research. Blueboar (talk) 02:31, 20 August 2015 (UTC)

Agreed, and I was not trying to oversimplify application, just make the point that "primary doesn't mean it's banned", people WP:DONTPANIC. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 20:39, 25 August 2015 (UTC)

As I've said in other places, I see very few policies and guidelines in Wikipedia mentioning the primary/secondary distinction that wouldn't be immediately improved by removing that distinction. The variation in what sources are primary, from a medieval manuscript in a dusty archive to a research paper in Nature, is so great that practically any guideline using the generic concept of a primary source is necessarily broken. Unfortunately, such guidelines were written by people who generally did not have the full gamut of Wikipedia coverage in mind. Take a look at our mathematics articles; if we restricted ourselves to sources that a typical editor could understand, that entire part of Wikipedia would disappear. A tragedy. We have excellent articles in mathematics because of our mathematically trained editors; we should not pretend otherwise. And banning use of recently published research would serve only to ensure that we are always out of date. What we should be doing is to mostly throw away the primary/secondary distinction and instead focus on what actually matters: reliability of sources and avoidance of original research. Zero^talk 02:21, 20 August 2015 (UTC)

The problem with the math articles is that non-mathematicians can't follow them...forget the sources. Geogene (talk) 02:28, 20 August 2015 (UTC)

A thing to remember is that a secondary source does not need to be mainstream, just transformative. For a complex math theory, a secondary source can come from a larger review article published later that places it in context in the field, yet still not "accessible" in terms of understanding to the layman. --MASEM (t) 04:06, 20 August 2015 (UTC)

Right. Most biology literature reviews are gobbledygook to non-specialists but are completely crucial for our sourcing of biology articles here. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 21:47, 26 August 2015 (UTC)

I agree with Geogene (except where he says "forget the sources"). Maths should be treated like a foreign-language source. If I cite a mathematical equation, that's like citing a source in the original Patristic Greek ---- you've got to be a specialist to understand it. In fact there's another issue as well. There are particular problems with maths articles because examples are vital to the reader's understanding but you can't just copy/paste an example from a textbook because of copyright. Therefore we need editors who're maths graduates to produce original calculations; I've always felt that there should be a topic-specific exception to the limits of WP:CALC in specialist mathematical subjects, where WP:V is satisfied by producing an equation that other mathematician editors can check. In practice we act as if there is one.—S Marshall T/C 07:25, 20 August 2015 (UTC)

How could there be a mathematics topic-specific exception to WP:CALC, which is by definition about mathematics? That would seem to be the same as a WP:BLP exception for articles that are really about living people. Rather, WP:V demands that claims be verifiable; anyone competent at math can verify that an example in which terms have been changed per WP:Copyright / WP:Plagiarism does illustrate the mathematical point in question. We do the same thing in linguistics articles, substituting comparable terms so we don't rip off examples from linguistic texts. If a particular instance so unbelievably complex or specific that we can't construct an example without OR (does such a case exist?), we can directly quote and attribute one. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 21:47, 26 August 2015 (UTC)

Well, specific examples help here. I like to think of myself as competent at maths but I couldn't verify the examples in, say, Kronecker product. Do you believe that copying a textbook example but changing the terms avoids copyright/plagiarism problems? Because I would challenge that.

I also think of myself as not completely incompetent at languages and I would say that that's a somewhat different thing. In a linguistics article that needs examples, such as case government (and, wow, completely unsourced, want to help me fix that?) ---- you could substitute zu mir for nach mir, or even more thoroughly, zu mir and the dative for um mich and the accusative, and that would be okay because the potential range of examples is so much more restricted. (I don't think you could write about case government on the English Wikipedia without mentioning German as it's the most familiar example for so many English speakers.)—S Marshall T/C 22:41, 26 August 2015 (UTC)

I don't disagree that examples would be helpful to figuring this out and what to say about it. Not sure we need examples in the policy, if we clearly enough state what is intended. I'd ask a professional, academic mathematician if they think the Kronecker product examples are verifiable or not with the standard materials and processes in their field. My point about linguistics being like mathematics was that if you understand the verifiably sourced algorithm in question, you can generate (and others can verify) new examples without it triggering WP:OR concerns. As for copyright: Changing textbook examples is or is not a copyright problem depending on the nature, complexity, and originality of the examples, in part, and the extent to which we do it from a single source, and replicate its presentation, in part. There are multiple copyright provisions that would affect this, part of them within fair use rules, part of them not. There's absolutely nothing wrong with changing "x / c = y" to "a / c = b", where x and y are just variables. That's not comparable to taking a complex, original word problem and doing nothing but changing "Alice" to "Bob". — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 01:11, 27 August 2015 (UTC)

In a previous message Zero wrote,

"What we should be doing is to mostly throw away the primary/secondary distinction and instead focus on what actually matters: reliability of sources and avoidance of original research."

So far, I think this is the best comment in the discussion. Does anyone disagree with Zero's comment? --Bob K31416 (talk) 20:44, 27 August 2015 (UTC)

SMcCandlish, Blueboar, Peter coxhead, S Marshall, MASEM, Zero, Geogene, or anyone else:

Since it's been a couple of days with no one disagreeing with Zero's comment, would anyone care to make a concrete proposal that implements it or any part of it? --Bob K31416 (talk) 15:29, 29 August 2015 (UTC)

We have to be careful to simply "drop" P/S as other guidelines do depend on it like notability. I think it is important to establish that secondary means some type of appropriate transformation which cannot be otherwise done without engaging in OR (eg critical review of a work) . --MASEM (t) 18:03, 29 August 2015 (UTC)

True (although WP:Notability has somewhat moved away from a simple P/S distinction in recent years... it now focuses more on a dependent / independent distinction). But that begs the question... if P/S is important to notability ... shouldn't it be covered at WP:NOTE? Is WP:NOR the right place to cover P/S? Blueboar (talk) 12:22, 30 August 2015 (UTC)

No, notability (at least in terms of the GNG) weighs on secondary coverage of topics. Independence is a requirement per WP:V. I think the P/S/T distinction should be a neutral guideline page along with explaining the other axes we use to evaluate sources like independence, first/third party, etc. so that when other policies and guidelines reference these terms, we have them as standard. --MASEM (t) 12:59, 30 August 2015 (UTC)

Masem, How would you like to proceed on creating such a guideline page? --Bob K31416 (talk) 13:48, 1 September 2015 (UTC)

It would probably start with what we have here at PSTS, but add in the other axes with similar explanations of how to identify them, related gotchas, and the like. A "field guide to source identification" of sorts. It should be a guideline because these are not always hard-line, and also taking into account Blueboar's point that a single source can be both primary and secondary at times, so it's not a final line for any source, either. --MASEM (t) 13:54, 1 September 2015 (UTC)

I'm not familiar with the process so correct me if I'm wrong, but I think the first step is for you to create an essay page. You might want to create a section here first, outlining what you want to do, to generate or see what interest there is and get useful ideas. --Bob K31416 (talk) 14:09, 1 September 2015 (UTC)

Just a further thought... we keep talking about primary/secondary/tertiary sources... rather than primary/secondary/tertiary source material. This may account for some of the problems we have with definition... after all, a single source might contain material that is primary (original data) as well as material that is secondary (analysis/interpretations of that data, or conclusions drawn from such analysis) and material that is tertiary (summation of the conclusions of others). Blueboar (talk) 12:22, 30 August 2015 (UTC)

Blueboar, Would you care to propose an edit for WP:NOR regarding that? --Bob K31416 (talk) 14:09, 1 September 2015 (UTC)

I would love to... but at this point I am not ready to do so. I have to think about my concerns and ideas further before I could propose a change. I can identify flaws in PSTS, but I have not yet fully decided how to address them... what is important is to fix the flaws... but isn't important that we do so right now. Blueboar (talk) 14:32, 1 September 2015 (UTC)

In your statement of the proposal, you could qualify it by saying that it is just a tentative idea to get useful comments from others, including possible alterations. In other words, you could still be noncommittal. --Bob K31416 (talk) 15:37, 1 September 2015 (UTC)

We've got to think more about this "independence" point. In, for example, articles about drugs, the best source may well be one that's funded by the pharmaceutical corporation that manufactures the drug. I think we've got to say that when the author is an academic scientist, it's editorial independence rather than financial independence that matters.—S Marshall T/C 09:17, 31 August 2015 (UTC)

See Masem's suggestion above for a guideline page. --Bob K31416 (talk) 13:56, 1 September 2015 (UTC)

Scientific sources, WP:SYNTH and OR in current events articles

I have been having a conversation at Talk:2015 Gold King Mine waste water spill about synthesis. Example of content deleted: diff

The material in the diff is said to be synthesis, I believe for the following reasons:

Common knowledge relating to acid mine drainage; SYNTH claim: stating that this water is acid mine drainage
The USGS reference to the average flow rate of the river; SYNTH claim: comparing the flow rate of the river during the spill
Research about acid mine drainage at mines in the area around Summitville, Colorado, and elsewhere; SYNTH claim: applicability to the mine in Silverton, CO, that was involved in the spill
USGS description of polluted rivers in 1996-97 + recently released test results for Fe and Cu levels in the river; SYNTH claim: connection between conditions 20 years ago and current conditions (i.e. it's irrelevant that the river was frequently severely polluted in the past, was cleaned up, and has become polluted again)

Editors objecting to the content have done so by saying (paraphrased) "we can't use that" and "we need reliable secondary sources." My understanding of WP:SYNTH is that it is permissible to state a conclusion that is supported by the sources. In the natural sciences, it is possible that the sources support a conclusion that they do not expressly state. (I have commented out an example here.)

I don't see why it's necessary for a basic scientific fact to be linked to a specific event by a journalist before it can be on Wikipedia. This is the one place where there are people who actually want to add science to news articles that need it, and they are usually met with condescension and dismissed out of hand. Editors are told that "I know it's true!" isn't acceptable, when they are actually saying "This is a verifiable part of the scientific consensus that I have supported with references."

I hope everyone will agree that it is in the best interest of WP and in the spirit of the pillars that articles contain scientific facts. If you don't agree, please try to understand that I am only referring to basic, simple facts where there is scientific consensus. And what I am objecting to is the application of SYNTH to things that are common knowledge, not original research. What can be done to help good-faith efforts to explain the science behind current events?

Statements: I have no CoI here. I have notified involved editors on the article and placed a note on the OR noticeboard. Roches (talk) 02:08, 16 August 2015 (UTC)

Summary of this section: application of OR to scientific concepts that represent a consensus within the relevant field but have not been mentioned in secondary source coverage of the event. Roches (talk) 02:19, 16 August 2015 (UTC)

Are you proposing a change in policy? If so what change are you proposing? If you aren't proposing a change in policy, I can't see why this needs to be discussed here rather than at WP:NORN. AndyTheGrump (talk) 02:40, 16 August 2015 (UTC)

I think you cannot compare waste pools as you did in your dif during a current event, and you require data on actual water quality tests to make any judgement. However you could link to the page, acid mine drainage and mention the common contaminants. And i see that the article mentions this under the background section. prokaryotes (talk) 04:06, 16 August 2015 (UTC)

Scary to say, but I think I am proposing a change in policy. I chose here instead of NORN because this has happened before. I'm not particularly interested on the dispute given in the example; it's just the most recent case I've discussed. You're right that the specific article shouldn't be discussed here. @Prokaryotes:I wasn't involved in the diff and shouldn't say much about the details, but there were water quality tests that showed high Fe and Cu levels. Some of the things in the diff might be unacceptable synthesis. Here, mostly, I wanted to discuss WP:SYNTH itself before I get exhausted over a single article, as I'd intended to do that last time something like this happened.

What I'm thinking is something like: "Verifiable, reliably sourced scientific information may be used in articles about current events, provided the relevance of the sources used is unambiguous, even if the information has not been cited by a news outlet or other secondary source." I'm trying to be brief, but I think that's the essence of it. Roches (talk) 04:37, 16 August 2015 (UTC)

Maybe this was a bad example but like in the background section mentioned, you can use related info in current events. The question might be how extensive and imho this can be done without any synthesis, just to give background relevance or point to similar events of the past. prokaryotes (talk) 05:03, 16 August 2015 (UTC)

Roches... in your comments above you say: "My understanding of WP:SYNTH is that it is permissible to state a conclusion that is supported by the sources." Not quite... we have a higher standard than that... we require that the conclusion be directly stated in the sources. The sources may provide all the pieces of the puzzle (the evidence that supposedly supports the conclusion), but if it is a Wikipedia editor who puts those pieces together (and forms the conclusion), then it most definitely is Original Research. Wikipedia becomes the first place to state the conclusion.

In another comment you say: "I don't see why it's necessary for a basic scientific fact to be linked to a specific event by a journalist before it can be on Wikipedia.". You are partly correct. It doesn't have to be stated by a journalist... but it does have to be stated by someone we consider reliable (for example, a statement by a noted scientist would be quite acceptable). The link must be made by a reliable published source outside of Wikipedia. Again, if Wikipedia is the first place where the fact is linked to the event... that's Original Research. Blueboar (talk) 11:53, 17 August 2015 (UTC)

Also note that the current policy already allows basic calculations. However, as soon as averages come in, no calculation is basic anymore. An average in this case implies that you do not have the exact volumes of the river flow or the leakage (up to gallons precise) and the exact concentration of chemical in the leakage at the exact moment (up to second precise) of the spill. This means that your calculation will provide an estimate, and estimation is original research. (For example you would have to provide assumption about the distribution about the means, and how this impact confident intervals of your calculations, all of which is essential to interpret your outcomes but is way beyond what can, or should be done in Wikipedia) Arnoutf (talk) 14:01, 17 August 2015 (UTC)

I don't disagree with the conclusion reached about this particular example, but be careful with statements like estimation is original research. In some cases it will be, but as per WP:SYNTH, it's not the estimation that's the problem, but whether it is used to advance an argument, rather than, say, to provide a succinct description. Peter coxhead (talk) 14:56, 17 August 2015 (UTC)

You are right I may have been overly generalising about estimation not being simple calculation (and hence OR). In this specific case the selection of the numbers used as input for the estimation and the way they are combined are the problem (I would agree for example that the estimate that for a fair 6 sided die 1, on average 1 in 6 casts would results in a 6 would not be original research - and that the estimated average value for such a die would be 3.5 is also no synthesis) Arnoutf (talk) 15:12, 17 August 2015 (UTC)

Support policy change proposed above by Roches 04:37, 16 August 2015 (UTC)

Education is learning to see what you're looking at. That's true for sports fans who know the players and patterns of play. It's true for birders who can identify species from a momentary glimpse or a fragment of sound. And it's true for scientists who often can see what process dominates a situation from reports of nonprofessional observations and a knowledge of context. Immediate examples include acid mine drainage as in the 2015 Gold King Mine waste water spill, earthquakes caused by human action, categorization of landslide causes, and the uncertainty of recurrence intervals during a flood event.

News media necessarily focus on the most extreme aspect of any event, the abrupt disruption of people's lives, their reactions to loss, and their fears for the future. Yet in environmental science, an event can be spectacularly abnormal at local scale, fairly normal at intermediate scale, and undetectable at large scale. Every event tapers away to zero at some distance from the most intense point. An abrupt arrival is followed by gradual dissipation. Both parts are necessary to the story, but the return towards normal is not ″news.″ When scientific context is excluded, the story easily becomes one of disaster that induces more fear in the audience, as falsely portrayed in the Gold King edit history. Half-true stories do not serve our readers well.

Policy should encourage experts to contribute to news stories in three ways—process identification, scale, and context—to communicate expert knowledge to the large majority who are not expert. Once an expert sees what's going on, the first step is to identify the foundation concepts and language of the topic area, to avoid wasting time reinventing wheels. The second step is to summarize the scale and mass balance of the process. Whether brewing beer, cooking eggs, or mixing concrete, knowing both batch size and the active ingredients is important. Finally, context should provide pointers to relevant literature within or beyond WP.

As presently implemented, all three contributions are forbidden as OR or Synth. We are wasting important brain power by requiring experts to find a journalist, explain the details, hope the journalist will publish without messing it up too badly, and only then cite the news report in WP. A really old quip says don't try to teach a bear to dance. He'll never be very good at it, the audience won't like such a bad show, and it makes the bear cranky. Instead, let's assume good-faith for technically expert editors as well as all others.

We can convey lots of useful information within limits of identification, scale, and context. I agree that going beyond those three is OR, but the fact that one editor is not familiar with a particular topic area should not make another editor's content in that area automatically OR. Justaxn (talk) 20:53, 27 August 2015 (UTC)

@Roches, Prokaryotes, Blueboar, and Arnoutf: Continuing the idea of the previous sentence, I suggest that science statements in news articles should only be flagged as possible OR by an editor not familiar with that science topic rather than removed immediately. In the article's Talk, that editor should identify the concern and invite other editors who are familiar with the science to weigh in. If one or two others chime in that the questioned content is in fact limited to identification, scale, and context, then the flag can be removed. But if another editor says ″I know this topic, and you're stretching a bit too far,″ the content goes away. Justaxn (talk) 22:08, 29 August 2015 (UTC)

That seems to be the current policy anyway - only we should treat reports as primary source until it is confirmed in another source (this does not have to a newspaper - e.g. a final report on the accident may also do). The main problem with the edits of the original poster here seem to have been that sources from a range of different reports, collected for different purposes at different times and most likely with different methods. While each individual source is probably true in its own context and for it own puropose here they were combined outside that purpose to form a statement about the incident that was not verbatim stated by anyone except the Wikipedia editor. That makes it OR/SYNTH - and the there is no way to judge whether the conclusion is true (and if there would be there would have been experts making such claims openly).

Also going back to your previous post. Wikipedia is a tertiary source; i.e. a source that neither presents new facts (primary) nor new inferences based on existing facts (secondary). That means by definition that a lot of the creativity in inference making of our editors is lost. This is a good thing for an encyclopedia. If you want to be creative in dealing with this kind of issue take up a career as journalist or scientist and publish your own primary and secondary papers (that, if sufficiently representative of mainstream agreement, main then influence Wikipedia). 09:06, 30 August 2015 (UTC)

If flagging and discussing possible OR were in fact current policy as practiced, this discussion never would have started. Instead, content was deleted without evidence the deleting editor followed any of the cited links. One of those links was to bibliography of >150 articles on acid mine drainage by one research group at sites mostly in one state (Colorado), where the disputed incident occurred. I personally did studies of AMD for 15 years in AZ and WV. AMD has been well known for centuries around the world, eg Rio Tinto (river). Any person with a modest education in AMD could look at the photos and news reports and say with high confidence ″this is acid mine drainage.″ That is no different from a person with a modest education in classic cars looking at film of a vehicle on the highway and saying ″this is a Ford Mustang from 1968-72.″ Both statements are essentially trivial, self-demonstrating facts. Thus my suggestion above, that disputes be referred to others, such as any 2 or 3 editors whose substantive edits appear in the history of acid mine drainage, in this instance.

The alternatives proposed just above are absurd in context of a current events article. The ″final report on the accident″ becomes available not less than a year later, and publishing new science articles in any venue has an even longer lead time. Further, the new knowledge that might be available through this incident is far less than the Least Publishable Unit that any science journal editor would require. This incident adds nothing to science that would be worth scientific effort to publish, but existing established science does provide context that could be useful to citizens in the region. But WP editors need to be open to seeing what they're looking at. And I'll sign this. Best wishes to all. Justaxn (talk) 02:34, 31 August 2015 (UTC)

The content I removed was this [2], which contains a lot more than a mere statement that it was "acid mine drainage", it was a POV-push towards "nothing to see here, move along" that was based on SYN and seemed completely at odds with the sense of the reliable sources, that this was more or less a disaster. Following the removal, I left the following explanation on the article's talk page [3]. Geogene (talk) 21:47, 31 August 2015 (UTC)

Thank you Geogene for vigorous editing to make the article better. Your second citation just above includes my original comment in the article Talk ″Can't we tone this down a little?″ At that point, the article had a strong POV ″Everybody's screaming--this must be a disaster.″ My disputed edit was less "nothing to see here, move along" than a recognition that every topic has at least two sides. I stated the basics of AMD to move the article from a strong disaster POV more towards NPOV. I agree my initial edit was imperfect, but merely deleting it with a shorthand OR was an inefficient way to move forward (thus this extended discussion). To avoid Synth, I moved individual statements and citations closer to related content. I linked to AMD a bit too often, and others appropriately trimmed that back. Similarly, I trimmed back the too-frequent use of ″toxic,″ not eliminating the word entirely but discomfiting editors who prefer that POV. When the article came to include 150-year-old conflicts between Navajos and Anglos, we clearly had a pickup-truck article trying to carry a trainload of POV.

The discussion has crystallized my present understanding that current events articles to which science is relevant should welcome content from science-competent editors to report process identification, incident scale, and science context. In case of dispute, the content should be flagged, though not removed, until it can be reviewed by other science-competent editors. Those reviewers should actively be invited from among editors who have made substantive edits to the relevant WP articles the first editor linked as part of science context. I couldn't have expressed that summary without this extended discussion, so thanks for that as well, Geogene. Justaxn (talk) 18:49, 3 September 2015 (UTC)

Transcribing

Hi all. I have some concern about the section WP:TRANSCRIPTION. It is not clear whether it includes IPA transcriptions or just transcribing into the language writing system from a video or audio file. I think that transcription in IPA would constitute an original research. Could someone clarify on this point? --SynConlanger (talk) 20:04, 4 September 2015 (UTC)

In my view, provided we can verify that this is a real sample of the language by a native speaker, then transcribing a real world language into IPA is not original research. The video or audio file would need to be published by a reliable source.—S Marshall T/C 09:57, 5 September 2015 (UTC)

so if I measure the Colosseum myself, I can use that measurement in Wikipedia and it won't be original research? I don't think so. But if there is consensus that IPA is ok we should state it clearly on the page. As it is, is ambiguous and subject to interpretation. 😊 --SynConlanger (talk) 11:32, 5 September 2015 (UTC)

I'm sorry, I'm lost. How is measuring the colosseum yourself comparable with performing a repeatable operation we can check, like transcribing an audio file into IPA?—S Marshall T/C 11:40, 5 September 2015 (UTC)

IPA transcription is highly subjective, so actually it is not replicable in the scientific sense. While measuring the Colosseum is. Anyway, if the consensus is towards giving IPA transcription as an exception to the OR restriction, please, state it clearly. Transcription could just well be "orthographic" transcription (which I consider not OR). --SynConlanger (talk) 12:48, 5 September 2015 (UTC)

Folks... this isn't what WP:TRANSCRIPTION is talking about... it's talking about something much more basic... if you have an audio or visual recording of someone speaking, it is not Original research to transcribe it into written English. Nor is it original research to take an audio or visual recording of someone speaking French, and translate that into written English. Your transcription and/or translation might still be challenged as being inaccurate ... but it isn't considered Original research. Blueboar (talk) 13:58, 5 September 2015 (UTC)

yes, this is exactly what I thought. --SynConlanger (talk) 16:02, 5 September 2015 (UTC)

Elaboration on "routine calculations"

I would like to add the following sentence to the "routine calculations" section: "This includes simple derivations of abstract mathematical formulas and illustrations of standard mathematical proofs." The reason I want to add this is that derivations are something different that a "calculation" which I consider to be more of a numerical result. I believe that my addition would be in the spirit of this policy, and certainly is in the spirit of "verifiable even if not verified." I also see this in practice all the time. I just would like to see it clarified here in the actual policy, but before I make a change I would like to hear if there are objections. LaurentianShield (talk) 19:15, 4 September 2015 (UTC)

I would say no... derivations would not be considered "routine" enough. The exemption is to allow very basic calculations... computing a simple average, converting miles into kilometers, etc. ... the sort of thing that would be considered routine by the average non-mathematician. (A lot more basic than what would be considered "routine" by mathematicians). To be safe... ask if someone who stopped taking math courses after one year of basic Algebra might be able to follow the calculation. That's what we mean by routine. Blueboar (talk) 01:50, 5 September 2015 (UTC)

I politely, but vehemently, disagree with Blueboar about this. I feel that the exemption should be for calculations checkable by undergraduate-level mathematicians. To argue by parallel with foreign languages: if I use a French source, then it's considered acceptable even if the source uses the past anterior and the subjunctive, because a person who properly speaks French can check its meaning. These constructions are (or were when I was learning French) taught to 17- or 18-year-old students who were going on to advanced study of the language, which puts them roughly parallel with the calculus in complexity. Why would we prefer linguists over mathematicians in this way? We do often prefer humanities over science in WP:NOR but I feel this is because the policy was mostly written and interpreted by humanities graduates rather than scientists. It makes no sense to me.

I feel that it can't possibly be original research if Newton and Leibniz proved it in the 17th century, so in my opinion calculus-level mathematics has got to get past the bar. And if calculus is over the bar then many derivations should be!—S Marshall T/C 09:50, 5 September 2015 (UTC)

Hi S Marshall, Would it possible to provide an example of such a calculation and a context in which it might be used? - Ryk72 ^'c.s.n.s.' 10:17, 5 September 2015 (UTC)

Sure.

For example, if an object's position on a line is given by
$x(t)=-16t^{2}+16t+32,\,\!$

then the object's velocity is

${\dot {x}}(t)=x'(t)=-32t+16,\,\!$

and the object's acceleration is

${\ddot {x}}(t)=x''(t)=-32,\,\!$

which is constant.

This is taken verbatim from differential calculus#Physics. I hope, and believe, that a Wikipedian has performed this calculation. If not then it's a copy/paste from a textbook and therefore a copyvio, so it would need to be replaced with an original calculation. We do need examples in mathematical articles because they significantly enhance the reader's understanding of the topic.—S Marshall T/C 11:23, 5 September 2015 (UTC)

S Marshall... there is no need to present an original (editor created) calculation in our mathematics articles. Any mathematics text book will contain examples which can be copied in our math articles (and cited to the text book).

That said... The routine calculation exemption is not talking about presenting an example of a calculation in mathematics articles... its about performing calculations to analyze or interpret data to reach a numerical conclusion. If you need to use differential calculus to reach your conclusion, that conclusion would be considered Original Research (you need to find a source that reaches the conclusion)... if you can reach your conclusion using very basic, grade school math - it's not. Blueboar (talk) 13:39, 5 September 2015 (UTC)

Hi S Marshall & LaurentianShield, I'm inclined to concur with Blueboar on this, at least for the type of "example" use provide above - referenced in the original post as simple derivations of abstract mathematical formulas and illustrations of standard mathematical proofs. This type of information is standard content and can be sourced if required. The WP:NOR policy, by my reading, refers to the use of calculations to determine content; an specific numerical calculation, not the abstract. Hope this helps. - Ryk72 ^'c.s.n.s.' 13:59, 5 September 2015 (UTC)

I certainly don't think it's a good idea to copy/paste examples from maths textbooks. An equation can't be copyright (although it can be patented in the US: State Street Bank & Trust Co. v. Signature Financial Group, Inc.). However, an application of the equation certainly can be. Our articles must consist of previously published concepts, but these concepts must be phrased in fresh and original language. And our examples must of course be based on previously published equations, but the numbers we plug into those equations would have to be fresh and original.
I agree that Wikipedians should not use mathematics to reach a non-trivial novel conclusion, but again this gets quite difficult in detail. Can I discuss a real world example? On Talk:HIP 56948: I had written an article about a star, and I had its relative magnitude from reliable sources. I also knew its distance. I did not have a source for its absolute magnitude, but I was able to derive the absolute magnitude from the relative magnitude and the distance using a standard equation:

M=m+5(1+\log _{10}{p})\!\,

Now, I realise that some humanities graduates are a bit daunted by things that look like this, but that's high school mathematics. There's nothing difficult about it and it's well below the difficulty level of some of our sample calculations from articles within the scope of WikiProject Mathematics! If I can cite Goethe in the original German on Wikipedia and that's considered verifiable and checkable, then I certainly ought to be able to calculate a logarithm.

Luckily for me, with HIP 56948 the whole problem was solved by a reliable source doing the same calculation I'd just done.—S Marshall T/C 14:52, 5 September 2015 (UTC)

I would argue that (in the case where no source gave the absolute magnitude) that knowing that that equation could be applied to give the absolute from relative + distance is original research. It may be a standard calculation for astrophysics/astronomy but its a non-standard calculation to the layperson. And it's not because of the mathematical operators involved, it's simply where that formal derived from. --MASEM (t) 15:01, 5 September 2015 (UTC)

To expand, I think as suggested above, we can only reasonably expect readers to understand routine calculations that would apply through a typical first-year high school algebra course. Even if an equation is standard or even "by definition" (for example, ph being -log10(H+) for chemistry) in a field, it should not be assumed that that is common knowledge enough to be appropriate for falling under OR's "routine calculation". Hence why at best we can talk routine calculations of basic mathematics, very basic statistics (median and mean but not st. deviation), and unit conversions but little else. --MASEM (t) 15:22, 5 September 2015 (UTC)

Masem, it's interesting to me that you say we can only reasonably expect readers to understand. By longstanding consensus I'm permitted to cite sources in foreign languages. To be fair, most people educated in the British state school system or the US public school system don't read any foreign languages at all (and a fair few of them struggle with English). But we offer no concessions to what we can expect them to understand. As long as the source is published, it's verifiable.

I can't check a source in Spanish. I don't speak a word of it. If I want to check a source in Spanish I have to ask a Spanish speaker. It should be the same rule for maths: if I want to check an equation that's beyond my personal competence, I should have to ask someone with a maths or hard science degree. Otherwise we're creating a tyranny of ignorance, where a suspicious editor is able to create a whole lot of work for me because they lack the skills to understand the evidence I'm citing.—S Marshall T/C 16:45, 6 September 2015 (UTC)

I strongly agree with S Marshall about the "tyranny of ignorance". We are constructing an encyclopedia; the editors of an encyclopedia not only can, but should, collectively, have a higher degree of competence than we expect of readers. "Verifiability" does not mean that anyone will be able to verify anything. I mainly write articles about plants; the sources are often written in highly technical botanical language. The "average reader" isn't supposed to be able to understand this (and anyway, those who do won't need to read the Wikipedia article). What's important is that an editor who does understand the language of the source can verify that what is written is an accurate account. This applies whether the source is in the abstruse language of the topic, whether maths, physics, botany or sociology, or a foreign language (or both, e.g. botanical Latin). Peter coxhead (talk) 19:09, 6 September 2015 (UTC)

The foreign language source is a different thing, I would say, closer to unit conversion than application of a field-specific equation. Yes, there are a lot of nuances in translating a language that, for example, Google Translate will not properly capture, and where a foreign language source is being used for a controversial claim we do strongly suggest the editor seek out an expert in the translation rather than relying on their own approach, which is effectively the use of a secondary expert source. However, at the end of the day, the foreign language translation should not be introducing new information that wasn't already in the source), simply information in English. On the other hand, using field-specific calculation to bring in new information that wasn't reported by a source, like the above absolute vs relative magnitude, that's where OR could be coming in, and why we'd want to avoid that case. --MASEM (t) 20:04, 6 September 2015 (UTC)

Well, new information? In that case I had a whole platoon of reliable sources saying this is the most sun-like star that's yet been discovered by man ---- which is an extremely strong claim, given the number of stars out there, so it absolutely did need lots of sources, before it could be said. If I'd reached that conclusion on the basis of my own spectrographic analysis then would unquestionably be original research and, indeed, horribly poor encyclopaedic judgment on my part to have written on the basis of the calculation, given that I'm not an astronomer.

But the situation was that it had been established beyond all doubt by academic astronomers as an extremely sun-like star. It had been shown by reliable sources to have a sun-like volume, density, temperature, colour and chemical makeup. My "novel step" was to perform a fairly simple calculation showing it was also of a sun-like spectrographic class. And I think that's one of the features that applies to WP:CALC ---- it can't reach an unexpected result. It can only reach a result consistent with what the reliable sources say, otherwise it's not a WP:CALC.

Let's try another example. I can easily prove by reliable sources that Mars is an oblate spheroid closely approximating a sphere, with an equatorial diameter of about 6,800km, but (let's pretend) I've got no source for its area. So I use one of Archimedes' calculations and say, oh look, the surface area of Mars is 1.45x10^8 km². I think that's a WP:CALC-style trivial calculation. However, if I then take that and go on to say "The land surface of Mars is about the same size as the land surface of the Earth", then that's not a WP:CALC-compliant trivial calculation because it's reaching a surprising conclusion (even though it's perfectly accurate). I think that's where we go beyond WP:CALC into new information.—S Marshall T/C 20:58, 6 September 2015 (UTC)

And while "My "novel step" was to perform a fairly simple calculation showing it was also of a sun-like spectrographic class" is something that I'm sure all sources would have agreed too as well (including the one you apparently found later), that is a potential bit of OR if no other source made that assessment. I would agree the Mars/Earth calculation is similarly one that is introducing a novel thought that if not already in a source, is original research, even though I would agree that the calculations and comparisons to get to it would fall under CALC. It starts to push correlation vs causation-type arguments, even if you can justify the comparison. I am not say that the result of your calculations in this are not correct, just that the resulting conclusion appears novel, even if the novel aspect is interesting or the like. --MASEM (t) 21:07, 6 September 2015 (UTC)

Wrong rules about primary vs. secondary sources

This is all wrong. You should not use UNREVIEWED primary AND secondary sources. It remains a mistery to me why you want to keep reviewed primary sources out of Wikipedia.

In fact, reviewed primary sources are much, much more reliable than any secondary source. Today, everybody can publish a book and write everything in it. Your "rules" really don't make sense and aren't scientific at all - rather ideological (...or economical...).

As such you should mention the real truth:

Do not analyze, evaluate, interpret, or synthesize material found in a primary source yourself; instead, refer to reliable secondary sources that do so. Do not base an entire article on primary sources, and be cautious about basing large passages on them. Do not add unsourced material from your personal experience, because that would make Wikipedia a primary source of that material. And do not think on your own, be arrogant and unscientific just repeating what some other guy wrote and especially pharmaceutical companies tell you to do. And they're basically telling you this: "Don't use primary sources aka objective and independent research on Wikipedia." I only answer: "Nice try!" --81.6.59.42 (talk) 20:19, 14 September 2015 (UTC)

Where did you get the idea that primary sources are necessarily objective or independent? And while anyone can write a book, anyone can also publish a primary research paper in a bottom-feeding journal and get the "peer reviewed" star attached to it. Problems of reliability are why we restrict sourcing to publishers with reputations for fact checking and accuracy. Restricting controversial claims to secondary sources is largely to handle issues of significance. You can find otherwise-reliable primary sources for essentially every viewpoint - not all of them need to be mentioned. Someguy1221 (talk) 01:21, 15 September 2015 (UTC)

Also where did you get the idea that we keep primary sources out of wikipedia? We do use them, with caution. The reasons are explained in great detail in our guidelines. Staszek Lem (talk) 02:15, 15 September 2015 (UTC)

Come on boys, don't be rhetorical, there are many people on Wikipedia argumenting to remove any primary ressources "according to the guidelines", and also this guideline is basically suggesting it. One example can be read here (on the bottom): https://en.wikipedia.org/wiki/Talk:Pyrroloquinoline_quinone#Use_medical_reviews (it's about a neglected vitamin, proves are here: https://books.google.ch/books?id=FoDtCAAAQBAJ&pg=PA161).

I am also not arguing about the quality of peer reviewed primary studies versus not-at-all reviewed secondary sources...it's too obvious that the latter is of much less quality. It doesn't even matter who wrote the secondary source, because nowadays even University professors have ties to the pharmaceutical industry which are sponsoring some of their labs or projects. In fact, only the quality of a study and its reproducibility is important, only this is real, empirical science.

Finally, please no personal discussion about me and rhetorical questions like "where do I get these ideas from?", because it's not about me but rather about you and the Wikipedia. But again: "Nice try..." --81.6.59.42 (talk) 08:10, 15 September 2015 (UTC)

Who said we accept unreviewed secondary sources? Someguy1221 (talk) 08:35, 15 September 2015 (UTC)

For the record, that article you link to is god-awful - almost entirely based on primary research papers. It should be gutted and replaced with something that reflects the medical consensus, not a hodgepodge of preliminary studies. Someguy1221 (talk) 08:36, 15 September 2015 (UTC)

No, the Wiki article isn't "god-awful" but good science. And what about the Google-book link I gave you? Even with mouse and rat experiments and fotos for the blinders like you, Someguy1221. Just empirics, you know...furthermore you didn't even mention WHY this article should be "so bad" except that it is "only primary sourced" (for the record: Above you guys were asking me the question WHO is neglecting primary research and now YOU are doing all the same), but that's typical for unscientifical rhetoricians (anyway: Who are you working for? Since you were incorrectly using the term "preliminary research", it's probably a multi-billion dollar pharmaceutical company, right?). But if you wanted a secondary source about the topic (there are hundreds): https://books.google.ch/books?id=Whu_zOWPaTQC&pg=PA415&lpg=PA415&dq=pqq+%22vitamin+q%22

To answer your question: Because there ARE no reviewed secondary sources, and that's exactly the reason why I criticized these guidelines here as wrong and unscientific (the only exceptions are tertiary sources, but again: Any random guy can write a "tertiary" book!). You don't really know how medical publishers and scientifical magazins work and review empirical primary research, do you? And beside this: Why don't you do your own studies to PROVE that PQQ is no vitamin instead of wrongly judging basic science and useful scientific research??

The truth is: Although produced by intestinal flora, biotin and vitamin K are both considered vitamins, and as such also PQQ is a vitamin. There are no doubts about it, just people like you neglecting scientifical research and fundamental facts. So for the third time: "Nice try!" --81.6.59.42 (talk) 09:35, 15 September 2015 (UTC)

I don't understand why you keep saying "nice try". You don't seem to understand how review work is generated. Review articles are indeed subject to peer review. And once again, we do not accept a source simply because it is secondary. You may wish to read our guideline on reliable sources. We know full well that anyone can write a book. That's why we only accept content from authors or publishers with a reputation for fact-checking and accuracy. You will not find an experienced editor on Wikipedia saying that something is a reliable source simply because it is secondary. Finally, you seem to deeply misunderstand the purpose of an encyclopedia. An encyclopedia is a tertiary source. It functions to share with readers the significant viewpoints on a topic as presented in reliable, secondary sources. It's function is most certainly not to mine the primary literature for what the editors personally feel is "good science". If you want to write a good secondary review of PQQ research, you'll have to do it somewhere other than Wikipedia. Ultimately, your beef seems to be with how Wikipedia works and why it exists, not this particular guideline, which isn't even a sensible place for your complaints anyway - the Original Research policy is to prevent Wikipedia itself from being the publisher of primary research content, not a restriction on what sources we can cite. Someguy1221 (talk) 10:15, 15 September 2015 (UTC)

I still don't understand how people could think primary sources are never allowed. The policy explicitly states that we may use primary sources, as long as we do so with caution. The trick is to use primary sources appropriately. If a reliable primary source states something important... something that should be mentioned in an article... we can mention it. However, we should do so with in text attribution. This hedges the statement, and lets the reader know that any interpretation, analysis, conclusions or claims are the opinion of the author of the source, and not necessarily widely accepted fact. (Or to put it another way - we can report the fact that source X says "Y"... but Y itself is presented as opinion and not as fact). Blueboar (talk) 13:12, 15 September 2015 (UTC)

"People" think this simply because some Wikipedia editors are argumenting that way, and these guidelines here are also basically saying it (please consider reading what I've written at the beginning of this thread). However, if it's wrong to argument that way, then please change these guidelines here and start forbidding the rhetorical argumentating tricks on Wikipedia by some editors, immediately. Before this, I have to look at your answer as just rhetorical, too. --81.6.59.42 (talk) 13:39, 15 September 2015 (UTC)

I'm sorry, but I think that actually you - Someguy1221 - don't understand how reviews work and how magazines review articles before publishing it: In fact, reviewed "primary" sources are always some kind of secondary sources because of the intense review they were going through. On the contrary, this isn't happening with secondary sources, they are reviewed - in this case - only by Wikipedia editors, who I mostly don't consider to be capable of doing correct scientific work (except if an editor has an University degree in the field he's active on Wikipedia). Furthermore, your answer clearly shows me that you didn't either read or understand what I wrote above. You're just blinding out all my scientific arguments and why these Wikipedia guidelines are wrong and unscientific. So again: Any scientific magazine certainly does better work than Wikipedia in reviewing "primary" research, only the Wikipedia editors themselves obviously think that they can do it better, unfortunately relying only on mostly unreviewed secondary sources. In the end we're ending up with rhetorical philosophy but not empirical science - the pure destruction of science, if you will.

There are three stages of scientific discoveries: Unreviewed (published), Reviewed (published) and Reproduced. There might be a fourth one, which is Patented, but patents often aren't reviewed very well at all. Beside these stages, there are four evolutionary stages of discoveries: Ignored, Disputed, Accepted and Truncated. For instance, the history of quanta mechanics is a nice example of the first three stages, and if the QM theories would turn out to be correct, it would never be truncated. Now, PQQ is in stage 2 for about 20 years or so. This is interesting, because it shows us that modern science stalls and is not even able to integrate a newly discovered vitamin. It looks as if today there are too much unimportant informations cycling around and being teached, and the mass media has its unfaithful responsabilities, too (I personally would forbid TV by law for children under 14...because it destroys the brain of childs).

But back to the Wikipedia: I have to disagree with you what the Wikipedia is: It's not just an encyclopedia as e.g. Brockhaus, but an open online-encyclopedia, where "online" stands for "more than just an encyclopedia". As such, it could be used to transfer real scientific knowledge (aka "primary sources") and not just viewpoints of people (aka "secondary/tertiary sources"). By forbidding primary sources you're basically truncating the hole idea of a free and open internet information exchange platform. And as such you're simply not using all the possibilities and advantages the internet could provide for scientific researchers.

Science, it could be used to solve all our problems. But instead of that, we're using it to create more and more problems (of which I'm not going into details here). I know that there are a group of psychopaths behind most of it. But we could use science even to produce pills for these people, containing compounds and neurotransmitters which they're lacking, or enzymes activating their brains aso., to heal them and make them healthy humans again. Though, the problem is that psychopaths have not only a lack of emotions and empathy, but also a lack of learning and understanding (they just memorize everything, even have very good memories giving them some advantages in - mostly memory-oriented - school tests, but are incapable of being creative in most cases, as such totally relying on the creativity of others and parasiting them, including stealing ideas aso.). And that's the main source of all problems in our world: People who don't ask themselves the questions which they should ask themselves: "What am I doing, really? Why am I doing this? What could I do to improve myself? Why do I always give other people the blame instead of myself? Maybe it's just me who is all wrong and doing the wrong things? Maybe I should invest my time and efforts into improving myself and this world instead of destroying it? Maybe there are normal humans out there who would like to help me and my condition, instead of blaming me and holding me responsible for the bad things I'm doing? Maybe there are scientists who would like to develop pills against psychopathy and I should invest money into their efforts, for that they can produce substances and pills that help me and improve my insane condition?"

However, even as a normal person it takes a long, long time of training and thinking to reach a point where you can ask yourself such questions. I'm not even saying that if I would be a psychopath, that I would be a better person than other psychopaths. Because I can understand now how difficult it is to overcome certain diseases - of body and mind - and how difficult it is to overcome the hatred of other, normal and feeling people. So, I'm just saying: "Use your time and money for the advantage of yourself AND other people, not for the disadvantage." Anything else is just short-sighted, stupid and pure (self-)destruction. --81.6.59.42 (talk) 13:43, 15 September 2015 (UTC)

Sigh... I suspect that this may be yet another case of confusion caused by the fact that different academic disciplines use the term "Primary source" in different ways... It is important to understand that this policy was primarily written by editors familiar with the terminology as used in the Humanities, and not by editors familiar with the terminology as used in the Sciences. In the Humanities, the type of source under discussion (peer reviewed journal articles) are not considered Primary sources at all... they are by definition secondary sources.

To understand the intent of this policy, you have to understand which definition is being used. Blueboar (talk) 15:43, 15 September 2015 (UTC)

It's true, there is a misunderstanding. I'll carefully re-read the guidelines again but as a first impression I must say that primary sources as a concept is not entirely clear here. For me primary source is the first (for Wikipedia, purpose, published) work that stated something. 😊 But this is not the case here if I'm not mistaken? I come from linguistics, and this is the meaning we intend for "primary source". I think what is meant here is "primary data" in general scientific terminology, am I wrong? --SynConlanger (talk) 21:00, 15 September 2015 (UTC)

It's unclear. Certain things are certainly primary sources, such as experimental data, historical documents, legal proceedings, statutes, legislative transcripts, etc. Other things are almost certainly secondary sources, such as literature reviews and meta-analyses. It can be less clear if a publication analyzes primary sources, but makes completely novel claims that appear nowhere else. Sources may certainly be a bit of both: a scientific paper, which I would normally refer to as a primary source, typically contains both experimental data as well as analysis of that data, and a brief introduction to the subject citing other primary and secondary sources. So it certainly has aspects of both a primary and a secondary source. A source can even be primary in some contexts and secondary in others - if we have an article about a book, for example, then that book would be treated as a primary source on its own article, no matter what the book would be treated as elsewhere. And finally, primary/secondary/tertiary aside, we are always left with the question of significance. If a viewpoint appears in precisely one secondary source, it may be deemed insignificant and left out of Wikipedia, much the same as if it had only appeared in primary sources, regardless of its verifiability. Someguy1221 (talk) 22:05, 15 September 2015 (UTC)

In the field of history, an old manuscript would be considered Primary (the manuscript is the historical equivalent of raw data in the sciences)... a published journal article that analyzes and interprets that manuscript and states conclusions about it would be considered Secondary (even if the conclusions are novel). When you understand that distinction, I think the policy is much clearer. Blueboar (talk) 00:42, 16 September 2015 (UTC)

Right, the definition is slightly distinct in my own field of Biology (see [4]). The distinction may be caused by the fact that the people who write the article analyzing the primary data are the ones generating the primary data in the first place. But our IP friend is speaking in the context of a medical article to begin with. Someguy1221 (talk) 01:19, 16 September 2015 (UTC)

So if the IP cited primary sources in the science sense that give raw data and analyse them, and the analysis given in the Wikipedia article is the one from the articles, that should be ok, right? --SynConlanger (talk) 09:20, 16 September 2015 (UTC)

Correct... as far as WP:NOR is concerned. Our sources can engage in original research, analyze data, and form novel conclusions (in fact, we rely on them to do so)... NOR says that we (Wikipedia's editors) can not. In other words... We can not analyse the data for ourselves... but if an analysis or conclusion is contained within a source, it is not Original Research for us to mention that analysis in an article... as it did not originate on Wikipedia.

That said, remember that passing WP:NOR is not the be-all-and-end-all of inclusion... The information also has to pass all of our other policies and guidelines... If an analysis is published in a journal with a poor reputation, we might reject it on reliability grounds. And if the analysis of data contained in a source is too novel (and challenges accepted wisdom) it may be considered fringe (in which case, mentioning it at all would give it UNUDE weight). And finally... novel conclusions (even those published in reliable journals) are best treated as opinion, and not accepted fact... and opinion should always be attributed. Blueboar (talk) 12:28, 16 September 2015 (UTC)

Not only can we not "engage in original research, analyze data, and form novel conclusions", we cannot line up a bunch of research papers to imply a conclusion that is not explicit in any one of the sources. -- TRPoD aka The Red Pen of Doom 17:40, 5 October 2015 (UTC)

Proposal to add a note to tertiary sources section

The section on tertiary sources currently reads as follows: "Tertiary sources are publications such as encyclopedias and other compendia that summarize primary and secondary sources. Wikipedia is a tertiary source.[8] Many introductory undergraduate-level textbooks are regarded as tertiary sources because they sum up multiple secondary sources." I think that this excerpt may cause confusion to new editors because they may assume that since Wikipedia is a tertiary source, then they can cite one Wikipedia article as a source for a contribution to a second Wikipedia article. To prevent this confusion, I propose the following addition (emphasis added in bold, just for ease of reading here on the Talk page):"Tertiary sources are publications such as encyclopedias and other compendia that summarize primary and secondary sources. Wikipedia is a tertiary source.[8](Note that one Wikipedia article cannot be cited as a source for a contribution to another Wikipedia article) Many introductory undergraduate-level textbooks are regarded as tertiary sources because they sum up multiple secondary sources." OnBeyondZebrax • TALK 20:19, 25 August 2015 (UTC)

Except that there are a few situations where we can cite one Wikipedia article in another - specifically in articles where Wikipedia itself is the article topic. Of course (just to confuse things) in such articles the cited Wikipedia article would be considered a Primary source for the information, not a Tertiary source. Yet another example of how the definitions are flawed. Blueboar (talk) 00:29, 26 August 2015 (UTC)

It already has footnote [8], which says essentially the same thing: "While it is a tertiary source, Wikipedia is not considered a reliable source for Wikipedia articles." – Margin1522 (talk) 09:32, 26 August 2015 (UTC)

Yeah, doesn't seem to be a flaw in the definitions. WP writing about WP is an entity writing about itself. That's always a primary source, no matter what it is. If the New York Times (which mostly consists, aside from primary op-eds and movie/book reviews, of most secondary source material) publishes an overview of it's own history, that's a primary source, period. That is, the "primaryness" trumps the "teriariness" or "secondariness" of whatever the source might be otherwise. This holds generally for primary sources, not just autobiographical ones. For example, Geoffrey of Monmouth's Historia Regum Britanniae was essentially a secondary source when it was written, synthesizing Gildas, Bede, Nennius, etc., but it's so old (i.e., to our our eyes it is so close to the events on which it reports) that it's become a primary source, so we can't cite it as a secondary one even for points where it's really, really clear that it's a synthesis of previous work (the very nature of how to do that kind of work has changed since Geoffrey's time). There's no reason for confusion about these matters if one remembers that reliability, verifiability, and avoidance of original research are the goals: If something about a source introduces doubt about it, go with the doubt, don't try to hide the doubt.

Anyway, If it were seen as necessary to clarify these points, OnBeyondZebrax's note would could be merged with footnote 8, and Blueboar's point, inline in the article or in a revised footnote, as something like: "Note that Wikipedia (and reuses of Wikipedia content) cannot be cited as such a tertiary source in our own articles, though article content about Wikipedia itself may cite a Wikipedia article, policy page, or other material as a primary source about the project." — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 21:36, 26 August 2015 (UTC)

If uncited information in one article is supported by citations in another article that is wikilinked from the first, verifiability is satisfied. It would be better to have citations in both articles, but WP:PRESERVE and WP:NORUSH would apply. Rhoark (talk) 13:27, 30 August 2015 (UTC)

Re: "If uncited information in one article is supported by citations in another article that is wikilinked from the first, verifiability is satisfied." Actually no, it's not... I don't remember if it written down anywhere, but the "it's cited at the linked article" issue has been discussed many times before (I know it has been discussed in the context of WP:SAL) and there is strong consensus that a citation in one article does not satisfy WP:V in another article... citation in both articles is required. This is not an onerous requirement... simply copy the citation from one article and paste it into the other. Blueboar (talk) 14:13, 5 September 2015 (UTC)

"In Wikipedia, verifiability means that anyone using the encyclopedia can check that the information comes from a reliable source." At the most fundamental level that does not mean a citation on the same page, or necessarily any page at all. However, material that is "challenged or likely to be challenged" needs an inline citation, presumably on the same page. However in the case that its on a linked page, I reiterate WP:NORUSH and WP:COMMONSENSE for good measure. Rhoark (talk) 01:35, 11 October 2015 (UTC)

This page appears blank on my browser

What is going on?--Tomwsulcer (talk) 18:20, 14 October 2015 (UTC)

I had pages like that at the same time as you; it's gone now – a server bug?? Peter coxhead (talk) 18:28, 14 October 2015 (UTC)

Yes, gone now -- maybe it was a server bug. I had been considering reverting an edit.--Tomwsulcer (talk) 18:30, 14 October 2015 (UTC)

I disagree with the "No original research" policy

I doubt that I'll convince anyone, but I think this policy impoverishes Wikipedia, rather than making it better and more reliable. It leads to ridiculous situations in which, for example, the author of a book cannot say what is in his or her own book; some third party has to say it.

Original research is anything but unverifiable. If I wrote, based on my own research, "combining compound A with compound B at a temperature of X degrees produces resulting compound C", that is something someone could definitely verify. Or if I read the correspondence of some person and wrote "X frequently discusses Y in his/her correspondence", and said what correspondence I read and where I read it, that is also verifiable. It's not the same as looking it up in a published article, but it's still verifiable.

There is no doubt the fear that if this policy were changed Wikpedia would be inundated with garbage. I don't know what the best solution would be, but I want to state my dissatisfaction with the current policy. deisenbe (talk) 04:13, 11 October 2015 (UTC)

This is not something special to Wikipedia, that is just how encyclopedias work. We are a tertiary source. HighInBC 04:17, 11 October 2015 (UTC)

That's not true at all. Encyclopedia articles are written by experts and frequently contain their own research, or their own perspective. Then there is the real question if whether Wikipedia is in fact still an encyclopedia, or should conform to practices of other encyclopedias, or whether it should be a tertiary source and nothing more. deisenbe (talk) 04:26, 11 October 2015 (UTC)

Besides that discussion, original research should be of sufficient quality. Wikipedia has no quality system in place to review original research at anything close to the level expertise that would be required, nor have experts the authority to remove substandard original research against majority opinion. The generally accepted and enforced original research policy guarantees at least some quality. Unless we are willing to identify experts (and who would that be? Requiring at least a PhD on topic, or even a facultry position, and how would we get sufficient of those) and give those extra powers (and how would that look like). Without such quality system in place we basically open ourselves up for fringe theories and outright nonsense if we allow original research. Therefore we should accept that the current system is sub-optimal but that the solution would be worse than the problem.

Your complaint that Wikipedia has been taken up by others, who have created alternatives like Scholarpedia or Citizendium, both of which only host a few thousand articles and are struggling to keep editorship active. Arnoutf (talk) 07:46, 11 October 2015 (UTC)

@Deisenbe: I agree with you that, contrary to the view expressed by HighInBC, traditional encyclopedias do frequently contain expert views that would here be called "original research". However, this isn't a traditional encyclopedia. The key point is that anyone can edit Wikipedia, with the result that expertise in the topic concerned is frequently (in some areas almost always) lacking. Arnoutf is right: the present policy is the best we can do given Wikipedia's structure.

However, it is important that the policy is applied sensibly. I wish that WP:NOTOR were also policy, an not just an essay. Peter coxhead (talk) 08:16, 11 October 2015 (UTC)

I stand corrected. HighInBC 14:26, 11 October 2015 (UTC)

Thanks for the support and for the reference to WP:NOTOR, which I was unaware of. As a practical matter, WP is chockablock full of OR, if you get off the heavily scrutinized articles. And I want to go on record as saying that within those limited areas in which I have personal expertise, I see the policy doing a lot of harm, and much less good. And I'm not talking about things I've written myself. deisenbe (talk) 10:40, 11 October 2015 (UTC)

I do see the problem. Even in heavily scrutinized articles the OR-pushers who manage to find allies are often happier to edit-war than concede the point; while on the other hand important text edits or paraphrases are frequently butchered incorrectly invoking the OR rules. Nevertheless, imperfect as Wikipedia is, I still think that allowing original research in would make the project worse rather than better. Arnoutf (talk) 11:40, 11 October 2015 (UTC)

Deisenbe, WP:Original research does not simply mean "unsourced." See Wikipedia talk:Verifiability/Archive 63#How wide is the "original research" exception for articles on fictional works?, which is very clear about that. So if you wrote, based on your own research, "combining compound A with compound B at a temperature of X degrees produces resulting compound C", and that is supported by a WP:Reliable source somewhere out there, even if not on Wikipedia, that is not WP:Original research. The WP:Original research policy has a note/reference right after the word exist in its introduction; that note/reference states, "By 'exists', the community means that the reliable source must have been published and still exist—somewhere in the world, in any language, whether or not it is reachable online—even if no source is currently named in the article. Articles that currently name zero references of any type may be fully compliant with this policy—so long as there is a reasonable expectation that every bit of material is supported by a published, reliable source."

Furthermore, you already addressed your disagreement with the WP:Original research policy earlier this year, and got your answers then; see Wikipedia talk:No original research/Archive 60#No OR is a bad policy, in my view.. In fact, that was the last time I commented at this policy talk page...until now. Flyer22 (talk) 20:21, 11 October 2015 (UTC)

User:Flyer22: Your memory is much better than mine. Congratulations! As it is a WP principle to "assume the best", I'm going to assume that you're not intending to publicly embarrass and chastise me yet again.

You don't need to restate the policy to me. I know what the policy is, though you no doubt know its details much better than I. My point was that I disagree with the policy. But I'm just one lowly editor. deisenbe (talk) 23:20, 11 October 2015 (UTC)

Deisenbe, pointing out Wikipedia policies and guidelines to you and expecting you to follow them, especially after you've been repeatedly pointed to them either on your talk page or elsewhere on Wikipedia, and getting annoyed/frustrated when you repeatedly ignore/disregard them, has never been about trying to embarrass you. I understand that you feel that I've been condescending to you, but we've already been over that on my talk page; I see no need to rehash it. Flyer22 (talk) 04:27, 12 October 2015 (UTC)

Your disagreement is duly noted, but since you are suggesting nothing newer and better, I don't see the point in this discussion beyond personal blogging. Staszek Lem (talk) 17:12, 12 October 2015 (UTC)

I agree, let's either talk relevantly about content and not contributors or stop this. The only thing I learned is, that although I had forgotten the previous discussion flagged up by Flyer22, my comments there are remarkedly similar to the ones here. So I do have a consistent opinion on this :-) Arnoutf (talk) 17:18, 12 October 2015 (UTC)

I agree 100% with the original comment at the top above and have written the following on a talk page:

"I think that this issue highlights one of the problems of Wikipedia. There is (thus far) an almost robotic adherence to the policy of referencing published resources. However, the publication of statements in magazine or book form does not in itself constitute evidence of reliability or accuracy. Far from it !

Ultimately the most reliable sources of information for matters of a historical nature are original, contemporary documents. There is no real substitute for these. Even then, there are times when, for political or selfish reasons, individuals may lie, but in this case, the context of the document and the date is evidence enough of reliability.

I think, indeed I hope that it will in time be recognised that original research, backed up (as in this case) by references that can easily be checked, is not only acceptable, but extremely valuable in improving the reputation of Wikipedia as an information resource.

I am very disappointed in general with Wikipedians and I find posting on Wikipedia - despite the fact that it is supposedly a place where any interested parties are free to edit articles - an extremely frustrating and generally very depressing experience."

I emphasise strongly again:

"the publication of statements in magazine or book form does not in itself constitute evidence of reliability or accuracy"

The quality and reliability of published material is extremely variable. It depends on many factors. The age of the material for one. Can a book first published in 1700 be a reliable source just because it is printed? And who should be the arbiter of whether it is a reliable source?

Are magazine articles a reliable source of information? Again, who should be the judge of that?

In matters of a historical nature, original research, especially research carried out by private individuals can be of great value. But all the same it must be referenced clearly. For example, a particular piece of research may highlight a previously unknown fact about a historical figure. (Probably a minor historical figure). If that research can be verified by a clear reference to a document that is held in a public place (such as an archive or library) and can easily be either read or digitally copied for verification by anyone then it is just as acceptable as a published source! Indeed it may be more so.

A published source is reliable - why? A published source may be very old. It may have been published privately. The appearance of a piece of research in book form does not make it reliable. There are not bunches of Wikipedia Editors installed in publishing houses making sure everything is done properly! There seems to be an assumption that editors in publishing houses are paragons of virtue. Custodians of absolute truth and accuracy. Well, are they? I doubt it!

I found a recent case of where the opinion (a controversial opinion) of an individual who had written an article in a magazine was being used in a Wikipedia article. That is not acceptable. Opinion is not fact. Its was publication in a magazine does not make that statement reliable or truth.

There is much more I could say. I hope that my comments will at least be read and noted. John2o2o2o (talk) 09:28, 28 October 2015 (UTC)

@John2o2o2o: you are actually making two points, I think.

Material which has been published but is wrong should not be included. On this we can agree entirely. You can challenge an unreliable source, based on reliable sources, and remove such material (or include text to the effect that although X has been claimed to be true^{ref bad source}, it is not^{ref good sources}).
It should be possible to include material based on original research, provided this is sourced. I'm not sure how you are interpreting "original research". If you mean by "research" searching though sources to find factual information, like someone's date of birth, which can then be attributed to a source, then there's usually no problem in adding it to an article. But this isn't really "research". The prohibition on "original research" is directed at original analysis, interpretation, evaluation, and the like. It's this that is rightly prohibited in Wikipedia, because there's no control on the expertise of the editor carrying out the analysis, interpretation, evaluation, etc.

Peter coxhead (talk) 12:39, 28 October 2015 (UTC)

RfC announce: What does Wikipedia:Identifying reliable sources (medicine) cover?

There is a request for comments at [ Wikipedia talk:Identifying reliable sources (medicine)#What does MEDRS cover? ].

At issue is whether the lead paragraph OF WP:MEDRS should remain...

"Wikipedia's articles are not medical advice, but are a widely used source of health information. For this reason it is vital that any biomedical information is based on reliable, third-party, published secondary sources and that it accurately reflects current knowledge."

...or whether it should be changed to...

"Wikipedia's articles are not medical advice, but are a widely used source of health information. For this reason it is vital that any biomedical and health information is based on reliable, third-party, published secondary sources and that it accurately reflects current knowledge."

This has the potential to change the sourcing policy from WP:RS to WP:MEDRS on a large number of Wikipedia pages, so please help us to arrive at a consensus on this issue. --Guy Macon (talk) 06:09, 3 November 2015 (UTC)

Maybe I'm being dumb here, but better to ask and be the wiser

Am I correct in assuming that a working scientist in some field or other, who has published peer-revieved articles in respected journals or books in that field, may quote them? That is original research, but has gone through the academic vetting process. Cheers Sorte Slyngel (talk) 20:24, 9 December 2015 (UTC)

Original Research refers to anything that an editor includes in an article that can not be backed by any published work. It is okay to quote a working scientist provided that their quote comes from a published source and is not, say, a quote that you wrote down in a private interview.Moira98 (talk) 03:01, 24 December 2015 (UTC)

To be clear our OR rule states that people editing Wikipedia can't make unsourced analysis. Linking to a reliable source coving a scientific study without going beyond what the soure covered does not violate this rule since it is not a Wikipedia editor committing original research.--67.68.23.129 (talk) 06:09, 26 December 2015 (UTC)

Removing citations to "apparently predatory publishers"

Moved to Wikipedia:Reliable sources/Noticeboard § Scientific Research Publishing

unintentional forumshopping Jytdog (talk) 17:04, 3 March 2016 (UTC)

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Please weigh in: Talk:Predatory open access publishing#Removing citations to "apparently predatory publishers". fgnievinski (talk) 16:16, 3 March 2016 (UTC)

Fgnievinski I've followed behind you at WT:MEDRS and at JzG's Talk page and at that article Talk page. That section is not an appropriate use of an article Talk page and I closed it. The issue has been under discussion at RSN since Feb 26 here. Jytdog (talk) 17:04, 3 March 2016 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Sections for discussing types of OR to serve as cautionary examples

Brought this up at Wikipedia_talk:Editors_are_not_mindreaders#Use of short-cut but I figure it's more related to the purpose of this policy... if we observe other editors engaging in what looks like OR by inserting unsourced claims, it would be very useful to be able to have some kind of shortcut to link to explain their particular problem to them.

Like not that it's the broad concept of OR, but a specific kind of OR so that they can better identify their edit with examples which more closely relate.

Like one problem I experience is one editor keeps claiming a fictional yellow ogre "likes to hug". It seems simple enough. If I had to guess at the character's motives I would think that too. He often initiates hugs with others with a big smile, after all. But it's still OR to assume that he likes what he's doing. To make that claim and not have it OR I think we would need a reliable source (like maybe a char description via a distributor/creator) which explains that he likes hugging. Those would be experts, but simple show-watchers would not be experts.

Basically if it's a type of OR where someone makes unsourced assumptions about how a character feels, what they like, I want to call it WP:EMPATHY. Whereas if it's a type of OR here someone makes unsourced assumptions about what a character's opinion is, what they are thinking about, that would be WP:TELEPATHY.

We already have something like this in WP:BALL (aka CRYSTAL/FUTURE) where users are warned against believing they can predict the future. This relates to cognitive distortion#Jumping to conclusions in this case ball/crystal/future relates to the second one, Fortune-telling, while what I am proposing is a section discussing the first problem, which is Mind-reading. Both are things Wikipedians should avoid doing so both should be warned against.

I believe the WP:RUMOUR (aka SPECULATION) tag could be appropriate to represent a larger section, seeing as how warning against rumour-mongering/speculating warngs against more ills than making guesses about the future, but also guesses about the past/present. In reation to the above CD concepts, I think these two tags should represent the broader taboo of JTC (jumping to conclusions) while BCF represets fortune-telling and WP:EMPA/WP:TELE could point to a new section warning against mind-reading thoughts or emotions.

In this case I am not sure if it is "what Wikipedia is not" or "what Wikipedians are not". The use of BALL/CRYSTAL avoids personifying Wikipedia by saying it is not the TOOL of a fortune-teller. (ie it is not saying Wikipedian is not a seer, but rather it is not the tool of seerS). I do not know any tools of an empath/telepath to substitute here though. At least not one easily identifiable. Even if the focus is on what we ought not to do I think we could still fit it into the WWIN page. 184.145.18.50 (talk) 13:56, 9 February 2016 (UTC)

This isn't a big enough deal that it would need to be covered in NOR policy. Policies are general, and are not meant to drill down in detail to every conceivable example. If what you're talking about is a common case (I believe it is), it would be better addressed at a topical guideline. I would suggest raising the idea at WT:MOSFIC (and more cogently; it does not help your argument that you are self-doubting yourself throughout it and pointing out why it's not as a good an analogy as you thought it was). It's also not going to help that you're trying to propose multiple shortcuts and are splitting hairs subjectively. It's practical to have a simple proposal to add (at perhaps MOS:EMPATHY in MOS:FIC) a warning against the tendency to imagine and impute to fictional characters various feelings, motivations, beliefs, or intents that are not actually explicit in the work. That could literally be the text of the proposal, perhaps with some examples, that range from obvious to subtle. (And "telepathy" doesn't mean that; it means the alleged ability to mentally communicate over a distance). — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 09:38, 4 March 2016 (UTC)

Primary, secondary and tertiary sources

"Wikipedia articles should be based on reliable, published secondary sources and, to a lesser extent, on tertiary sources and primary sources. Secondary or tertiary sources are needed to establish the topic's notability and to avoid novel interpretations of primary sources."

It would appear that under this policy a piece of scientific research, published in a peer-reviewed journal by an expert in the field, is to be less relied upon than secondary sources? Someone who has produced the latest, break-through discovery is overridden by someone else's musing in a non-peer reviewed book that might follow up several years later? Wikipedia is behind the times in many of its science-related topics; we have non-science background editors overriding the latest research due to a call to some populist ideas that were published a decade earlier in a book (some bordering on novels). I do not seek a response to this statement as it has been raised before - and very poorly addressed from what I have read in the archives - however Wikipedia editors that have put together this policy need to think about its long-term impacts on the encyclopedia. (NB: I am not a scientist and am not pushing any particular point of view, but I have noticed the impact of this policy on a number of Talk pages of the science-related Wikipedia subjects - those subjects remain about a decade behind what we know today. Users expect an encyclopedia to be up to date else they begin to rely on other, more current sources for their information (e.g. press releases), with all of its inherent issues.) Regards, William Harris • talk • 20:18, 26 December 2015 (UTC)

I am a scientist, and I do support the statement as it is. An encyclopedia should present mainstream consensus views. A breakthrough discovery will, if it is indeed the claimed breakthrough, very rapidly become mainstream. However, in sciences(as in many disciplines) mistakes are sometimes made, and "lucky" finds are often published. Therefore we should be extremely careful with a single paper showing an (experimental) effect.

However, we do not rely only on non-peer reviewed books. We generally also consider the literature review in peer review scientific articles, review and meta-analysis peer reviewed papers as secondart literature. In fact, I think that many of the books you are referring to would be tertiary sources, rather than secondary sources - so not a preferred source either.

I do realise that the borders between all of this are rather thin and hard to understand, and most of the discussions about this have been about these kinds of borders. Arnoutf (talk) 21:07, 26 December 2015 (UTC)

"Under this policy a piece of scientific research, published in a peer-reviewed journal by an expert in the field, is to be less relied upon than secondary sources" is precisely what it means, and is intentional. It's a perennial "Wikipedia adjustment" issue for academics and professional researchers (and even science grad students) when they first encounter this. Especially in the sciences, academic journals are mostly sought by specialists in the field for the hot new research and ideas, with literature reviews seen as sort of dull "catch-up homework", or dismissed as people who are not cutting-edge themselves just summarizing the real work of the eminent. Academic publication, within academe, is tied closely to credit, reputation, and advancement.

For encyclopedic purposes, primary research papers are low-quality sources for facts, because what they are presenting is untested, unverified, unrepeated hypotheses, experiments/fieldwork, data, and conclusions. They've only been "sanity checked" to the extent that a peer review committee thought them interesting enough and their methodology sound enough to be worth presenting to their readers. Depending on the journal, a paper might have been been selected because it is likely to generate controversy and in-the-field "buzz", not because it is felt certain to be correct.

WP relies on secondary (and sometimes tertiary) sources because they reflect real-world review, acceptance, and integration of what the primary material has presented, generally in combination with other research. That is the analysis, evaluation, interpretation, and synthesis that Wikipedians can't do on our own with this material, or rely upon primary sources for, as a matter of "no original research" policy.

Obviously, peer-reviewed secondary sources like literature and systematic reviews are of more value for science that lower-quality secondary sources like newspapers or monographic books by a single author or team, as Arnoutf touches on, above. "Secondary" and "primary" aren't tied to the genre of the publication, but relate to whether it's presenting something new or something arrived at by review of multiple sources. Whether sources (primary or secondary) are of high value or not is matter of author/publisher reputability, and level of editorial review. It may be helpful to review WP:PSTS for WP's internal definitions of primary, secondary and tertiary for purposes of our content policies and sourcing guidelines. These terms can have radically different off-Wikipedia meanings in different fields, and it's unworkable to try to apply one particular field's internal definitions of them to Wikipedia.

Tertiary publications like encyclopedias, including Wikipedia, are always "behind the times" of the bleeding-edge research, because it often turns out to be wrong, and such publications are not in a position to say ourselves which new ideas and conclusions will prove to be correct according to broader scientific consensus. It's just the nature of the publication, its editors, and its purpose.

PS: I have to write pretty much the same explanation of these matters anywhere from 1 to 10 times per month in various places, and I know other editors do, too. We should probably try to improve the relevant pages to make these points (more succinctly than I can off the cuff like this) to forestall these misunderstandings and necessity to keep explaining them in the same terms.) — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 01:27, 4 March 2016 (UTC)

... but there's a lot of scope for editorial judgment in all this, and a lot of argument about secondary sources is really about notability. There's a tension between notability and reliability here. If, for example, a scientist publishes a piece of bleeding-edge research in the scientific literature, and the BBC then report on it, then the fact that the BBC has reported it is strong evidence that the research is important and needs its own page. But the content of that page should be checked carefully against the published research, because we would need our page to reflect the actual research, not the BBC journalist's misunderstanding of that research. In terms of notability the journalist's article is key but in the context of WP:NOR, the scientist's expertise in their own field should trump the journalist's report 100% of the time.—S Marshall T/C 08:47, 4 March 2016 (UTC)

Notability: The requirement that a topic have significant coverage in multiple, independent, reliable, secondary sources – notability – is tangential to everything in WP:NOR about secondary sources, other than this is the policy that defines what they are. And the BBC by itself reporting something does not lend it notability (note the "multiple" component that's required – it's the confluence of multiple RS in providing similar coverage that confers notability). Just in case, it may be worth addressing some other common "notability"+"sources" areas of confusion: A newspaper having a larger circulation doesn't automatically make it more notable; that's often a factor of population, target market, advertising, etc. The reputability of a source is a combination of factors, and circulation isn't a major one; how often it's cited by others is a much stronger indication, as is what other RS say explicitly about that writer/publisher/periodical. Many of the world's worst tabloids have enormous circulations and are very notable (mostly in a negative way).

The research: Oh, I agree. But as a matter of policy, all WP:AIES material must come from secondary sources. The obvious solution is you cite the secondary sources for the claims/conclusions, to satisfy policy, and for the general readership to look at, and you cite the original primary research papers for details in the data, for specialist readers to look at. This is how most of our good science article are already constructed, out of necessity. We don't even have to have a guideline saying to do it this way, it just evolves/emerges automatically. I already address the scientists vs. journos matter, and someone else did it before I did, so I'm not certain why that's still your concern; to wit: "Obviously, peer-reviewed secondary sources like literature and systematic reviews are of more value for science that lower-quality secondary sources like newspapers or monographic books by a single author or team, as Arnoutf touches on, above.". I kind of looks like you are saying "whatever a virologist ever says about virology must always be trusted over what a news source says about, even if the scientist's paper is drawing fire because of methodological errors and the irreproducibility of its results, the scientists claims actually came from their personal blog anyway, and the journalism is based entirely on top-flight systematic reviews, and was put out by a publisher with an unusually high reputation for fact-checking with experts"; but surely that can't be what you mean. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 09:25, 4 March 2016 (UTC)

@SMcCandlish: the key reason this has to be explained over and over again lies in what you wrote above: "Secondary" and "primary" aren't tied to the genre of the publication, but relate to whether it's presenting something new or something arrived at by review of multiple sources. This means, rightly, that it's not possible to decide whether something is "primary" or "secondary" based on a simplistic rule, such as the medium in which it was published, but rather that editors have to make decisions in relation to the part of the article that the relevant part of the source is being used to support. So cases have to be taken on their merits and discussed as such, and this will continue to be the case. Proper revision of WP:PSTS should result in an increase in the need for discussion and consensus, not a reduction.

Take one piece from what you wrote above: For encyclopedic purposes, primary research papers are low-quality sources for facts, because what they are presenting is untested, unverified, unrepeated hypotheses, experiments/fieldwork, data, and conclusions. This is actually circular logic, given that you don't equate "research papers" with "primary", since what makes a "research" journal article "primary" is its being used to support reporting in Wikipedia on new and hence unreplicated "hypotheses, experiments/fieldwork, data, and conclusions". To decide that a particular use isn't justified in a Wikipedia article, you have to check carefully both the material it is supporting and the quality of the relevant part of the source. If this can be explained better, great; but don't expect it to diminish the need to keep on repeating the explanation in particular cases. Peter coxhead (talk) 08:48, 4 March 2016 (UTC)

I wrote unclearly, but thought it was obvious that I was referring to the untested, unverified hypotheses, data, and conclusions in primary research papers (mentioned in same sentence), not any secondary intro material they might contain. I was engaging in some shorthand here, in response to the OP, not drafting new guideline language. I agree with your premise that "Proper revision of WP:PSTS should result in an increase in the need for discussion and consensus, not a reduction." It just shouldn't be this exact same thing over and over again, this confusion of "well in my field, here's how we do it and here's what we think a secondary source is and what value it has" with how WP operates including in the topic of that person's field. It's so, so tiresome to have to keep covering this, month after month, year after year. The very reason we have guidelines is so we don't have to do that. If we do not address it in WP:NOR and WP:RS themselves there will never, ever be a day when this situation improves, because new editors with the same preconceived notions will always continue to arrive. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 09:02, 4 March 2016 (UTC)

I'm sure that fundamentally we don't disagree, and I won't want to needlessly prolong this thread, but some primary "data" is acceptable in some contexts. For example, when a new species is named by a taxonomist, a diagnosis must be attached for the naming to have effect. Facts in that diagnosis are reliable by definition: "the new species is differentiated from all others in the genus by the possession of a retrolateral tibial apophysis" in the original publication supports "the species has a retrolateral tibial apophysis" in a WP article. However, a secondary source is essential to demonstrate the acceptance of the species by taxonomists generally. This is well-explained in the project pages of various Tree of Life descendant projects, but doesn't prevent the need to keep re-explaining it over and over again to new editors.

Perhaps I've become weary and/or cynical (which I observe to be a common consequence of spending too much time on Wikipedia), but I doubt the wisdom of trying to achieve ever tighter language in relation to complex issues like this one (and I think it's essentially the same issue as in the recent MEDRS discussion). When an issue is subtle, complex and subjective and needs careful explanation, the resulting text will be greeted with "tl;dr" (just think how often the length of your posts is regarded as self-evidently bad) or "I don't understand it so it must be wrong". Even those editors that do read such text will misinterpret it to suit their own ends. Whatever is written down in WP:NOR and WP:RS, new editors with preconceived notions will always continue to arrive, and unlike us ( :-) ), many of their notions will be wrong. Peter coxhead (talk) 09:37, 4 March 2016 (UTC)

Meanwhile, the advanced science and mathematics parts of Wikipedia are excellent and getting better precisely because of our highly qualified editors who are able to understand the latest research and summarise it for us mortals. If they had to rely on secondary sources for all their material, a large part of it would become useless and in the case of mathematics a lot of important information would disappear completely. Zero^talk 00:37, 5 March 2016 (UTC)

Also note from WP:V, no less a policy page than this one: "If available, academic and peer-reviewed publications are usually the most reliable sources, such as in history, medicine, and science." This seems to be in direct contradiction to the claim here that "Under this policy a piece of scientific research, published in a peer-reviewed journal by an expert in the field, is to be less relied upon than secondary sources". The rules on this page were written in an attempt to define and exclude Original Research (which means original on the part of the editor). They were never intended to prevent editors from seeking out the most reliable sources, as specified by WP:V and WP:RS. Note that we also need to observe WP:NPOV, which pretty much covers the gap. If a newspaper article contradicts a paper in Nature without explaining why, we can just ignore the newspaper. However if the newspaper article documents a controversy over the Nature article, quotes other scientists as disagreeing, etc, then we should include that material as well. In all cases, the idea is to present readers with an accurate and balanced summary of what all the experts are saying. Zero^talk 02:20, 5 March 2016 (UTC)

Exactly so, and well put. It's necessary to make balanced judgements, using sources that are most reliable and appropriate for the material being supported by them. No simplistic rule like "journals bad, encyclopedias good" can replace editor judgement and consensus within the over-riding requirements, such as WP:NPOV. Peter coxhead (talk) 09:53, 5 March 2016 (UTC)

Implied conclusions

If one source claims the first queen of England was Mary I and is on the Mary I page, and another source claims it is Elizabeth II on the Elizabeth II page, would it count as implying a new conclusion to add the Mary I source on the Elizabeth II page with "Mary I preceded Elizabeth II as queen, and therefore Elizabeth II was not the first" because the Mary I source does not specifically say "and therefore Queen Elizabeth II was not the first queen of England"? That would seem like stretching the spirit of this caveat to me, but a pedantic reading would discount using the sources in this way. Yb2 (talk) 20:55, 9 March 2016 (UTC)

That would indeed be original research. Such combinations are explicitly addressed in the section wp:synth. (PS I guess you mean Elizabeth I not the current queen) Arnoutf (talk) 17:40, 10 March 2016 (UTC)

Especially since there are yet other sources that claim that the first Queen of England was Matilda. Blueboar (talk) 19:06, 10 March 2016 (UTC)

Open science

Is referring to open science considered as original research? --PJ Geest (talk) 11:02, 27 March 2016 (UTC)

WP:PRIMARY is abused and out of control

In the past few days I have been running across editors who arrive on a page and seek to delete content based solely upon WP:PRIMARY. They will state that the New York Times or wsj is a primary source, etc. Then they will use this justification to delete sections, or call an overall article into question.

Example: An editor sought to section blank https://en.wikipedia.org/wiki/Ethereum#Implementations because he/she said it wasn't properly cited. Then a revert war started and the page ended up marked up with citation problems for non-primary sources, etc. A healthy discussion started on the talk page https://en.wikipedia.org/wiki/Talk:Ethereum#Sources and eventually the matter was closed.

Is the standard that everything that is included on wikipedia to be the secondary sources, only allowing cases when the NYT reviews an article written in the WSJ?... is the bar we are going to hold for inclusion of content? I argue this is taking it too far. I understand and agree with the premise of sources, but it has gotten out of control in the past couple of years. Yes, it is fully needed in WP:MEDRS and in cases of debunking. Maybe the editors and admins can think of adding some more clear guidance to primary sources, and clearly state that third party reporting by large news organizations is different from a witness who reports seeing something on the street. A strict reading on WP:PRIMARY is detrimental to the overall purpose of wikipedia to accurately describe everything， because everything would never have secondary sources. Jtbobwaysf (talk) 05:05, 9 April 2016 (UTC)

So if the NY Times reports that the temperature in New York City reached 64 degrees on April 1 that can't be cited in a Wikipedia article until some other source corroborates the value? OK, so it's not the NY Times, it's a weather web service. Same thing: what they provide is always primary material. The same holds for a primary scientific article. There is always reason for a degree of skepticism but that doesn't seem to demote the majority of scientific publication to being regarded as too unreliable to cite. If it is so unreliable there's a huge problem because within science such citation - without waiting for secondary source corroboration - is routine. (Is necessary, or the progress of knowledge would be much slower.) Well, sure, I can understand the concept that a few Wikipedia editors know better than the entire scientific community but golly: I don't think there's a single source of any category (primary, secondary, or tertiary) I can cite as evidence of that concept being valid.

I was taught that the experimental section of a journal article was likely to be accurate, the analysis not so much so. It seems that some insist that what is published as the result of research by someone isn't as reliable as a secondary source. I'm aware that every few years it comes out that an author intentionally deceived in describing an experiment and its results. That's an incredibly minute portion of what is published. The policy seems to say that primary sources describing experiments can't be cited but the problem is that if a second researcher does the same experiment and gets the same results that very often results in no publication, there being no reason to publish. It's only the cases in which the second researcher cannot duplicate the results that something is published. (I can't help suspecting that the policy is targeted at history and social science material, not physical science material, but as worded it hits all with the same force. That is unfortunate.)

I see objections now very much like the objections I saw seven years ago, when I participated more vigorously in this discussion. I predicted then that the same objections would continue to be raised. There is very ardent enforcement and insistence upon of a very badly constructed policy. There are reliable and unreliable primary sources, there are reliable and unreliable secondary sources. Even if some statistical statement can be made about primary and secondary sources as a whole it is a misuse of statistics to judge individual sources on the basis of broad categorization of all sources of that type unless a category is either 0% reliable or 100% reliable, which neither category is. The standard has to be care in selection and interpretation of sources (with interpretation being very limited, naturally) and a good dose of wisdom. I can foresee that this policy will never be shortened to "use wisdom" but it wouldn't be bad for that to be the general philosophy. Add detail and exposition as needed - but respect wisdom. Minasbeede (talk) 19:39, 9 April 2016 (UTC)

@Minasbeede: The policy does not say that primary sources can't be cited; quite the opposite: (1) Deciding whether primary, secondary or tertiary sources are appropriate in any given instance is a matter of good editorial judgment and common sense, and should be discussed on article talk pages. (2) ... primary sources that have been reputably published may be used in Wikipedia; but only with care. Secondary sources are required to judge notability and avoid synthetic interpretation of primary sources. But primary sources are certainly appropriate in many instances, as in the case of peer-reviewed scientific primary literature—and note again that, generally, scientific and other academic articles are typically secondary sources as well, depending on what claim you're citing, since they will almost always refer back to the extant literature. —Nizolan ^(talk) 02:46, 12 April 2016 (UTC)

I think the problem does not lie with the policy but, as Minasbeede states, the abuse of the policy to change article content. The problem is that too few Wikipedia editors show sufficient wisdom; but I am not sure that a change in policy can enforce wisdom. Arnoutf (talk) 11:57, 12 April 2016 (UTC)

I agree it can be abused, but yeah, that's not going to change by changing the policy. The policy as it stands is fine. It seemed to me that Minasbeede was taking the policy itself to sanction that kind of action, which it doesn't. —Nizolan ^(talk) 17:32, 12 April 2016 (UTC)

I note Jtbobwaysf inadvertently failed to notify anyone at Talk:Ethereum of the attempt here to influence the discussion there. After a blatant recruitment on Reddit, Ethereum has suffered a tremendous influx of editors who have a COI - either a strong financial position in ETH, or (in one case) an Ethereum Foundation board member editing while COI without notification, even after having the WMF TOS pointed out to them. So we have a financially-motivated group of advocates on the article, including a series of personal attacks on me on the talk page (which Jtbobwaysf has been participating in). So, more eyes on the article and the talk page would be most welcomed, particularly in relation to sourcing - David Gerard (talk) 21:59, 17 April 2016 (UTC)

Use of primary sources and secondary sources.

Just recently on the on-going article 2016 Stanley Cup playoffs, there has been an edit war between me and another editor on the use of primary sources being used to reference the general recaps of each individual series. The other editor's argument being that an article cannot have a majority of primary sources, but the policy (or guideline...?) says that an article should be based on secondary sources. Now this article is mostly fact-based, with the primary source being the most reliable in telling game recaps, but I do agree that this article should not be based entirely off the primary source.

My edits are the series recaps: New York-Pittsburgh, Tampa Bay-Detroit, Los Angeles-San Jose, etc. Of which I supply the game recap references to each individual game, with a paraphrased sentence that precedes the reference. This is not an interpretation of the events that happened, it is almost a direct word-for-word. The primary source being NHL.com.

Can an article be based with a majority of primary sources?Conyo14 (talk) 07:20, 24 April 2016 (UTC)

First, we use NHL.com a lot in our other Hockey playoff related articles (for example, see:2015 Stanley Cup playoffs) so I see no problem using it in a similar way in this one.

That said... NHL.com contains both primary statistics and secondary reporting on the games. Using the raw statistics to create a game recap would be OR... Citing the reporting to support a game recap is not. Blueboar (talk) 11:34, 24 April 2016 (UTC)

Gregorian centuries, routine calculations, and WP:SYNTH

This section is motivated by a disagreement at the List of oldest living people regarding the presentation of Emma Morano as the last known living person from the 1800s. The fact is cited, and gives every indication of being true. However, it is trivially easy to find sources that refer to her as being the last living person from the nineteenth century,[5] yet it would clearly be in violation of Wikipedia's standards to declare that nakedly since it would not be true as could be verified by a routine calculation, regardless of the existence of sources. The underlying problem is that many journalists are ignorant of the fact that 1800s ≠ nineteenth century (and so with other centuries), and thus participate in a proliferation of confusion. I would like to add a sentence of clarification that Morano is not the last known living person from the nineteenth century. Unfortunately, I have not been able to find a citation for this outside of web forums,[6] which may violate WP:SOURCES, although I probably will once she and one of Violet Brown or Nabi Tajima are deceased, so this entire matter may be better left to WP:NORUSH. However, we would then be waiting around for people to die in order to create situations which allow for nonconfusing wording as filtered through low-quality journalism, and we would be failing to settle the matter for similar cases.

I propose, since the misunderstanding of start and end years of centuries is so widespread in journalist sources yet entirely settled by the relevant authorities, that clarifications as to this matter be made explicit exceptions to WP:OR and WP:SYNTH, perhaps under "routine calculations" and in WP:What SYNTH is not. In other words, if a claim as to the first or last of anything in a given nth century or '00s exists on Wikipedia, that it always be understood NOT to be in violation of either policy to clarify that that claim is different from the first or last of anything in a given '00s or nth century, respectively. 0nlyth3truth (talk) 17:16, 16 May 2016 (UTC)

I don't think that fixing well-known misconceptions is WP:SYNTH. We fix/ignore reporters' minor errors all the time. (and I guess we don't cite the ones with grave errors) And I agree it will be helpful to warn unsuspecting readers by mentioning the corresponding cases of incorrect usage, e.g., in the form "sometimes she is <nowiki>[[Century#Start_and_end_in_the_Gregorian_Calendar|incorrectly]]</nowiki> she is called the last living person from the 19th century,[1]". Staszek Lem (talk) 00:10, 17 May 2016 (UTC)

I think that an addition like "sometimes she is incorrectly..." would be OR unless supported directly by a source. In this example, I would use "1800s" in the article and ignore "19th century" altogether after getting talk page consensus. Btw, I see no talk page discussion there; it isn't good to manage disagreement only via counter-edits and counter-counter-edits. Zero^talk 00:36, 17 May 2016 (UTC)

See Wikipedia talk:WikiProject Years/Archive 11#When do centuries and millennia begin?. I interpret the outcome of that discussion as most academic and official sources consider the last year of a century to end in the digits 00, but the general population and a few academic sources consider them to end in the digits 99. Also, the sitting pope and president of the United States took care to avoid stating whether 1999 or 2000 was the last year of the 2nd millennium. So it is impossible to answer whether the last year of the 19th century was 1899 or 1900. Jc3s5h (talk) 17:19, 17 May 2016 (UTC)

The text in question may be changed from "debunking" to positive form: "As of 2016, she together with XXX (born 1900) and YYY (born 1900) is among the last living persons born in 19th century [Footnote: The 19th <nowiki> century in Gregorian calendar<nowiki> runs from Jan 1 1801 to year Dec 31 1900] (wikilink or ext ref to the fact)". Staszek Lem (talk) 20:11, 17 May 2016 (UTC)

Website information

The following post is an excerpt from Wikipedia:Articles for deletion/List of multilingual websites.

*Comments. Wikipedia requires that an article be supported by reliable sources. Perhaps the research involved in finding reliably sourced information is what some of the commenters count as original research. The article "Facebook" (version of 11:21, 15 May 2016) says that the website is available in 140 languages, but the website https://www.facebook.com lists only 91 available languages. If an editor revises the article "Facebook" to make it agree with the website https://www.facebook.com, perhaps some of the commenters would say that that is a violation of the policy against original research. No commenter has explained what aspect(s) of "List of multilingual websites" he or she believes involve(s) original research. Different editors have different interpretations of what counts as original research, and different people have different interpretations of what counts as forbidden work. —Wavelength (talk) 00:32, 22 May 2016 (UTC)

What does this policy say about using information on the website https://www.facebook.com to correct information in the article "Facebook"?
—Wavelength (talk) 21:11, 22 May 2016 (UTC)

RfC notice: image montages of individual faces on ethnicity and other demonym articles

FYI

– Pointer to relevant discussion elsewhere.

Please participate in the discussion at Wikipedia talk:Manual of Style/Images#Proposed repeal of WP:NOETHNICGALLERIES. It is a proposal to vacate the previous consensus reached in the February 2016 RfC that resulted in the creation of the MOS:NOETHNICGALLERIES provision at MOS:IMAGES, and also relates thematically to Wikipedia:Village pump (policy)/Archive 127#RfC: Ethnicity in infoboxes (all of these discussions are ultimately about using infoboxes to identify individuals as members of particular ethnicities, and this relates also to MOS:IDENTITY). Notifying this policy talk page because NPoV and NOR are the two most frequently raised policies (pro or con) in relation to the topic of the RfC. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 17:24, 29 May 2016 (UTC)

primary vs. secondary sources for "age of consent" articles

Paraphrasing the wording of Template:Age of consent pages discussion header, statutes and other authoritative sources should be cited for the applicable "age of consent" articles, in recognition of the high standard of verification appropriate for these articles.

Such authoritative sources (e.g. statutes and case law) would generally be primary sources, and I would like to get some feedback whether it's proper that the statement in the template should be intrepreted to override the preference for using secondary sources as indicated in this article. Fabrickator (talk) 17:37, 11 July 2016 (UTC)

Author!

What about when a published author is copying his own works into a Wikipedia page? Where is the line drawn between encyclopedia and self-publishing on the web? Kortoso (talk) 21:09, 18 July 2016 (UTC)

That would in my view not be problematic for being OR (as some external quality control was applied). But it still may be problematic because it might be copyright violation WP:COPYVIO, or more likely a conflict of interest WP:COI. Arnoutf (talk) 16:55, 1 August 2016 (UTC)

This is covered at WP:SELFCITE. VQuakr (talk) 19:12, 2 August 2016 (UTC)

Plot summaries?

Shouldn't this policy article mention the fact that plot summaries (of books, movies, plays, etc) do not require sources or references, because the work being summarized is considered to be the source? HandsomeMrToad (talk) 09:38, 13 July 2016 (UTC)

Nope. This issue was discussed at length some years ago. Very simple plot elements (name of main characters, setting) can be described, but any analysis has to be sourced. Plot summaries usually require some degree of analysis of the work, therefore they should be sourced. LK (talk) 10:37, 13 July 2016 (UTC)

I disagree with Lawrencekhoo's summary of previous discussion. Summarizing a plot after having read a book or seen a movie is really no different than summarizing any of the other sources we use. Jc3s5h (talk) 11:42, 13 July 2016 (UTC)

It depends on what is contained in the "summary". A basic descriptive summation of the plot can be supported by the primary source... the work itself (and the citation can be assumed, and thus omitted). If, however, the "summary" crosses over into character or plot analysis then the work itself is not enough to support that analysis, and we need a different source (which does need to be cited).

To make this easy... a brief outline of what happens in a book/movie/TV show/ etc. can be supported by the (citation assumed) work itself... But any discussion of why it happens, or what that means, needs a separate (cited) source. Blueboar (talk) 12:03, 13 July 2016 (UTC)

Seconded. Here is a simple example. "She saw him and gasped." is OK "She saw him and gasped, scared." - is a judgement, because maybe she gasped surprised. If the source directly say she was scared, the it is OK, but if the source simply describes that she acted in a way the reader understands she was scared, then it is original research. I deliberately made a trivial example, but in fact many plots in wikipedia have much larger problems with interpreting novels and films. Staszek Lem (talk) 23:27, 13 July 2016 (UTC)

That might not be the best example, because if in context if the character acted scared (such as then turning and fleeing or screaming or some other action), describing the gasping as "scared" is reasonable. But if that was the scene, and the shot cut away and we never get any further context, then that's a problem to describe it as "scared". The metric is a reasonable interpretation that an average reader or viewer would get in considering the whole work in context, and need not be devoid of emotions as long as those are reasonably obvious. --MASEM (t) 00:09, 14 July 2016 (UTC)

That makes sense. For example, we can have a scene at a bank where a man walks up to a teller pulls out a gun and then tells the teller that they will help him make a large "withdrawl" or they will soon have more holes in them than with swiss cheese. In case an average viewer can easily conclude that this person is a robber who is threatening the teller with a gun without the character explicitly stating I am robbing this bank and threatning you with this gun.--174.91.187.248 (talk) 05:38, 4 August 2016 (UTC)

RfC on expanded use of pull-quote templates with "giant quotation marks" for any quotations

FYI

– Pointer to relevant discussion elsewhere.

The RfC at Template talk:Pull quote#Request for comments on use and documentation is relevant to the do-not-steer-readers'-interpretation aspects of this policy (as well as WP:UNDUE in WP:NPOV policy). — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 07:09, 18 August 2016 (UTC)

Terminology and WP:Synthesis

Opinions are needed on the following matter: Talk:Slut-shaming#Scope. A WP:Permalink for it is here. The matter concerns whether or not we should stick to sources that use the term slut-shaming and if not doing so can be a WP:Synthesis violation. How do we judge what is on-topic or is not synthesis if sources don't use the term slut-shaming? Flyer22 Reborn (talk) 03:17, 16 August 2016 (UTC)

The matter is now an RfC: Talk:Slut-shaming#RfC: Is it WP:Synthesis to use sources that do not identify the topic as slut-shaming to make claims about slut-shaming?. Flyer22 Reborn (talk) 23:26, 18 August 2016 (UTC)

RfC: Placing films in the Category:Films about hebephilia

Opinions are needed on the following matter: Category talk:Films about hebephilia#RfC: Should films be removed if not based on reliable sources or not fitting the hebephilia definition?. Flyer22 Reborn (talk) 03:00, 29 October 2016 (UTC)

Transcription of English audio

A Wikipedia editor, User:Rms125a@hotmail.com, said that one of my contributions counted as undue weight and original research in their edit summary. Their edit summary was "rv as OR and UNDUE" in this edit. For the undue weight part of that editor's claim, less of Garcia's quotes could be included to only keep the most relevant parts, but I was wondering about the original research part of that editor's claim. Specifically, I was wondering if transcribing English audio from YouTube videos into written English counted as original research when the audio is the English words of the person themselves in the case of Mikey Garcia or an English translation of their message by their translator in the case of Ruslan Provodnikov. I was also wondering if stating obvious things that happens in a video counts as original research. For example, to give context to a quote by Garcia I mentioned that Garcia was looking over a strawberry field in the video when he said the quote, "You know I know this, this where my dad, worked.", and I also gave context to a quote by Garcia when I mentioned that Garcia pretended to pick strawberries in the field before standing up in the video and saying "Back-breaking work man'". Were these contextual descriptions that framed the relevant quotes the original research and not the transcription of the English audio or were these contextual descriptions also not original research, because they were describing obvious things that happened in the video in order to give context for the quotes?--Ephert (talk) 20:58, 30 November 2016 (UTC)

It's not original research to transcribe information from an audio source, so long as the audio is a reliable source, and so long as the passage isn't out of context in a way that creates a meaning in the Wikipedia article that wasn't intended in the source. Jc3s5h (talk) 21:39, 30 November 2016 (UTC)

It's definitely undue; there's no reason to add a long quote like that. As for it being OR, if you took the quote from text (a book or website, etc.) then the quote could be verified against the text. In this case, we'd have to agree that your transcription is correct. If it's a translation, what if there's dispute between what the subject actually says in their native tongue and what the translator says? You claim Garcia was bending over to pick strawberries; that's definitely OR because now it gets into how the content is subjectively seen. No Wikipedian is allowed to do that. I could make a case that claiming "2+2=4" is OR if you don't have a source to prove it. What you say is "describing obvious things that happened in the video in order to give context" is inappropriate. Chris Troutman (talk) 22:12, 30 November 2016 (UTC)

Chris, if you look at the video cited, then you'll see that at 7:34, Garcia bends over to pick strawberries. There is absolutely no question about this ---- absolutely no other way to interpret what happens; using your eyeballs is not OR. As you also quite rightly say, it's definitely undue weight, and I'd thoroughly endorse the decision to revert the edit on those grounds alone, but the OR argument isn't really tenable.—S Marshall T/C 23:24, 30 November 2016 (UTC)

My Spidey-sense tingled a little which made me suspect at least some portion of the text in question is/was UNDUE and OR, but I concede that OR may not be applicable and withdraw that. If a consensus merges that I am also wrong re UNDUE, by all means restore the text in question. Yours, Quis separabit? 03:00, 1 December 2016 (UTC)

It's always difficult to distinguish what's OR / Udue. It requires epic judgemental approach. — Preceding unsigned comment added by AksheKumar (talk • contribs) 06:58, 4 December 2016 (UTC)

Lack of volunteers in Wikipedia:No original research/Noticeboard?

Most cases raised in the Noticeboard may have lacked volunteers or the third-party or something. What caused the insufficient amount of volunteers? --George Ho (talk) 22:49, 11 December 2016 (UTC)

Article about an amateur who researched stone circles

My uncle did original research on stone circles throughout the British Isles over thirty years..His work is in a library of Scottish Heritage.I doubt it has been looked at much.Would an article about my uncle be okay? HairyAlikk (talk) 15:28, 16 December 2016 (UTC)

@HairyAlikk: "No original research" here means that wikipedians cannot add text in wikipedia articles without providing references in reliable independent sources. Now, are there any publications about your uncle? Please see our guideline WP:NBIO on how to start articles about people. - Staszek Lem (talk) 18:07, 16 December 2016 (UTC)

Depending on what the level of expertise on the subject the Uncle has, this would be more a question of WP:WEIGHT than original research.--64.229.167.158 (talk) 07:48, 29 December 2016 (UTC)

If you are talking about creating a bio article about your uncle, you will need to support the article with independent sources that talk about your uncle (and his work). You would have to show that he is notable enough for us to have an article about him.

If you are talking about using his research as a source in an article about stone circles... if his work has been published, then it might be usable (but see the comment above, concerning WP:WEIGHT... a lot depends on how others view your uncle's work - his expertise and reputation). Blueboar (talk) 13:20, 29 December 2016 (UTC)

Copyright, plagiarism, and original research

This change on 11 November 2010 from

Rewriting source material in your own words, while substantially retaining the meaning of the references, is not considered to be original research.

to

Articles should be written in your own words while substantially retaining the meaning of the source material.

by SV is not a reasonable summary of the WP:PLAGIARISM guideline and is much more in keeping with SV's own personal opinion of using text copied without modification from PD sources such as the EB1911. The previous sentence is much more accurate and useful description for this specific policy. For this reason I am reverting to the previous wording. -- PBS (talk) 19:00, 22 January 2017 (UTC)

Glossary of Indian Classical Music

Glossary of Indian Classical Music Source : Harmonium Learning Center This is an alphabetical list of the more important and frequently used terms in North Indian classical music.This category is about musical terminology of Hindustani classical music, a principal sub genre of Indian classical music. Alaap:Elaboration of a melody without accompaniments

amad:arrival into the orbit of the tala leading to the sam

andolan:swinging (around a swar)

ang:lit. part, body

antar:the latter half of a khyaal composition

antra:The second half of a song based on the higher notes of the scale.

anuvaadi:neutral (refers to swar)

Amrohi:Descend of the notes Example : Ni, Dha, Pa, Ma, Ga, Re, Sa

arohi:ascending contour of a raga or ascend of the notes. Here each note is higher than the preceding note. Example : Sa, Re, Ga, Ma, Pa, Dha, Ni

asthayee:the first half of a khyaal of song compositioin or the first part of the composition. Mainly develops in the lower and the middle octave.

audov:pentatonic

bada khyaal:vilambit khayal composition

bandish:that which is bound or constrained; refers to a melodic composition fixed with words to the tala cycle; may also refer to the superstructure of the kyaal presentation. While alaap is the revelation of the raga, bandish is its design or display. Here the modes are explained in tune and words. Bandish is a composition (vocal or instrumental) fixed in a rhythmic pattern.

besura:out of tune

bhajan:devotional song

bol: syllables of text or tAla compositions

bol-alaap: Alaap done with words of text bol-taan:taans with words of text. Use of words in the improvisation of notes in medium or fast tempo chalan:lit. gait, a tonal sentence showing movement within a raga dhamar:a traditional dhrupad-like form usually set in the 14-beat dhamar taal or a style of compositionin 14 beats of a Taal dhrupad:ancient Indian classical form evolved from prabandha. 'Dhrupad' and pada (word). It has a formal structure, the details of which are beyond the scope of this glossary. A style of composition in 12beats Drut:Fast Tempo of the music ektaal:a 12-beat tala gamak:embellishment imparted to a swara; there are several types of gamakas gandhar:G or g, the third swar of the scale gat: refers to a composition played on the tabla Gat:A fixed composition of instrumental music gharana:a stylistic tradition or school in Hindustani music ghazal:a lyrical form in Urdu poetry; there are precise criteria by which a form is considered a ghazal guru:Traditional teacher or preceptor

jati:Model Scale jhaptaal:a 10-beat tala kalakar:artiste kana-swara: grace swar khraj:bass, movement in the lower parts of the mandar saptak khayal:lit. thought, the dominant form in Hindustani music of the day or composition in Hindustani music, usually in a slow tempo, in which the artist uses a great deal of improvisation komal:flattened kshudra: refers to a raga low in significance lakshan:distinguishing character laya:tempo madh:middle, qualifies the octave and the tempo madhyam: M or m, the fourth swar of the scale mandar: the lower register matra:time interval between beats of a tala. One beat of the rhythm meend:a smooth (typically descending) glide from one swar to the other melakarta:the raga grouping scheme found in the Karnataka paddhati comprising 72 sampooran scales mukhra:lit. face; the opening textual and melodic phrase of a composition or the first line of the composition. murki:a short, rapid trill nishad:N or n, the seventh swar of the scale odho:Pentatonic mode emphasizing any five notes pakkad:the 'handle' (quintessence) of a raga pancham:P, the fifth swara of the scale poorvi ang:the lower half (tetrachord) of the scale poorvi ragas:Ragas sung between the hours of 12 midnights and 12 noons raga:the melodic structure basic to Indian music or combination of notes that conveys a definite emotion raganga:raga+anga, the core feature of the raga rasa:The flavour to be realized in the atheistic emotion rishab, rikhab:R or r, the second swar of the scale roopak:a 7-beat Hindustani tala sahitya:prose, text of a composition sampoorn:complete, refers to a raga employing 7 swaras samvadi: the sub-dominant swara in a raga saptak:octave or nn octave of natural notes sam:the first beat of a tala sargam:presentation of a melody in actual notes shabad:lit. word, also refers to the hymns in the Sikh holy book, Gurbani shadav:of six swaras shadaj:S, the tonic Sa shruti:Shruti or śruti (ʃrut̪i) is a Sanskrit term, which in the context of Indian music, is the smallest interval of pitch that the human ear can detect. swar; also refers to microtones or a musical interval shuddh:It refers to the 7 basic notes of the octave or a pure note samay :Each raga has a specific time at which it an be performed. This is so as those notes are supposed to be more effective at that particular time. sthana:location swar:a musical note plus its graces swa roop:appearance, profile taiyyari:preparation (in terms of training) of a musician taal:the rhythmic cycle with matras and bols. Time measure of rhythmic beat. taan:rapid musical passage. Improvisations of notes in medium and fast tempo teentaal:a 16-beat tala tiver:augmented or sharpened, qualifies the madhyam thaat:group of sampoorn scales, taxonomical scheme for Hindustani ragas uttara ang:the upper tetrachord varja:eliminated, banned (refers to swars absent from a raga) vilambit:slow, qualifies pace of a composition or elaboration vivaadi:a foreign swar occasionally injected into a raga for ornamental or surprise value. The meaning of vivadi is "one which produces dissonance ".Which is not present in the raga. But still a vivadi swar is used in a raga by able singers in such a way that it enhances the beauty of the raga. This is done very rarely. For example some times in the raga "aiman" Shuddh Madhyam is used in between

— Preceding unsigned comment added by Bilad0032 (talk • contribs) 09:33, 7 February 2017 (UTC)

Is 222 approximately 200?

In Talk:Executive_Order_13769#200_million an anonymous wikiholic (=me) asked: Was there a community consensus (closed RfC) to supplement policy WP:SYNTH with the WP:NOTSYNTH essay intended to be an explanatory supplement, or was it added as some alternative fact? –2A03:2267:0:0:3D0D:4FD7:7B82:8DF5 (talk) 07:53, 3 February 2017 (UTC)
This interesting WP:SYNTH case was archived later. –2A03:2267:0:0:E8B1:9A50:813C:1999 (talk) 00:48, 26 February 2017 (UTC)

Evaluation of sources as OR?

I've been running into a situation where the evaluation of sources is described as OR, see for example below:

The above is a combination of original research which we dont do and actual genuine concerns via RSN discussion, permalink, plus the one below.
Once more, you are using a non-expert source to attack the credibility of biographer. That is OR and Tendentious.
You've given nothing in the above to show that he is. All looks like opinion and OR to me via RSN discussion, Archives and so on

As far as I know, evaluation of sources is not considered WP:OR in Wikipedia terms. I would like to clarify this. If this is not the case, should a statement to this effect be added to the page? K.e.coffman (talk) 02:31, 26 January 2017 (UTC)

You are correct. The first paragraph of this policy states, "This policy of no original research does not apply to talk pages." Research is often an essential part of determining whether a source is reliable or a point of view significant. RS would be totally nonfunctional if we were not allowed to argue over whether a source was reliable. Someguy1221 (talk) 02:44, 26 January 2017 (UTC)

@Someguy1221: Thank you for the clarification. Since this issue comes up on a regular basis, would there be any objection to modifying the statement as follows:

Currently: "This policy of no original research does not apply to talk pages."
~~Suggested: "This policy of no original research does not apply to talk pages; specifically, evaluation of sources by editors is not original research. (See Wikipedia:Identifying reliable sources)."~~

Or perhaps as a note, following "...does not apply to talk pages"? K.e.coffman (talk) 19:52, 27 January 2017 (UTC)

I object to the change suggested by K.e.coffman because it could be viewed as a limitation; if the change were made, someone would use it to delete discussion on a talk page because it is some form of original research on a talk page other than evaluating a source. Jc3s5h (talk) 20:11, 27 January 2017 (UTC)

That makes sense. How about: "This policy of no original research does not apply to talk pages.^[1]?

References

^ ~~For example, talk page discussions evaluating sources or identifying appropriate weight to be given to a sources are not original research.~~

K.e.coffman (talk) 20:20, 27 January 2017 (UTC)

Maybe make it more a positive example (which then means it probably should be placed somewhere in the body as to explain such allowed OR usage. Something akin to "Wikipedia talk pages are generally not held to the NOR requirement. Some activities related to the development of an article that may seem like original research, such as source evaluation, discussions on weight and merit of inclusion, and wording choices, are considered necessary elements towards building an encyclopedia and are thus allowed on talk pages. Editors are reminded that all other policies and guidelines (in particular WP:V, WP:NPOV, and WP:BLP) still apply to these talk pages, and talk pages are not to be used as forums for aspects not related towards building of an encyclopedia." --MASEM (t) 20:29, 27 January 2017 (UTC)

This looks good to me. Perhaps anchor it to WP:OKOR (as in "when OR is okay") or WP:TPOR ("talk page OR")? K.e.coffman (talk) 20:38, 27 January 2017 (UTC)

While I agree Masem's looks good, I would rather see it as an instruction creep and prefer the shorter version, a bit expanded:

"This policy of no original research does not apply to talk pages and other pages which evaluate article content, such as deletion discussions or policy noticeboards."

There is so much to say which arguments are OK and which are not. For some situations we even have essays to this end, like "Which arguments are not to use in AfD discussions." Staszek Lem (talk) 21:01, 27 January 2017 (UTC)

The shorter version works as well. K.e.coffman (talk) 23:44, 27 January 2017 (UTC)

Since there's be no further input, I implemented the last suggestion: diff, with my addition (in bold) "...evaluate article content and sources...". Please let me know if there are any concerns. K.e.coffman (talk) 00:13, 29 January 2017 (UTC)

Current consensus

Courtesy ping to Dapi89 who apparently disagrees that evaluation of sources is not OR (diff, so that they can review the above discussion. Please also see WP:OR which states:

"This policy of no original research does not apply to talk pages and other pages which evaluate article content and sources, such as deletion discussions or policy noticeboards.

K.e.coffman (talk) 18:23, 25 February 2017 (UTC)

It appears that the consequence of K.e.coffman's proposal above to insert "and sources" is that greater weight now appears to be given to a Wikipedian's original research than scholarly book reviews published in journals. --Nug (talk) 08:49, 26 February 2017 (UTC)

Coffmann, you've misunderstood, unintentionally though, I couldn't say. I object to these personal "evaluations". So far you've offered no sources. Dapi89 (talk) 12:27, 26 February 2017 (UTC)

I'm undecided, agreeing with nearly every argument for the two nearest versions. Two comments on Masem's version. I think that such an explanation is a good approach could be a good addition. However, respectfully, I think that it has a huge flaw in saying that Wp:Ver, WP:NOR apply to talk pages. Generally they don't, and this even self-conflicts with the paragraph that it is in. North8000 (talk) 14:52, 26 February 2017 (UTC)

Continued discussion

Editors Nug and Dapi89 possibly refer to the recent discussion at WP:RSN where the three of us participated: Colin Heaton's biography of Hans-Joachim Marseille, specifically when used for claims of Marseille being an anti-Nazi: Hans-Joachim_Marseille#Marseille_and_Nazism. The following 3rd party evaluations were offered as part of the discussion:

General commentary on Marseille's alleged anti-Nazi stance:

Center for Military History and Social Sciences of the Bundeswehr issued a statement in 2013 pertaining to Marseille, which states: "occasional attempts in the popular literature to suggest [...] an ideological distance to National Socialism are misleading" (for full quote & PDF, see: Talk:Der Stern von Afrika#Rabidly anti-Nazi)

General review of Heaton's work:

Horst Boog, one of the most respected German authorities on aerial warfare during WW II, devotes a whole paragraph of his review to a list of errors, concluding that there were even more errors. The preceding is re: Boog's review of Heaton's Night Fighters (which is his MA thesis at Temple Univ.) in the Militärgeschichtliche Zeitschrift (2010).

Specific reviews of the biography in question:

a review of Heaton's book by the journal Military Review in the March-April 2015 edition states "A well-written, insightful, quality book, it entertains while it educates; it is highly recommended."[7]
a review of Heaton's book on Marseille from Aviation History. Mar 2013, Vol. 23 Issue 4, p62-62. 1/2p.. It reads in part: "Writing the biography of a 22-year-old, most of whose life remains undocumented, isn't easy."

Neither of the specific reviews addressed the subject of whether Marseille was an anti-Nazi or not.

Editor Dapi offered a personal opinion:

This claim of unreliability is just an opinion of one editor. (...) Heaton qualifies as reliable. I did not see third party evidence from Dapi that Heaton is reliable when it comes to Marseille being anti-Nazi.

Editor Nug also offered a personal opinion:

I think Heaton is a reliable source for his own opinion that "Marseille was perhaps the most openly anti-Nazi warrior in the Third Reich."

Re: So far you've offered no sources -- based on the above, the sources have been offered. I believe that it's appropriate to offer personal opinions regarding sources based on available evidence. K.e.coffman (talk) 19:00, 26 February 2017 (UTC)

K.e.coffman has been raising essentially the same issue of Heaton's reliability on multiple noticeboards, first on multiiple talk pages, then on RSN, now here. This has been going on since at least July 2016. Looks like a case of WP:FORUMSHOP. --Nug (talk) 09:18, 27 February 2017 (UTC)
This discussion got restarted because editor Dapi89 disagreed that evaluation was not original research. You expressed concerns as well; see Wikipedia_talk:No_original_research#Current_consensus: ...greater weight now appears to be given to a Wikipedian's original research than scholarly book reviews published in journals.

Do you agree that evaluation of sources is not OR? Then (if we can convince Dapi) we could close this discussion. K.e.coffman (talk) 09:42, 27 February 2017 (UTC)

I agree that the evaluation of sources ought place more weight upon scholarly reviews published in professional journals in preference to personal viewpoints without reference to anything. --Nug (talk) 09:53, 27 February 2017 (UTC)

The issue is, he is using the most indirect commentary to discredit Heaton. Where do these people deal with the work at hand? Dapi89 (talk) 21:40, 28 February 2017 (UTC)

Sure, I agree with the statement that more weight should be given to scholarly reviews. Center for Military History and Social Sciences of the Bundeswehr and historian Horst Boog are indeed scholarly sources. That the two sources that reviewed Heaton's work in question are "scholarly" is debatable, as their scholarly credentials have not been established.

For editor Dapi, I point out this guideline: Wikipedia:Identifying_reliable_sources#Definition_of_a_source:

The word "source" when citing sources on Wikipedia has three related meanings:

The piece of work itself (the article, book)
The creator of the work (the writer, journalist)
The publisher of the work (for example, Random House or Cambridge University Press)

The credibility of the "creator" of the work (Heaton) has been called into question by Horst Boog in a scholarly review. No OR about this, no? K.e.coffman (talk) 02:47, 2 March 2017 (UTC)

Newspapers as primary or secondary sources

Ww2censor has stated newspapers are NOT primary sources. diff

I think that's far too sweeping a statement. In fact I think that the specific newspaper notices that are the subject of the discussion in question are indeed examples of primary sources.

Perhaps the policy and/or Wikipedia:Identifying and using primary sources needs some clarification on this point? Ww2censor has gone to a lot of trouble to find these sources, so I can understand their objecting to any suggestion that they may not be reliable secondary sources.

But my understanding is that many things published in newspapers would be primary sources... most obituary notices as an example. It's this principle I'm most interested in exploring.

Comments? TIA Andrewa (talk) 21:00, 24 February 2017 (UTC)

Unfortunately I may have written this in too dogmatic a form for Andrewa and have responded to him here pointing out the generally accepted view that newspapers are secondary sources and contrary to his suggestion, be consigned to talk pages and not used in articles. ww2censor (talk) 23:31, 24 February 2017 (UTC)

In my view, it is sufficiently dogmatic to be highly misleading. (It was posted to the talk page of a newbie, remember.) That was my reason for coming here.

I agree that, generally, newspaper articles are reliable secondary sources. That is a far weaker statement. I have no problem with it at all.

(And as an aside, it appears quite irrelevant to the discussion in which your dogmatic statement was made... a discussion in which, by contrast, your dogmatic statement if true would be pivotal.) Andrewa (talk) 08:42, 25 February 2017 (UTC)

I didn't look into the policies right now, but to say that newspapers are sources is a bit sloppy in this context. Articles in newspapers are our sources, and as such, they may be either primary or secondary, heck even tertiary ones pop up sporadically. "Primary/secondary" classification is not according to publisher, but according to in what form information is delivered, raw or regurgitated. What is more, even an individual article may be both primary and secondary. For example, a research article in a journal may be mostly a primary source, but it may have a "Background" section, which would be a secondary source. Staszek Lem (talk) 23:52, 24 February 2017 (UTC)

Hi @Andrewa, Ww2censor, and Staszek Lem:. I'll speak specifically to obituaries in newspapers. I would not call these primary sources; rather, a primary source for a funeral or death notice would be an interview with the family, or information from a mortuary or, more frequently available, the often created paper handout distributed at viewings which detail birth, death, and a eulogy. The Newspaper account is necessarily a secondary to information held by the mortuary - which is most commonly the source, I believe - or family. My interest in obituaries relates to both Wikipedia and genealogical investigations.
on a broader topic ... what are primary newspaper sources? I agree that there a significant amount of newspaper content might be primary; for instance, pictures taken by staff and included in articles ... opinion pieces, which are equivalent to personal letters or interviews ... eye witness reporting which is done by newspaper staff and entered direct as content into an article. These might not be popular views, and I've not thought very hard on this. I think the more important matter than primary/secondary is reliable/unreliable; content in newspapers which is subject to editorial review (does not necessarily include Opinion pieces or letters to the editor) should be considered reliable and intrinsically usable as citation sources. Thanks for listening. --User:Ceyockey (talk to me) 00:39, 25 February 2017 (UTC)

Contemporaneous reporting

I have a related question. I tagged a 1944 (contemporaneous) newspaper source as primary, and the response was: Nonsense. Primary sources are accounts written by people directly involved, like diaries for example, the author of that article was obviously not involved in the battle. Source: Talk:Paul_Maitla#Unreliable_primary_source

I've always thought of war-time propaganda to be an unreliable primary source. What would be the feedback here? Courtesy ping to Nug. K.e.coffman (talk) 00:49, 25 February 2017 (UTC)

A primary source is by its very nature reliable, only its authenticity can be questioned. The fact that wartime propaganda said that enemy General XYZ had syphilis, for example, is established by citing a propaganda document itself as a primary source. But as to whether the General actually had syphilis, we would look for secondary sources, and I hope we'd regard that and other propaganda documents as quite worthless. Andrewa (talk) 08:42, 25 February 2017 (UTC)

Actually, primary sources (using the "secondary == transformative") can be unreliable too; an eyewitness reporter may misreport a fact they observed, for example. --MASEM (t) 14:42, 25 February 2017 (UTC)

Good point... I'm struggling with the wording here. The fact that General XYZ was claimed to have syphilis is reliably verified by a ref to the publication that made this claim. But with an eyewitness account, as you say, there's an inherent possibility of error. Andrewa (talk) 14:51, 25 February 2017 (UTC)

I would argue for purposes of terminology that a fictitious statement like the propaganda case, or say a tabloid "X is dating Y!" are primary sources, but by their nature the sources are unreliable (and in the propaganda case, there's also the dependency aspect), so their use as a factual statement should be questioned. In the case of mistaken reporter, that would likely be a one-off type case, and ideally other primary sources would be used to show that one account incorrect. All of this comes down to the fact that it is very difficult to slap a "all sources of type X are primary/secondary" and requires understanding what the sources provides (primary/secondary/tertiary info), who wrote the source (first or third-person), and their "COI"-aspect related to the topic (dependent or independent), all which have different applications to the different content policies of V, NOR, and NPOV (and WP:N too). --MASEM (t) 15:01, 25 February 2017 (UTC)

In this case a contemporaneous newspaper reports that a certain officer was decorated for his role in a particular battle. WP:CONTEXTMATTERS, by their very nature (decorations needing the approval of the military's chain of command) such reports are fundamentally more reliable than, say, reports of an enemy General XYZ was having syphilis (which is likely to be propaganda). I cannot think of a single instance of a report of a soldier's decoration being fictitious. --Nug (talk) 08:20, 26 February 2017 (UTC)

A Victoria Cross or a Medal of Honor citation would IMO be an excellent secondary source. The investigations carried out before these are awarded are quite rigorous. But I would not be confident that all militaries are so reliable in their awards of decorations. Andrewa (talk) 10:33, 26 February 2017 (UTC)

Why would you be less confident that the German military would be less reliable in their rigor? Do you have concrete evidence of a single instance that a contemporaneous newspaper controlled by the Germans fabricated a report of such an award? -Nug (talk) 20:09, 26 February 2017 (UTC)

At the time that I tagged the article, the 1944 newspaper was citing much more than date of the award: link, such as:

Maitla with his battalion succeeded in stopping the Red Army offensive in the Battle of Auvere...^[1]^{[non-primary source needed]}

References

^ Rüütlirist SS-Hauptsturmführer Paul Maitlale anti üle (Knight's Cross to SS-Hauptsturmführer Paul Maitla awarded; in Estonian). Eesti Sõna 11 October 1944

For reliability of contemporaneous materials re: the circumstance leading to the awarding of the Knight's Cross, please see this comment: link from the Wikipedia:Articles for deletion/Heinrich Debus (SS officer) discussion, including:

...as Roman Töppel has shown (Das Ritterkreuz des Eisernen Kreuzes und der Kampfwert militärischer Verbände, Ztschr Heereskunde 446/2012), to receive a Knight's Cross a soldier needed more than to be "extremely brave on the battlefield". To be considered for the award he also needed a superior officer who would propose him. In his biography of Jochen Peiper, historian Jens Westemeier described Peiper's favoritism concerning these proposals and spoke of "Ordenskameraderie". For example, Westemeier calls the award for Georg Preuß "a bad joke".

I would also note that the German press did not have independent news correspondents at the front lines. All press accounts pertaining to military action were based on pre-screened materials from the Wehrmacht Propaganda Department. See, for example, Wehrmachtbericht (Wehrmacht Report). So the 1944 source is not independent of Germany's war-time propaganda, and is, in fact, a part of it. K.e.coffman (talk) 22:44, 26 February 2017 (UTC)

Your response appears somewhat misleading. The fact is your tagging related to two cited sentences about the actual awarding of two decorations:

In April 1943, the Estonian Waffen SS brigade participated in the battles in Nevel, and Maitla received the Iron Cross II class on 8 December for bravery.^[1]^{[non-primary source needed]}
On 29 July, he and his battalion led a counter-attack at the Battle of Tannenberg Line for which he was awarded the Knights Cross on 23 August.^[1]^{[non-primary source needed]}

References

^ ^a ^b Rüütlirist SS-Hauptsturmführer Paul Maitlale anti üle (Knight's Cross to SS-Hauptsturmführer Paul Maitla awarded; in Estonian). Eesti Sõna 11 October 1944

Secondly, your link from the Wikipedia:Articles for deletion/Heinrich Debus (SS officer) discussion is totally irrelevant and off topic, what has Peiper's favoritism got to do with anything?

Thirdly, your contention that "German press did not have independent news correspondents at the front lines" is just plain silly. Given the need to maintain secrecy and thus military censorship in place the Allies didn't have "independent" news correspondents at the front lines either. But that is irrelevant since the cited sentences only identifies the battles and the dates of the decoration. Please read WP:CONTEXTMATTERS --Nug (talk) 08:30, 27 February 2017 (UTC)

Please see the entire paragraph [as tagged:

Maitla was appointed Hauptsturmführer and in April 1944 he was commander of the 1st Battalion of the 45th Regiment of the newly formed 20th Waffen Grenadier Division of the SS (1st Estonian). In the same year, Maitla with his battalion succeeded in stopping the Red Army offensive at Auvere and received the Iron Cross I class. On 29 July, he and his battalion led a counter-attack at the Battle of Tannenberg Line for which he was awarded the Knights Cross on 23 August.^[1]^{[non-primary source needed]}

References

^ Rüütlirist SS-Hauptsturmführer Paul Maitlale anti üle (Knight's Cross to SS-Hauptsturmführer Paul Maitla awarded; in Estonian). Eesti Sõna 11 October 1944

When a citation appears at the end of the para, it's assumed that the citation covers the entire paragraph, including the statement: "Maitla with his battalion succeeded in stopping the Red Army offensive ...". Or would you disagree? K.e.coffman (talk) 08:41, 27 February 2017 (UTC)

Given that the fully cited article Battle of Auvere describes the outcome as a "German defensive victory", it is not clear why you think the statement "Maitla with his battalion succeeded in stopping the Red Army offensive in the Battle of Auvere" is controversial given that he and his battalion where participants in the battle. In any case you also tagged a single sentence paragraph:

In April 1943, the Estonian Waffen SS brigade participated in the battles in Nevel, and Maitla received the Iron Cross II class on 8 December for bravery.^{[1][non-primary source needed]}

--Nug (talk) 09:40, 27 February 2017 (UTC)

Continued discussion

In reply to Staszek Lem above, I agree totally. Very well said.

Agree that some obits are secondary sources. But one placed by a funeral director in the classified advertisements of a newspaper, just listing the name, dates of death and birth, names of relatives and similar details of the deceased, is a valid primary source IMO. Such notices are very common in Australia.

All please note that my question above states the specific newspaper notices that are the subject of the discussion in question are indeed examples of primary sources. (emphasis added) But I don't think this is the place to discuss those particular sources, that is for WP:RSN. That's why I have not linked to them here. Anyone interested can follow the diff I provided to the discussion, and find them there.

But is it fair to say that there is consensus here that the statement newspapers are NOT primary sources is completely unhelpful? That is what I am seeking to decide here, in order to reassure the relative newcomer on whose user talk page the original claim was posted.

And I need to now further ask, is there a generally accepted view that newspapers are secondary sources as Ww2censor now claims above?

This seems to me to be no better than the original statement. They describe it as less dogmatic, which is probably true, but the errors are all still there.

And they have now posted on the user talk page To me the essence of your opinion seems to be that newspapers are primary sources when we generally hold newspapers are NOT primary sources and as such are OR. Sorry but I disagree with that interpretation. So to terminate this storm-in-a-teacup I won't respond again... diff which completely misrepresents my opinion, but more important, also IMO misrepresents Wikipedia policy. Andrewa (talk) 14:32, 25 February 2017 (UTC)

Newspapers can be either primary or secondary, it all depends on context. We are looking for some type of transformation of information (analysis, critique, interpretation, evaluation, synthesis, etc.) for secondary sources, and that often happens in a larger NY Times obit (why that person was important, their impact, etc.) as well as within op-eds and larger in-depth pieces. But most of a newspaper's day to day reporting is primary, simply re-stating facts said by others or reporting eye-witness observations without commentary. It is all context though, and no single statement covers the gambit. --MASEM (t) 14:39, 25 February 2017 (UTC)

Exactly.

This is my opinion too, and is I believe Wikipedia policy.

And the continued statements by Ww2censor plainly contradicting this view need to be corrected, particularly in view of the fact that they have been posted on the user talk page of an already rather stressed and confused newbie.

I note the dismissive storm in a teacup comment. Not impressed. Failure to get to the point, IMO. Andrewa (talk) 20:36, 25 February 2017 (UTC)

With respect, I think most of this discussion is neither here no there. The important thing is whether particular newspaper articles are reliable sources and how they can be cited without violating NOR. I'm not aware of any difference it makes whether the article is a "primary" or "secondary" source. We aren't forbidden from citing primary sources; we are only forbidden from making our own interpretations of them. But we aren't allowed to make our own interpretations of secondary sources either. Zero^talk 00:55, 26 February 2017 (UTC)

I'm not addressing the issue of NOR but I will say journalism is always a primary source. Secondary sources require analysis, which is found in academic journals and some books. I think with the advent of social media the term primary source has been misunderstood by some to mean only what someone says about themselves. Primary sources are any account (independent or otherwise) that is contemporaneous. There has to be some time and distance for a secondary source analysis. The fact that a newspaper is either reliable or independent does not make it secondary. Chris Troutman (talk) 03:08, 26 February 2017 (UTC)

Not true. Your day-to-day news reports are primary as they aren't transforming information. But in-depth journalism stories can be secondary as well. Most of the NYTimes obits that go indepth about a person (not the short form ones) are secondary for that person. --MASEM (t) 03:17, 26 February 2017 (UTC)

My undergrad is in history. Anything not written at least 50-100 years after the fact is mere journalism and hence, contemporaneous primary source. An obit lacks the academic distance required. Chris Troutman (talk) 03:43, 26 February 2017 (UTC)

I'm finding Anything not written at least 50-100 years after the fact is mere journalism rather difficult to swallow. Do you have anything to substantiate that claim? – Juliancolton | ^Talk 04:22, 26 February 2017 (UTC)

By that logic, there is no way that any topic <50 years could be notable (since notability requires secondary sources). You are using a very obtuse definition of "primary/secondary" here that is not reflective of practice. --MASEM (t) 04:33, 26 February 2017 (UTC)

I think this requires careful analysis on a case-by-case basis. If the journalist who wrote the article added any sort of commentary or analysis, the article would meet the literal definition of a secondary source. But you'll find that many newspapers and websites are padded with stories that are nothing more than regurgitation of press releases, speeches and interviews, all of which are primary sources. I like what other people have said about a secondary source transforming a primary source. Now, you can argue that if a writer seeks out and combines content from multiple reliable sources to make a story, the trimming and juxtaposition of sources has transformed the content. But I'll often find an article that contains no new information beyond the single primary source it is based on, not even an opinion about the content, and I do not myself consider that a secondary source. Some will argue that the newspaper has transformed the content by doing fact-checking, even if they didn't talk about that, but I find this a dangerous assumption. And that's not to mention news stories that are actually first-hand accounts of an event by the writer, making it a primary source. So, tldr, it's complicated, but ultimately I agree that the more important question than primary vs. secondary is, "what is the source reliable for, and how is it being used?". Someguy1221 (talk) 04:51, 26 February 2017 (UTC)

Agree totally. But doesn't this make the statement that newspapers are NOT primary sources (their emphasis) an utter nonsense? And the later less dogmatic restatement that there is a generally accepted view that newspapers are secondary sources equally so? Perhaps even more nonsensical in view of the divergence of opinions above? Andrewa (talk) 08:40, 26 February 2017 (UTC)

I'm not so sure. Say we have a traffic accident, The witness statements are primary sources, but the subsequent police report is a collation and analysis of witness statements resulting in a conclusion as to what actually occurred. Therefore a police report would be a secondary source, and a newspaper report of the accident which relied on the police report would thus be a tertiary source. --Nug (talk) 08:56, 26 February 2017 (UTC)

@Nug: I think you've missed the point of Andrewa's comment. Information in newspapers can be primary, secondary or tertiary, depending on what the information is. There should be no dogmatic statements about what newspapers are; rather there should be advice on when particular content falls into different categories, e.g. an advert or an editorial comment is primary; an article based on several primary sources is secondary; etc. Peter coxhead (talk) 10:57, 26 February 2017 (UTC)

I agree with all of this, but as Peter coxhead says, it completely misses the point. It seems to confirm my opinion that the specific statements on which I have asked for comment are a load of unhelpful rubbish.

But frankly I'm getting frustrated because everyone seems to agree on this, but there seems great reluctance to actually say so one way or another. We are instead waffling away on how difficult it can be to decide in specific cases. I know that, and again it seems to back up my opinion. The statement newspapers are NOT primary sources is completely indefensible (check the context if you must, diff supplied right at the top of the section where I first raised it and repeated below). Items in newspapers can be primary sources. Is that so hard?

And the later attempt at a less dogmatic statement of exactly the same misinterpretation of Wikipedia policy by the same user is no better. But one step at a time.

Does anyone other than Ww2censor think that newspapers are NOT primary sources (context)?

And we'll take it from there. My bottom line is to give the poor newbie on whose user talk page these claims have been made some help. It's taking a lot longer than I had hoped but we will get there. Andrewa (talk) 18:39, 26 February 2017 (UTC)

The more general question

There seems some diversity of opinion above as to what is a primary source as opposed to a secondary source. I think there's some very good stuff above that clarifies the distinction. In particular, I think the point that a single document can be a primary source for some facts and a secondary source for others needs further exploration, and perhaps some tweaks to the policy to reflect this.

See Wikipedia:Articles for deletion/Usman W. Chohan for an ongoing discussion to which exactly this principle seems relevant. Andrewa (talk) 05:45, 1 March 2017 (UTC)

The single most problematic bit of advice on primary vs secondary is this line in WP:PSTS about secondary sources "A secondary source provides an author's own thinking based on primary sources, generally at least one step removed from an event." Many editors pay too much weight on the second clause (being a step removed) to make the assumption that "third-party =secondary", but the first clause is the more important one that defines the notion of "non-mechanical transformation of information" that we want. I would almost want that line to stick to the first clause, and then add a secodnary statement saying "In many cases, this is transformation of information one step removed from the event, but does not always need to be a third-party." or something like that. --MASEM (t) 07:04, 1 March 2017 (UTC)

Yes, good points all, particularly like the way you have identified the problem, and also like the proposed solution. Andrewa (talk) 08:01, 1 March 2017 (UTC)

Andrewa and MASEM are raising very timely and relevant points, and this issue is seen quite often among some other users, who tend to fixate on arbitrary distinctions between primary and secondary sources, and oftentimes lose sight of the notability of the subject at hand. Thank you for raising these pertinent points. --عثمان وقاص چوہان 12:14, 1 March 2017 (UTC) User:Uchohan

One must realize that these these terms have variable meanings in real life. So what we are talking about here is constructed Wiki-meanings, and also this in the context of their place in the Wikipedia policies and procedures. And IMHO that framework places an over-reliance on just two metrics (traditional RS criteria and primary/secondary/tertiary distinction ) which in turn relies on two myths, one of these being that a clear Wiki:primary/secondary/tertiary classification is always possible. Kudos to those who are doing their best to accomplish the latter. IMO the big answer is to add 2 more metrics (expertise and objectivity with respect to the statement which cited it) in the same prominence as the current two, have them all go into an overall "strength of sourcing" metric (rather than separate mandates) which needs to be appropriately strong for the particular situation which cited it. This would inherently slightly de-emphasize the primary/secondary/distinction and thus the struggles to create the "always possible". Sincerely, North8000 (talk) 13:16, 1 March 2017 (UTC)

Keeping in mind that most source evaluation is something done published article-to-article, things like expertise and objectivity are more broad terms that apply specifically to a work relative to a field that are determined at the point they are assesses as RSes for that field. Yes, these can change article-to-article - inclusion of controversial op-eds should be determined if we're talking a noted person in the topic area or some freshman, for example. I still feel a lot of the arguments when it comes to primary vs secondary boils down to editors mistaking WP's established approach of using transformation as "one step removed" with or without transformation, while at the same time we do have two other axes of source evaluation already in play: first vs third-party, and independent vs dependent sources. --MASEM (t) 14:38, 1 March 2017 (UTC)

Agree with everything that you just said except that it is not really as enshrined in Wikipedia as you think and in fact the most prominent areas of policy tend to lean against what you just said. North8000 (talk) 15:01, 1 March 2017 (UTC)

Agree that these these terms have variable meanings in real life but I think it's overstating the case to say that our use of them as technical terms makes them constructed Wiki-meanings. They are the best terms available to us in English, our use of them is reasonable in terms of their meanings elsewhere, and our policies and procedures etc should clarify exactly what is meant. Andrewa (talk) 19:32, 1 March 2017 (UTC)

Agree with you 100% and I did not intend to indicate otherwise. My point is that since the meanings are created in Wikipedia for Wikipedia, the the answers are being created here, not discovered here. North8000 (talk) 19:39, 1 March 2017 (UTC)

There is an element of both. At the risk of sounding like a philosopher (and with a degree in formal logic and a career in audit I am both academically and professionally qualified to split hairs) it's a bit like asking, do we create mathematics, or discover it? And my answer is, we do both at once. Andrewa (talk) 19:47, 1 March 2017 (UTC)

@Andrewa: Well, math is a lot more complicated in that respect. Here we're just talking about Wikipedia definitions of terms applicable only to internal Wikipedia purposes. For that would you not agree that the details are create-able and created here?North8000 (talk) 20:42, 1 March 2017 (UTC)

Agree. Andrewa (talk) 04:07, 2 March 2017 (UTC)

I have referred to this discussion at Talk:Oleg Sentsov#Is it time for another move discussion? in reply to the very reasonable point there There is a political or wartime aspect to this that goes beyond just counting the number of sources - the Russian sources which say he is a criminal call him "Oleg" and other sources which oppose the Annexation of Crimea by the Russian Federation call him "Oleh".

This is very much part of the larger issue here, IMO. It is rather complicated, but the bottom line is reader experience, and I think we are making progress but need to focus on some individual cases. (This one is of course further complicated by WP:BLP.) Andrewa (talk) 20:00, 1 March 2017 (UTC)

I think the best advice is to recognize that things like PSTS and other means to classify a source are not always objective, and that how sources apply to a topic will change; outside of a few limited cases, it is never a good idea to say a given source is always primary/secondary/whatever (eg going back to the original issue on this thread about calling all newspapers sources as primary or secondary). There are some fallacies that do happen when these come into play and the guidance should avoid those but it's harder to create objective rules to say when things do truly fall into nice boxes here. --MASEM (t) 20:12, 1 March 2017 (UTC)

Finding a bunch of instances and then saying there's "many" of them - is that WP:SYNTH

The article K. B. Sainis had this:

His researches have been documented by way of several articles^[1]^{[note 1]} and his work has been cited by many researchers.^[2]^[3]^[4]^[5] Besides, he has contributed chapters to many books published by others.^[6]^[7]^[8]

The citations are not sources saying Sainis' work has been cited by many researchers, or that he has contributed chapters to many books - rather they are actual instances of his work cited by researches or his contributions to books. Is it WP:SYNTH? What if it gets rid of the "many" wording, e.g.

His studies have been documented by way of several articles^[1]^{[note 2]} and his work has been cited in a number of publications.^[2]^[3]^[4]^[5] Besides, he has written chapters in books published by others.^[6]^[7]^[8]

Would that be an improvement, or still synth/OR? HaEr48 (talk) 16:20, 17 March 2017 (UTC)

References

^ ^a ^b "Sainis KB [Author]". US National Library of Medicine. 2017.
^ ^a ^b Atta-ur-Rahman (17 February 2017). Studies in Natural Products Chemistry. Elsevier Science. pp. 486–. ISBN 978-0-444-63937-0.
^ ^a ^b United Nations. Scientific Committee on the Effects of Atomic Radiation (2009). Effects of Ionizing Radiation: UNSCEAR 2006 Report to the General Assembly, with Scientific Annexes. United Nations Publications. pp. 193–. ISBN 978-92-1-142270-2.
^ ^a ^b Thomas Bechtold; Rita Mussak (6 April 2009). Handbook of Natural Colorants. John Wiley & Sons. pp. 253–. ISBN 978-0-470-74496-3.
^ ^a ^b Dr. Mohd Yusuf. GREEN DYES AND PIGMENTS: CLASSES AND APPLICATIONS. Lulu.com. pp. 163–. ISBN 978-1-329-93291-3.
^ ^a ^b Shakti N. Upadhyay (1999). Immunopharmacology: Strategies for Immunotherapy. CRC Press. pp. 10–. ISBN 978-0-8493-0951-9.
^ ^a ^b Vibha Rani; Umesh Chand Singh Yadav (14 November 2014). Free Radicals in Human Health and Disease. Springer. pp. 349–. ISBN 978-81-322-2035-0.
^ ^a ^b K. B. Sainis, Sumariwalla, P. F., Goel, A., Chintalwar, G. J., Sipahimalani, A. T., and Banerji, A. (1997). S.N. Upadhyay (ed.). Immunomodulation - Immunomodulatory properties of stem extracts of Tinospora cordifolia: cell targets and active principles. Narosa Publishing House. p. 155.{{cite book}}: CS1 maint: multiple names: authors list (link)

"several" and "many" scream POV-ish (even if that's not the intent here). It's not OR to recognize by way of examples that others have written on the studies or that his writings have been cited, but it is ORish to try to quantify that using "large" terms like "several" and "many", since if you consider the entire space of scientific publication, these are drops in the bucket. So basically you just need the wording that takes out the implicit quantity issue. "His research and papers have been documented by others." would remove the ORness. --MASEM (t) 16:36, 17 March 2017 (UTC)

Everything on wikipedia is original synthesis

Everything on wikipedia is original synthesis, so in practice the no original synthesis rule is just used to censor anything an editor doesn't like. — Preceding unsigned comment added by Jwray (talk • contribs) 21:02, 20 March 2017 (UTC)

You seemed confused. Did you intend to post this comment to Facebook, instead? Chris Troutman (talk) 22:02, 20 March 2017 (UTC)

Another RfC discussion at Talk:Cold War II

I started another RfC discussion, Talk:Cold War II#RfC: "Novel risks and measures for preventing escalation" section. I invite you to comment. --George Ho (talk) 07:26, 5 March 2017 (UTC)

Update: the RfC discussion is relisted. George Ho (talk) 03:53, 3 April 2017 (UTC)

Slight modification on translation policy

https://en.wikipedia.org/w/index.php?title=Wikipedia%3ANo_original_research&type=revision&diff=772305842&oldid=767143006, TL,DR: If there is a readily-made translation from a translator, use it as a source instead of original. Erkinalp9035 (talk) 21:05, 5 April 2017 (UTC)

More detail needed on what this quote from source policy- "primary sources that have been reputably published may be used in Wikipedia..."

In the policy about Sources (primary, secondary and tertiary) the page says, "Policy: Unless restricted by another policy, primary sources that have been reputably published may be used in Wikipedia..."

There are three sources that I would like advice on:

1. The former secret documents posted at the The National Security Archive at George Washington University.

2. The former secret documents posted at the Central Intelligence Agency's (CIA) Public Reading Room.

3. The US National Archives and Records Administration's (NARA) Presidential Libraries primary documents.

Are these three repositories of unique documents considered by Wikipedia as "reputably published" per policy?

Is there a list of Wikipedia approved primary source entities, if not can a list be started? A ri gi bod (talk) 19:34, 12 April 2017 (UTC)

You might be interested in a failed proposal from 2006 on exactly this topic. My impression is that since that time the community has slipped a little in direction of allowing archival documents, but just about any example will provoke an argument. One important point remaining is that any use of an archival document that even slightly smells like a private interpretation is definitely not allowed. With some classes of archival documents, such as raw intelligence reports, the fact that we aren't allowed to even judge their reliability by ourselves practically rules out their use. On the other hand, if a reliable secondary source refers to an archival document as reliable for something it might be ok to quote from that part of the archival document. Zero^talk 06:21, 13 April 2017 (UTC)

RfC on the WP:ANDOR guideline

Hi, all. Opinions are needed on the following: Wikipedia talk:Manual of Style#RfC: Should the WP:ANDOR guideline be softened to begin with "Avoid unless" wording or similar?. A WP:Permalink for it is here. Flyer22 Reborn (talk) 23:11, 17 April 2017 (UTC)

Secondary vs non-primary

I changed the word "secondary" to "non-primary" in the phrase "All analyses and interpretive or synthetic claims about primary sources must be referenced to a secondary source." The implication that encyclopedias can't be used to source interpretive or synthetic claims clearly runs contrary to community consensus and appears to have been inadvertent. Eperoton (talk) 13:07, 29 March 2017 (UTC)

In a universe where the only possible sources are primary, secondary, and tertiary, the change could make sense, but there is no such universe. In the real world non-secondary could mean scrawled on a piece of toilet paper I found on the floor of a public W.C. Jc3s5h (talk) 17:27, 29 March 2017 (UTC)

That objection can easily fixed by saying "All analyses and interpretive or synthetic claims about primary sources must be referenced to a reliable non-primary source." Peter coxhead (talk) 17:48, 29 March 2017 (UTC)

I agree with Peter coxhead, and I would also note that the current version of the sentence doesn't explicitly refer to a reliable secondary source. Presumably, the reliability requirement is made clear by context, because Jc3s5h's example could be a non-reliable secondary source.

If we can't reach consensus here, I'll open an RFC, also because there's another problem with this section that should be fixed. Our policies state in multiple places that WP articles should be based mainly on secondary sources, but here we define encyclopedias as tertiary sources. The classification here is based on the UMD page [8], which is very clear on two points: 1) "what some define as a secondary source, others define as a tertiary source"; 2) bibliographies, dictionaries, encyclopedias and textbooks are considered both secondary and tertiary sources. We need to fix our description to make this clear. I'm actually in the middle of an exchange with an editor who's been trying to convince me that academic encyclopedias are not appropriate sources for history articles, so these details can lead to real misunderstanding. Eperoton (talk) 18:04, 29 March 2017 (UTC)

I think the amount of time that has passed since the topic of the article occurred influences what sources are appropriate. Current news events usually lack secondary sources in the fullest sense of the phrase, because not enough time has passed for academic journals, or even weekly or monthly magazines, to publish articles on the topic. At the other extreme, events that occurred hundreds of years ago where the issues are considered well-settled won't be mentioned in recent secondary sources, so one must go searching in old journals that are hard to access, or use tertiary sources. Jc3s5h (talk) 18:30, 29 March 2017 (UTC)

I don't entirely agree with this analysis, but I'm not sure how it relates to the proposed changes, so I won't pursue that angle unless you help me understand its relevance. Do you have objections to the following proposed changes?

Change "secondary" to "reliable non-primary" to reconcile this statement with the rest of the policies, which make it clear that reliable tertiary sources are allowed.
Change the current definition of tertiary sources to to read "Tertiary sources are publications that summarize primary and secondary sources. Some types of publications, such as encyclopedias and textbooks, are considered both secondary and tertiary sources.[cite UMD again] Wikipedia is a tertiary source.[8]"

Eperoton (talk) 18:41, 29 March 2017 (UTC)

There have been no follow-up objections since then, so I'll implement the first proposal. I'm not comfortable making the second one in the absence of any comments at all, so I'll flesh out that argument further in a separate talk section next. Eperoton (talk) 00:55, 19 April 2017 (UTC)

@Eperoton: I disagree with that change but I'll ignore it for the time being. I am opposed to your second suggestion. Yale shares your point of view while Penn and Cornell rightly consider encyclopedia and the like to be solely tertiary sources. My undergrad is in history and we would never consider using any tertiary source for a paper; it's just wrong. Tertiary sources are mere restatements of secondary-source analysis. They should never be performing their own analyses. I would doubt textbooks as reliable sources, anyway. Were any editor to use a textbook to cite something, I would assume they hadn't performed any real research and are probably ignorant about the topic. Chris Troutman (talk) 01:21, 19 April 2017 (UTC)

@Chris Troutman: I think we'll need a broader discussion of this issue, because the confusion between different definitions of secondary/tertiary source have potential to lead to confusion about policy. I'll take this opportunity to lay out a first draft of my thoughts. You've provided one definition: "Tertiary sources are mere restatements of secondary-source analysis." It's a reasonable concern that one should rely too much on such sources, but it describes only some encyclopedias. For example, Brill's Encyclopedia of Islam consists of original pieces, where each entry is generally commissioned to a leading academic authority on that specific subject, and it's routinely cited as a standard reference in academic sources. The suggestion that its use should be deprecated would strike any scholar in that field as bizarre. The New Grove Dictionary of Music and Musicians is another standard academic reference. On the other end of the spectrum, some encyclopedias out there don't meet RS criteria at all. Likewise, there's a broad range of quality among what one might call "textbooks". That's not really different from any other type of source. However, our definition doesn't make clear that the distinction between secondary and tertiary sources is restriction to mere summary. It may alternatively suggest that any source which draws not only on primary but also secondary sources, is tertiary, which would of course describe the vast majority of academic publications of all types. Nor can we use the "mere" qualifier in the definition -- because how would we determine which sources provide original analysis and which mere summary without some very major OR (mind-reading)?

Here are what I see as the actual points of community consensus which these tricky distinctions are, I believe, trying to get at:

1) WP editors should not analyze, etc primary sources.

2) Our presentation of material should not differ substantially from how it's summarized by RSs, either in terms WP:WEIGHT or WP:PROPORTION, and, in case of articles treating a general subject, the use of high-level surveys helps to ensure that.

3) Some sources are more authoritative than others, and weak tertiary sources (as weak secondary sources) should not be given excessive weight.

I'm not sure how best to express these principles in terms of secondary/tertiary distinctions, but since our policies/guidelines already give clear preference to secondary sources, I think changing the characterization of tertiary sources would be the easiest route. Otherwise, we seem to be encouraging something which is contrary to the spirit of these policies: summaries based on bits sources discussing specifics of primary sources, which are cobbled together in a way that doesn't reflect how the subjects are summarized in surveys written by experts. Too many of our articles have that very problem already. Eperoton (talk) 02:16, 19 April 2017 (UTC)

@Eperoton: I honestly don't understand your objection and I see no reason for the proposed change. Our definition of tertiary sources is fine and our discouragement of using them, while not absolute, is appropriate. If you have a specific case in mind, I'd be glad to discuss it. Our definition, of course, has to apply to all tertiary sources. If there are documents which some place unreasonable faith in, that are also compendia, then the authors might consider presenting their work as a different sort of book. We ought not change our definitions for odd cases. Further, consensus on a given talk page should represent the subject-specific knowledge of editors in that field, who would be sensitive to these sources.

In the future, please take care with your pings; I didn't receive this one because it is malformed. (I failed to press the shift key when I created this account and the "t" in my username, unlike my signature, is lowercase as a result.) You can make a change in preferences (under notifications) to have the system let you know about the broken ping. The red-link is also a clear indicator. Chris Troutman (talk) 02:51, 19 April 2017 (UTC)

@Chris troutman: Hmm, I find it hard to believe that you would call professional historians from an entire field "unreasonable" because an encyclopedia happens to be a standard reference in that field, but there seems to be no other way to read that sentence. I can see that this topic will need to be revisited more thoroughly. I'll set it aside for now and do an RFC on the original change. Eperoton (talk) 01:48, 20 April 2017 (UTC)

@Chris troutman:, I think your ideas about not using tertiary sources are not applicable to Wikipedia. Your statement "we would never consider using any tertiary source for a paper; it's just wrong" would apply to papers that are to be submitted to a journal, and by extension, as a class assignment in a university class. But the goal of such papers is to offer new insights to history, and in the case of university classes, develop the research skills of scholars-in-training. Wikipedia's goal is to summarize accepted knowledge, and it is the encyclopedia anyone can edit. We allow editors who don't live within bicycle distance of a major library, and who don't have access to a vast array of online journals through their university computer accounts. In the Wikipedia environment, tertiary sources are much more acceptable than when preparing a paper for publication in a scholarly journal. Jc3s5h (talk) 11:18, 20 April 2017 (UTC)

I agree with much of Jc3s5h's comment. I've been looking at discussions of secondary and tertiary sources given in various books. First, there's no consistency in how these are defined. Secondly, the books which recommend starting research with tertiary sources but not relying on them primarily are teaching students how to perform research (i.e., find and interpret the relevant primary sources and critically engage with other original research publications about these sources). It seems that these recommendations have been incorporated into our policies without careful consideration. On the upside, they don't seem to have too much actual impact on community standards. I've been using academic encyclopedias alongside academic books for content creation in Islam-related articles for a while now, and no one has yet objected to use of a citation just because the cited source in turn cites a secondary source. On the downside, if the generally accepted purpose of this recommendation is to teach original research skills, what is it doing in a policy which is supposed to teach how to avoid original research? Eperoton (talk) 15:06, 20 April 2017 (UTC)

To amplify my comment, here are my notes on the use of these terms beyond WP: User:Eperoton/SecondaryTertiaryDefinitions. Eperoton (talk) 18:30, 20 April 2017 (UTC)

@Chris troutman and Jc3s5h: Are there still outstanding objections to the minimal change I've been trying to make? [9] I'll do an RFC on the broader secondary/tertiary issue, but this one seems uncontroversial, so hopefully we can get a simple consensus here. Eperoton (talk) 01:33, 24 April 2017 (UTC)

I object to "non-primary" because it could mean anything; the meaning is not necessarily restricted to secondary and tertiary. Jc3s5h (talk) 11:36, 24 April 2017 (UTC)

@Jc3s5h: I'm not sure what you mean, but there's an easy way to address this concern: we can replace "secondary" by "secondary or tertiary" instead. Eperoton (talk) 12:13, 24 April 2017 (UTC)

That would be better. Jc3s5h (talk) 12:16, 24 April 2017 (UTC)

Neutral notice

This is a neutral notice that an RfC on citing regarding use of primary sources has begun at Talk:RuPaul's Drag Race#Request for comment. --Tenebrae (talk) 14:28, 28 April 2017 (UTC)

Implicit synthesis

"Synthesis of published material" says, "Do not combine material from multiple sources to reach or imply a conclusion not explicitly stated by any of the sources." (My emphasis.) But there is no explanation of what implicit means or any examples given. In many articles, some editors do not think there is synthesis unless an explicit connection is made. They argue that we leaving it to readers to decide for themselves. For example:

"In a January 2014 profile of controversial political scientist Charles Murray, the SPLC labeled Murray a "white nationalist." Following the disruption of a presentation by Murray at Middlebury College in March 2017, some commentators were critical of the SPLC's description of Murray."

The text implies that the disruption resulted from the SPLC description of Murray as a white nationalist made three years earlier.

I think it would be helpful to expand information about implicit synthesis and would welcome any comments.

TFD (talk) 17:35, 8 May 2017 (UTC)

The text implies nothing of the kind. Of course, nobody can forbid readers "to read between the lines" and "connect the dots" between any two random pieces of info. And there is no way you can prevent readers from "deciding for themselves" in the weirdest ways possible. This is how all conspiracy theories are crafted. On the other hand I do smell a piece of OR in the second sentence: I am pretty much sure that "some commentators" were "critical" way before Middlebury, i.e., the second sentence tries to squeeze a correlation between "disruption" and "were critical" by cherry-picking. And mentioning Middlebury may be WP:UNDUE. And "some commentators" is WP:WEASEL.

By this logic one may disrupt any wikipedia article. Eg. "J. Random Boss was named CFO of Zugenta" "Next year he was convicted for embezzlement" Can one imply that becoming CFO makes you prone to becoming a crook? Some people say so. So what?

Concluding: in this particular case there is to way to nail down the OR without looking into full context. Staszek Lem (talk) 18:58, 8 May 2017 (UTC)

The policy uses the word "imply" in its commonplace meaning, and whether or not a sentence implies something or not may well be subject of disagreement, and especially so if there are incentives for competing interpretations. For example, I think there's enough spurious implication happening in your example to merit a rewrite, though not the kind that you mention (I think it implies that the disruption had affected the judgement of commentators, which is just confusing). The policy is clear that unsourced conclusions should not be implied, and implying means making implicit statements -- again by commonplace English usage. Implicit is not explicit. If there's a dispute about the meaning of these words, it can be resolved using a dictionary. I don't think we need an example for that. As for leaving it to readers to decide whether content violates policy, that just seems silly. By the way, we have an explanatory supplement for synthesis, WP:NOTSYNTH, which I find pretty helpful. Eperoton (talk) 20:20, 8 May 2017 (UTC)

How to describe the Emmett Till case in the lead sentence of the Emmett Till article

Opinions are needed on the following matter: Talk:Emmett Till#RfC: Should we include the "accused of showing an interest in a white woman" aspect in the lead or specifically the lead sentence?. A WP:Permalink for it is here. Flyer22 Reborn (talk) 17:08, 24 May 2017 (UTC)

No published source?

Excuse me, but has this been brought up before? Could it be "No Published Source" or something similar, rather than "No Original Research"? Would this help stop endlessly recurring edit wars, simply by calling it by what is required, and not what it isn't? Also, look up 'Original Research'; the first thing you'll find, no lie is, "Original research is considered a primary source". It seems the terms been appropriated a long time ago; who would have the power to adjust thought and usage like that, if they can be made to see why such a change is favorable? It's not just that original research has some other obscure, academic meaning; Wouldn't people object less, really, if it were called "No Published Source", rather than be told they're engaging in research? Verifiability is required, and for that we need verifiable sources, why cloud the issue otherwise? Please start with a yes or no, answering my first question, simple enough. End there, too, if you feel compelled to start reciting rules; I'm just asking if it's been discussed. — Preceding unsigned comment added by 2604:6000:A340:AE00:59BB:9D07:49AB:3D2D (talk) 14:42, 5 June 2017 (UTC) Please disregard this inquiry; 'OR' seems to be most firmly entrenched, from a quick review of it and it's accompanying articles, no one will be able to remove it, and have 'verifiability' take it's place, in half the space and not offending anyone. — Preceding unsigned comment added by 2604:6000:A340:AE00:59BB:9D07:49AB:3D2D (talk) 19:51, 5 June 2017 (UTC)

==  "No Obscure Reference" to replace "No Original Research" ==
How about replacing "No Original Research" with "No Obscure Reference"? Shortcuts remain the same except WP:ORIGINAL would be replaced with  a slightly shorter, and much clearer and correct, WP:OBSCURE.This term more accurately fits in many ways, as does the term 'Reference', both singly and combined. 
What is objectionable is that rarely is an idea 'original', and to call it such implies the idea is somehow that person's baby, produced by that person and that's why their trying to include it! Obscure immediately 'divorces' the statement from the writer and points out the real problem, that it's somewhat obscure or obscured, difficult to find. 
Research implies unpublished original research of some professional level(college), as if someone were to post their preliminary findings on Wikipedia. Rarely is their any 'research'; references, or the lack thereof, is what should be referred to.
I hope I have explained it well, perhaps some proponent could add something?
I'm asking for a 'consensus', which opens this to debate. What do you think?
 --2604:6000:A340:AE00:D4DB:6BCE:5D23:FBA1 (talk) 06:16, 7 June 2017 (UTC)

Interpretation of images / images as source

In a recent discussion we came upon the following issue: can we use an archival photo (lacking description) as a claim that a subject was given specific awards? Namely, Aleksander Piotr Mohl lists some awards, which are sourced to this photo (higher res will load after clicking). Photo archival caption identifies the subject but not the awards. Can we identify the awards ourselves, or is it OR? --_{Piotr Konieczny aka Prokonsul Piotrus| reply here} 03:24, 21 June 2017 (UTC)

odissa & midnapur district

Before the some part Midnapur district was attched with odissa, but no wikipedia searh received by me. — Preceding unsigned comment added by 103.218.238.246 (talk) 06:04, 21 June 2017 (UTC)

Re-proposal to include Wikibooks as part of cross-wiki search results

Another proposal to display Wikibooks pages in search results is made at Wikipedia:Village pump (policy). Join in there. --George Ho (talk) 18:21, 25 June 2017 (UTC)

Peer reviewed scientific papers as primary sources?

I understand the policy about reliance on secondary sources and limited use of primary sources. It seems perfectly reasonable for topics involving events or political issues. The primary sources on these subjects are quite likely to be biased.

But should the same policy apply to scientific publications? At least in theory, the peer review process is supposed to remove biased or incorrect publications. I'm concerned about the tendency in technical Wikipedia entries, to rely on media reports and newspaper summaries of the actual research. I think the press summaries of the research may introduce inaccuracy and not really avoid biases by people involved.

Fcrary (talk) 22:03, 25 June 2017 (UTC)

Yes, the same policy should apply... they can be used, but should be used with caution, because they can easily be misused in ways that constitute original research. Blueboar (talk) 22:14, 25 June 2017 (UTC)

Literature review sections are generally consider secondary. On the other hand position papers, empirical primary data and discussion sections are definitely primary. The difference is often hard to judge, so I would support calling scientific papers primary at default (including my own 70-odd scientific papers ;-). Arnoutf (talk) 14:30, 26 June 2017 (UTC)

I suppose that's part of my concern. Review papers may not include all of the data, and can be misleading. I know of one case where the original paper very carefully pointed out what was actually measured, and what was interpretation based on certain assumptions. Subsequent review papers just stated the interpretation as fact. This is even more of an issue if the secondary reference isn't a review paper, but a media report of a press release which accompanied the (primary) publication. I see references like that on Wikipedia articles all the time, and I don't think it's preferable to citing the primary source. Fcrary (talk) 17:06, 26 June 2017 (UTC)

In a scientific research paper, I would call the methods and results sections primary, discussion secondary, and related work secondary/tertiary. If it's clear that the original paper says something which a review article doesn't report correctly, one can hopefully find a consensus on how to reflect that discrepancy. On the other hand, if editors disagree about interpretation of the primary source, using it directly gets into OR territory. Eperoton (talk) 23:53, 27 June 2017 (UTC)

I would call discussion also primary - as that is where the authors interpret their own findings against the status quo in the field. While you could consider this a secondary reflection on the results, I would still label this as primary as it is the authors themselves reflecting on their own work, presented in that specific paper. Regarding incorrect use of primary sources in a review. I would say that this makes it a low quality review paper, but still a secondary source (e.g. low quality journalists also publish secondary reports). Quality and a report being primary-secondary-tertiary are different things. Arnoutf (talk) 07:07, 28 June 2017 (UTC)

As always with this much discussed issue, it depends on what the source is being used to support. I mainly write about organisms. They are usually first described and given scientific names in a scientific paper. This paper:

must not be used to support the accepted use of the name, since the scientific consensus may be that the name wasn't original and so wasn't acceptable under the nomenclature codes or that the name corresponds to a taxonomic view that was wrong – at least one secondary source is needed
can be used to support the etymology of the name – if the namer says that the name honours Joe Bloggs or that it is based on the locality where the species is found then no other source can be more definitive.

My message is that discussing the use of sources in the abstract is rarely helpful; it all depends on the context. WP:MEDRS is a useful example of detailed guidance in one specific context (but of course should not be applied to other contexts). Peter coxhead (talk) 08:12, 28 June 2017 (UTC)

Obligation to discuss noncontroversial but unnecessary clarifications on which there is no actual disgreement?

@Flyer22 Reborn: I don't quite understand what you meant by this. User:UnitedStatesian and I agree that ancient works that incorporate other lost works are still to be treated as primary sources; I just don't think it's necessary to specify that, as Wikipedians (particularly those Wikipedians who read this far down in a long list contained in a footnote) are unlikely to misunderstand this. I've only once encountered an editor who tried to argue that an ancient historical work was not a primary source because its author was not a witness to the events in question, and crafting the policy wording to cover extreme cases like that would seem to go against the spirit of WP:CREEP.

USian and I actually might have a slight disagreement over the meaning of the word "cite" in this context, as I was taking it as covering vague allusions to "sources I have consulted" (the example I gave was the Gospel of Luke) whether or not scholars can confidently identify one or more of the "cited sources" anyway. It seems USian was taking it to refer to where an ancient sources references/names a specific earlier source they consulted, and while on that point he/she may be right or wrong, I just don't think it's a significant enough issue to add an otherwise redundant clarification that earlier sources incorporated into slightly less ancient writings should not be treated as secondary sources.

All in all, I reverted the good faith edit rather than taking it to the talk page precisely because I don't see it as a disagreement over content at all, much less a disagreement worth arguing over.

Hijiri 88 (聖やや) 14:32, 12 August 2017 (UTC)

Thanks for explaining. I was wondering if any editors disagreed or might disagree with your addition. It's always good to be cautious of changes made to our guidelines and policies.

And for the record, I know that you think I'm holding a grudge against you, but I'm not. I see you around enough and we won't always agree on matters. So if I disagree with you, it's not about me holding a grudge. Flyer22 Reborn (talk) 17:55, 12 August 2017 (UTC)

Semi-protected edit request on 16 August 2017

This edit request to Wikipedia:No original research has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.

Mathew baraka (talk) 11:04, 16 August 2017 (UTC)

NOT DONE - no indication of what text is requested to be changed. Blueboar (talk) 11:10, 16 August 2017 (UTC)

FreeBMD

Moved to Wikipedia:No_original_research/Noticeboard#FreeBMD - Sitush (talk) 04:23, 13 September 2017 (UTC)

A typed transcription of a scan of hard copy index is NOT a primary source. The primary source is the certificate of birth/marriage/death, or the registrars entry into the register. If while researching at University if someone had ask me what original research I had dome, and I had said "FreeBMD", I'd have been laughed out of the fucking building. Regards DynamoDegsy (talk) 12:44, 10 August 2017 (UTC)

A copy of a primary source is still the same primary source.

As I explained at User talk:Sitush, the births, marriages and death registers (which comprise the indexes and the individual entries) are the primary sources. Making a copy of a primary source into a different medium (which is all FreeBMD is, just a computer typed-up copy of the indexes without any individual interpretation or analysis allowed - I know, because I helped do some of it), is still the same primary source.

Searching those indexes and identifying the correct records for yourself is blatant original research, and it's frought with error - any genealogist worth their salt knows that there are masses of problems if you rely on index entries, some of which I have also explained at User talk:Sitush

Wikipedia requires secondary sources and I have also explained to you what that means, and I'll quote again here from WP:SECONDARY: "A secondary source provides an author's own thinking based on primary sources, generally at least one step removed from an event. It contains an author's analysis, evaluation, interpretation, or synthesis of the facts, evidence, concepts, and ideas taken from primary sources." In this case, I went on to suggest that a suitable source might be, for example, something by someone who had done the BMD research properly and had published it in what Wikipedia considers a reliable source. Boing! on Tour (talk) 13:10, 10 August 2017 (UTC)

Here's Johnny… A few points… 1. The Indexes are typed after the birth/marriage/death certificate is issued (sometimes years after the certificate is issued), and so are at least one step removed from the event, 2. The person creating the index is not the person that issued the birth/marriage/death certificate, and so are there is evaluation, and interpretation in the creation of the index, and 3. The index shows just the quarter of the year in which the birth/marriage/death was registered, not the actual birth/marriage/death registration date, and so the information is not the same as the primary source, i.e. the birth/marriage/death certificate. DynamoDegsy (talk) 16:59, 11 September 2017 (UTC)

The discussion on my talk page that is referred to above, and in which you stood alone in your opinion, is here. Nothing has changed. Drop it, please. - Sitush (talk) 17:19, 12 September 2017 (UTC)

Actually, this discussion should be at WP:NORN. - Sitush (talk) 17:56, 12 September 2017 (UTC)

WP:Secondary sources is quite clear, and I'll repeat what I copied for you above...

"A secondary source provides an author's own thinking based on primary sources, generally at least one step removed from an event. It contains an author's analysis, evaluation, interpretation, or synthesis of the facts, evidence, concepts, and ideas taken from primary sources. Secondary sources are not necessarily independent or third-party sources. They rely on primary sources for their material, making analytic or evaluative claims about them."

The BMD indexes are just that, indexes, and carry no analysis, evaluation, interpretation, or synthesis, and no analytic or evaluative claims (and the fact that the indexes are not typed by the same person who typed the register entry (not the "certficates" - a certificate is a copy of the register entry) does not change that). The indexes are simply partial copies of some of the information that's in the BMD registers themselves, filed under quarter. They are most definitely not secondary sources. For genealogical material, you would need to find someone who has done the research properly, verified that they have all the correct entries, etc, and published the results in a reliable source. Searching the indexes to try to find the correct entries yourself and adding the results of that to Wikipedia articles is clear original research (as per WP:OR). Boing! said Zebedee (talk) 18:03, 12 September 2017 (UTC)

To expand with an example. A Tom Askin is born, an entry is made in the birth register for him, and a certificate is copied from that entry for his parents. Then, sometime later, someone else goes through all the register entries, quarter by quarter, and types out the index. All they're doing is making a one-line entry in the index for each birth contained in the register, and that is still a primary source - the primary/secondary question is not about who who creates the document or when it is created chronologically, but is about the nature of its content. What the transcriber is not doing, obviously, is saying "This is the baby who will go on to be a cricketer" (or, in the case of a death index, "This is the cricketer") - they are not identifying which Tom Askin they're typing up an index entry for. Someone identifying the index entries as relating to a specific Tom Askin would be creating a secondary source, as they are analysing the primary source, consulting with other documentary evidence, and synthesising new information based on their primary sources - and if they publish it in a reliable source, we can use it. But you and I, as Wikipedia editors, are not allowed to do it ourselves, as it is Original research. We are not allowed to search the indexes and decide for ourselves which ones correspond to the people we are writing about - several of us have given examples of how badly wrong that can go, but even if it might seem conclusive in some cases, it's still prohibited by WP:OR. Boing! said Zebedee (talk) 18:34, 12 September 2017 (UTC)

Here's an "interesting" scenario… secondary sources have identified that Tom Askin married A. Stephenson of Whitley Bridge at Kellington Parish Church on 16 August 1933… FreeBMD has the marriage of Thomas C. Askin to Aubuary (née Stephenson) registered during July→September 1933 in Pontefract district… are we saying this FreeBMD can/cannot be used to augment other secondary references? Also, as it stands, I don't believe that asking questions on talk pages is a Wiki-crime just yet, so I dont believe that comments such as… "If you are going to keep using this crap, I will get you topic banned", and this "is quite likely to end badly for you" are actually appropriate. DynamoDegsy (talk) 21:07, 12 September 2017 (UTC)

Yes, that is OR. There are even instances - more common than I think you realise - where two children of the same parents have been given the same name, and registration districts were often huge (eg: the one for Sculcoates). There are also gaps in the FreeBMD data, which the site acknowledges and which actually applies to the example you give. The general issue has been explained to you previously and WP:IDHT alone is grounds for blocking etc. - Sitush (talk) 01:50, 13 September 2017 (UTC)

In general I agree with Situch (including that this discussion should be at NORN and not here). There is no blanket rule against use of primary sources, and it is commonplace to quote from primary sources in illustration of material from a reliable secondary source that refers to the primary document. However, the problem with BMD records is the identification of the primary record. In the examples given here, the possibility of mis-identification is sufficiently high that I don't think we can use the primary source at all. I know from my own knowledge of BMD records that people of the same name doing similar things around the same time in nearby places is a common occurrence. It would be fine if there was a reliable secondary source which identified a particular primary record as belonging to a particular person, but making such identifications ourselves is not allowed. I could be persuaded of an exception if the details were extremely precise, but I don't remember seeing any examples of that in the encyclopedia. Zero^talk 03:26, 13 September 2017 (UTC)

Estimate the year of birth…

Moved to Wikipedia:No_original_research/Noticeboard#Estimate_the_year_of_birth... - Sitush (talk) 04:24, 13 September 2017 (UTC)

Is it considered original research to… estimate the year of birth, by subtracting the age at death stated in a reference, from the date of death stated in the same (or different) reference? Best regards DynamoDegsy (talk) 17:03, 11 September 2017 (UTC)

Simple math(s) is ok. But beware of synthesising sources and, in the circumstances you are referring to (which is rugby league biographies), do not use websites unless you can be 100 per cent sure they have not mirrored the original research that you already introduced into those bios. - Sitush (talk) 17:37, 12 September 2017 (UTC)

Actually, this should be at WP:NORN. - Sitush (talk) 17:55, 12 September 2017 (UTC)

Also note that this is a non-trivial calculation, unless you also know the birthday of the person. Example: Person X died 12 September 2017 at the age of 100. The simple estimate would be "Year of birth = 1917" (ie 2017-100). However, if this person would have been born on 24 December 1916 the person would still be 100 in September 2017 and the estimated year of birth would be obviously wrong. (and if we know the birthday, chances high are we already know year of birth as well) Arnoutf (talk) 20:00, 12 September 2017 (UTC)

We have a template {{Birth date based on age at death}} that takes this into account and covers the 2016/17 issue. As long as people acknowledge that without the date of birth there is the one year variation and don't make a definite statement about the year of birth and use the range or c.2016/17 then simple maths is ok. Nthep (talk) 20:20, 12 September 2017 (UTC)

^^ Agree. - Sitush (talk) 01:52, 13 September 2017 (UTC)

Are press releases and media coverage primary or secondary sources?

A number of articles related to science or spacecraft use references to press releases or media coverage of press releases. Are these primary or secondary sources? A scientific paper is clearly original research and a primary source. I assume a press release from the author's home institution is as well (e.g. ``JPL announces that a JPL scientist has just discovered...) What about a media story or blog which simply repeats and summarized such a press release? If the whole point of using secondary sources is to have some independent filtering or fact-checking of the original research, I don't think a story which simply passes on the news counts.

Fcrary (talk) 05:08, 20 September 2017 (UTC)

The original paper is a vastly more reliable source than most journalists' attempts to summarize it. As has been said repeatedly in discussions of the nature of sources, the primary/secondary/tertiary distinction which works well for many topics doesn't work for scientific ones. Peter coxhead (talk) 09:01, 20 September 2017 (UTC)

Press releases are primary. In-depth media coverage is generally secondary, but regurgitation of a press release is still primary. Material doesn't "magically" become secondary just because it's in a newspaper, otherwise op-eds, movie reviews, advice columns, even advertisements would all be secondary sources. Secondary/primary depends on the nature of the material and its (and/or its author's) relation to what it's about; it has nothing to do with publication format. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 17:20, 20 September 2017 (UTC)

Press releases, as well as all mere journalism, are primary sources. Secondary sources provide independent analysis well after the fact. If someone in 50 years writes a book or academic journal article about something JPL did today, it could be a secondary source. Anyone regurgitating what JPL has said in the present is just giving an opinion. In regards to your dispute about Juno (spacecraft), I would remove all that info per WP:CRYSTAL. Wikipedia can say that JPL said they intend to have the mission end someday and we could specify that Congress' funding runs out at such a point, but we cannot say when something in the future will end. Let's allow it to end and for someone to write something responsible about it. One of the biggest problems with Wikipedia is the vain desire by many to write the narrative up to the minute, sources be damned. Chris Troutman (talk) 17:49, 20 September 2017 (UTC)

Does this count as "original research"?

On a page I created that was recently deleted (faster than I expected it to be... what's the normal timeframe for a non-speedy deletion?), and on another page I help maintain... both are pretty heavily based on various works of fiction. Is it "original research" to say things like "This book has vampires in it", or other things you could easily and pretty inarguably say by, well, reading the work in question? Not subtle matters of symbolism or anything like that, just things like "What types of being does this book have?", or "Do the vampires in this setting need to drink human blood?" or whatever? Tamtrible (talk) 19:06, 20 September 2017 (UTC)

Non-interpretive discussion of what a book's contents are about is completely acceptable and not a violation of OR. If the text clearly uses terms like "vampires" then you can say that there are vampires in the book. If the book has a character that drinks blood and bites people in the neck but never calls them a vampire, it would be OR to say so as that's interpretation and presumption; you'd probably find commentary on the book says "You are clearly meant to assume this character is a vampire even though they don't say that..." or the like. --MASEM (t) 19:42, 20 September 2017 (UTC)

Attention needed to clear up serious confusion about secondary sourcing on the part of the film wikiproject

Please see this sprawling discussion, in which an alarmingly large number of people are convinced that primary-source opinion pieces in the form of film (and book, etc.) reviews are secondary sources within WP's meaning because they're reviews of other works (i.e. that the work that is the subject of WP's article is the primary source, and that individual opinion magically transubstantiates into secondary sourcing because it's about a work instead of about, say, a mineral or a person). A secondary mistaken argument is that it must be secondary because it's in a newspaper/newssite, and this makes it secondary because everything in a news-focused publication is magically secondary. This is total bollocks, but they're actually trying to change an important site-wide guideline to reflect this nonsense. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 17:49, 21 September 2017 (UTC)

Well it is a secondary source as to the film, and a primary source as to the author's opinion. Alanscottwalker (talk) 18:23, 21 September 2017 (UTC)

Trying to direct people to that discussion not fork a new one, but that over-generalization is a part of the issue; a review can be (not "is") a secondary source for a quite limited subset of things about the content of a film, and there's a fine line on both sides: simple reguritation (e.g. of what the credits say) isn't secondary sourcing, nor is highly opinional "analysis"or "interpretation" of the content that's coming straight from the imagination of the reviewer. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 23:11, 21 September 2017 (UTC)

Serious and deep seated OR issue needing experts

Rommel myth, an article that you or your project may be interested in, has been nominated for a community good article reassessment. If you are interested in the discussion, please participate by adding your comments to the reassessment page. If concerns are not addressed during the review period, the good article status may be removed from the article. LargelyRecyclable (talk) 13:52, 22 September 2017 (UTC)

WP:PSTS, MOS:TONE/MOS:FICT, and reviews of fictional works

FYI

– Pointer to relevant discussion elsewhere.

Please see Wikipedia talk:Manual of Style/Film#Proposed clarification of reviews' relation to WP:PSTS and MOS:TONE. This is a request for comments in the general sense, but not a WP:RFC at this stage, being an initial discussion draft (broadened to cover writing about fiction generally), building on a lengthy discussion/dispute at the same page. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 15:27, 22 September 2017 (UTC)

Are interviews considered secondary sources?

If so, could that info be added to the secondary source section [10] as an example? TimTempleton ^(talk) ^(cont) 17:00, 1 November 2017 (UTC)

In my opinion: An interview report often consists of two parts: introduction and answers. Intro is definitely secondary. Answers may be both: they are primary if a the interviewed speaks about themselves, their observations, and their work. They are secondary when the interviewed speaks about work of others within their area of expertise. There rest is hearsay. Staszek Lem (talk) 18:16, 1 November 2017 (UTC)

No, it's all primary. There's no secondary source analysis in an interview (in any part) and there's no significant distance from the subject. Chris Troutman (talk) 18:17, 1 November 2017 (UTC)

When an expert summarizes work of others in the area of their expertise, how can this me a primary source? When an interviewer in the introduction writes "Professor Smith was born and raised in New Hampshire", how this is a primary source? Staszek Lem (talk) 18:47, 1 November 2017 (UTC)

The truth is in the middle, as usual. It depends on the case. Some interviews are preceded by secondary material that is regular journalism. Some are preceded by the interviewer's op-ed opinion, which is primary. The interview proper is primary by definition. — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 19:14, 1 November 2017 (UTC)

interview proper is primary by definition - which definition? It is one thing an interview of a bystander who witnessed a crime. An interview of a physicist about the modern state of the string theory is totally different. Staszek Lem (talk) 19:20, 1 November 2017 (UTC)

No it isn't, as during the interview the physicist is likely to be enticed to give their personal opinion on which they cannot reflect and would never write down as an overview of the field. Also take into a account that if the same physicist publishes a scientific article in which their own theory is proposed as pinnacle of modern state of string theory, that scientific article is also a primary source. Arnoutf (talk) 19:28, 1 November 2017 (UTC)

Agreed. What's happening here is a confusion of the primary/secondary/tertiary distinction, with the high/mid/low quality distinction. An interview of a physicist about their branch of physics is a high-quality primary source, while an interview of a rock star about physics is a low-quality one. The former is still primary because it involves no analysis/evaluation/interpretation/synthesis, and its just direct from the brain and mouth/keyboard of the interviewee (aside from minor copyediting, trimming for length, etc., before publication/broadcast). It is just a question-mediated editorial. And, yes, editorials about physics by physicists are also primary sources; all editorials are primary sources in nature, though some in-depth ones may also contain some secondary material, e.g. if they do something like a mini-literature review as background material. — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 19:50, 1 November 2017 (UTC)

A bit of context. When I write articles about companies or individuals, I find that interviews are convenient ways to source historical and biographical info. It's also an easy way for a journalist to not have to spend a lot of time researching and describing company/personal history, The downside is that statements in interviews are not necessarily vetted as would be the case if the interviewer wrote a traditional piece using material from the interview and combined it with other sources. When I participate in AfD reviews, I sometimes come across prolific deletionists arguing that interviews are not valid sources, that they are promotional, and they don't demonstrate notability. This delineation between primary and secondary sources seems to be a potential "official" policy guide regarding using interviews as sources, and I was hoping to get some clarity. TimTempleton ^(talk) ^(cont) 20:55, 1 November 2017 (UTC)

interview of a physicist about their branch of physics is a high-quality primary source editorials about physics by physicists are also primary sources --Sorry, this is your personal opinion I respectfully disagree. Please provide an evidence it is a common understanding. I suggest you to read our article primary source and references cited within. Any document which analyzes/summarizes primary sources is secondary source. An editorial about the current state of physics is primary source if it is clear that the author speculates off his head. If the author's narrative is built upon analysis of various scholarly works, it is a secondary source, per definition. Staszek Lem (talk) 21:24, 1 November 2017 (UTC)

WP's article Primary source isn't a reliable source and its policy; it doesn't define what WP means by that phrase internally; it's a descriptive article about what the term means (which varies) off of Wikipedia. I'll address the rest of this below. — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 22:37, 1 November 2017 (UTC)

Yes, I know that, and my full quote was primary source and references cited within. Staszek Lem (talk) 00:36, 2 November 2017 (UTC)

What's happening here is a confusion primary/secondary/tertiary distinction, with the high/mid/low quality distinction. No. What's happening is the confusion of the definition accepted in historical sciences and the definition used in wikipedia. If the community agrees that interviews are primary sources for the purposes of Wikipedia, I have no troubles with accepting that. Heck, an interview which summarizes existing knowledge without drawing novel conclusions is even more remote, tertiary source. Staszek Lem (talk) 21:36, 1 November 2017 (UTC)

This actually addresses your disagreement with me in the point immediately above this. WP defines a secondary source, for our purposes, as having the qualities of a secondary source (in terms described generally at that article, and more precisely defined in policy as having analysis, evaluation, interpretation, or synthesis of previous material) and from a reputable publisher, not self-published. An interviewee giving their on-the-spot view, even if it's "built upon analysis of various scholarly works" (which any physicist's interview opinion about physics matters will be) is effectively self-published; the interviewer and publisher of the interview is doing nothing to vet the material, and is simply conduit (i.e., effectively a |via=, though we don't cite it that way). A "Wikipedia-secondary" source requires some form of reputable editorial judgement to act up the factual material (not upon just "is this person interesting enough to interview?", the answer to which might be "Yes, because they crazy and full of crap, and people will get a good laugh out of their crazy ideas"). PS: Any time you say on a talk page "this is [just] your personal opinion" you are probably making a mistake, since virtually everything we do and say in "Wikipedia talk" namespace is editorial opinion, from which consensus forms. There is no external source for what WP policy (and WP policy interpretation) is; it emerges from the gestalt of editorial judgment (i.e., opinion) on the policy question at issue. — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 22:37, 1 November 2017 (UTC)

In other words, you are saying that an interview is not guaranteed to be a reliable source, i.e., this is not an answer in the original question. Staszek Lem (talk)

Any time you say on a talk page "this is [just] your personal opinion" you are probably making a mistake - No I am not. In talk pages we also have (a) references to policies/guidlines/consensuses and (b) logical arguments. Ye are right in the remaining, but until the consensus forms we are in disagreement, and my remark says that I am in a disagreement with your statement, not with the policy/consensus you may have alluded to, nor with your reasoning. Staszek Lem (talk) 00:34, 2 November 2017 (UTC)

Meh. If you disagree neither with my reasoning nor what I cite as backing it, then you're understanding me. I don't think its productive to continue arguing about exact wording no one's actually having difficulty with, especially since that particular phrasing isn't going to affect the content of this policy page or how it's applied. — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 11:17, 2 November 2017 (UTC)

An interview may contain secondary source material, transformative comment, on a third subject. May. But in general, if you need to ask, consider an interview to be a primary source. —SmokeyJoe (talk) 21:50, 1 November 2017 (UTC)

They are primary sources for Wikipedia purposes, because they are unreliable and require interpretation. Furthermore, many interviews are edited which can lead to further inaccuracy. TFD (talk) 23:10, 1 November 2017 (UTC)

No. Reliability issues are governed by WP:RS. In my opinion an interview is a form close to WP:SELFPUBLISHED (due to possible 3rd party post-editing). Staszek Lem (talk) 00:34, 2 November 2017 (UTC)

in general, if you need to ask, consider an interview to be a primary source - I would agree with this as a generic advice, because vast majority of interviews are about personal opinion or memories. However we should not rigidly exclude the first part of your sentence. Staszek Lem (talk) 00:43, 2 November 2017 (UTC)

Yes. It does happen that an elder stateman may be interviewed, question and answer, on an historic subject. If the topic doesn't concern that elder statesman, the verbal account s/he gives, albeit prompted by questions, constitutes an independent secondary source for the historic subject, and it may well be even considered reliable. Note however, it is very unlikely that that someone will be attempting to claim the interview as the evidence that the topic is notable, because in such cases the historic topic will be obviously notable. We are probably talking about the great depression, WWII, the moon landings, or some political crisis. What I think I notice is that when interviews are proffered as evidence of notability for borderline notable topics, the interview borders promotion, and non-independence. This is regardless of of the primary/secondary source-typing, which will not be the main question. There is good advice at Wikipedia:Identifying and using primary sources and Wikipedia:Party and person. Interviews are rarely good sources except as primary sources for the opinion of the interviewee. --SmokeyJoe (talk) 03:07, 2 November 2017 (UTC)

Agreed with your conclusion, but not the opener. How old someone is and what their job was doesn't have anything to do with whether their recollections and musings are primary or not. Recollections and musings are the product of subjective human memory and opinion formation, ergo primary. It's effectively impossible for an interview to be a secondary source, because it's off-the-cuff, and the interviewee is not performing WP:AEIS on other source material, but on their own brain, nor is the output of the interviewee subject to any meaningful editorial review. So, not secondary, twice over. — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 11:17, 2 November 2017 (UTC)

It depends. An interview with an eyewitness (primary source) to an event without asking anything further "opinions" is still primary. An interview with a movie star about their role in a movie that they have filmed for, where the actor can describe some of the behind-the-scenes, that's secondary. --MASEM (t) 03:22, 2 November 2017 (UTC)

No it is not. As the actor's account of the behind-the-scenes is clearly a primary report of the actor's perception of what went on behind the scenes. Probably very relevant to Wikipedia, but still a primary account. If, however, many people involved with filming would have been interviewed and published, and if someone, using an established method would have analysed and then concluded what actually went on behind the scenes by combining points of view across all those interviews - that would be a secondary source on what went on behind the screens.Arnoutf (talk) 10:15, 2 November 2017 (UTC)

Yep. And all that applies just as much to physicists talking about their insider perspective on issues in their field, or "elder statesmen" reminiscing about how an international conflict was managed. There is nothing comparable (that WP needs to care about) between an interview and a secondary-source publication (even by the same individual); in the latter case, we can judge the quality of the output based on what prior works they're attributing and synthesizing/analyzing, and on the fact that a real publishing company with a reputable editorial staff decided that the work passed muster after it was worked over. Interviewees are relying on their own brains, not prior works, and their commentary is almost entirely unfiltered (usually just trimmed for length, if even that). — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 12:39, 2 November 2017 (UTC)

Interviews where someone is talking about themselves or their experiences (including projects they were involved in, like movies) are primary. Alanscottwalker (talk) 11:51, 2 November 2017 (UTC)

It's very easy to make the claim that all interviews are primary sources without making any exceptions when WP:ALLPRIMARY proves that all sources are primary for something. However, ALLPRIMARY also states, "More importantly, many high-quality sources contain both primary and secondary material." Anybody can prove a source is primary for some reason or another. This requires no special knowledge or skill. It takes a great mind that thinks outside the box to be able to extract the valuable secondary source information from the exceptions that do exist. Huggums537 (talk) 01:00, 3 November 2017 (UTC)

Yes, it's very easy to know what a primary source is, if you know what a primary source is -- if you don't like the primary "box", that's just tough, it's like saying one does not like things that bother you. Alanscottwalker (talk) 14:58, 6 November 2017 (UTC)

Someone commenting on oneself, commenting on previously accepted facts about oneself, and ascribing interpretation or meaning to those facts, is a non-independent, possibly-reliable, primary-source and secondary-source on oneself. --SmokeyJoe (talk) 01:12, 3 November 2017 (UTC)

I forgot to add that just because anyone can easily prove a source is primary doesn't automatically disprove it as being secondary as well. This idea that a source must be either/or secondary/primary is the reason for so much debate. The users who had the right idea were:

Staszek Lem, who said, "Answers may be both"
SMcCandlish said, "The truth is in the middle, as usual."
SmokeyJoe said, "An interview may contain secondary source material, transformative comment, on a third subject. May."
Masem said, "It depends."

Huggums537 (talk) 03:05, 3 November 2017 (UTC)

@Huggums537: Not only you are dead wrong, you've also quoted more than one Wikipedian out of context. If a source is primary, it cannot be secondary. Furthermore, Staszek Lem is consistently wrong with everything they've said. To give a bit more clarity to my original answer (as others have stated), interviews are primary when they are contemporaneous about the subject. Secondary sources have to have temporal distance from the subject. They have provide an analysis, not just an observation or an opinion. For that reason, pretty much every interview you're going to see as a source here is a primary source. Chris Troutman (talk) 03:41, 3 November 2017 (UTC)

"If a source is primary, it cannot be secondary." Iffy, that phrasing Chris. A source, including an interview, can be primary for one purpose, and secondary for another purpose. These purposes may occur sequentially in the one paragraph. However, these will be rare examples among rare examples of interviews being secondary sources. --SmokeyJoe (talk) 04:59, 3 November 2017 (UTC)

(edit conflict) Well, it's important to distinguish between sources that are primary for something (e.g. a literature review that closes with a personal prediction of what research to keep an eye out for next quarter is secondary for the lit. rev. material, but primary for the future-prediction opinion), versus a primary source that is primary by its very nature, e.g. an eyewitness report, or one of Trump's tweets. That is, a) while WP:ALLPRIMARY is not reversible (it isn't true that all sources are secondary for something), b) "primary" has more than one relevant sense, and they're not equivalent in all ways. Agreed on the gist of you [Chris Troutman] say about temporal distance (which we don't talk about enough) and the implications this has for most interviews (and, though you didn't touch on this, most news reporting as opposed to journalism in the more formal, investigate sense). No opinion on the editor-related complaints. >;-) — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 05:04, 3 November 2017 (UTC)

If all sources are primary for something, then it only stands to reason that all secondary sources are also primary sources for something. Thus, all secondary sources are both secondary sources and primary sources for some reason or another. Logic dictates it, not me. Therefore some, but not all, primary sources are also secondary (or even tertiary) sources. I will admit, however, that some primary sources are strictly primary sources, and have no secondary or tertiary characteristics. At any rate, it's very easy to accuse someone of quoting out of context when the actual nature of quoting inherently demands that you take precise sections of words/phrases directly out of the original context from which they came. Quoting quite literally is taking words out of context. Besides, picking an easy battle about that dodges the real issue and focuses more on edit behaviour than it does on the debate about the topic anyway. Huggums537 (talk) 06:58, 3 November 2017 (UTC)

The silliness of this discussion illustrates once again the harm done by the primary-secondary distinction. Everything we are not allowed to do with primary sources is also not allowed with secondary sources. If we formulate the rules for sourcing without the primary-secondary distinction, they get simpler: the source has to be reliable, we have to report what it says without original analysis, and we can't combine multiple sources to draw a conclusion not drawn by any of them. (And I always get flak for expressing this opinion; go for it.) Zero^talk 08:03, 3 November 2017 (UTC)

Zero That is because your opinion to set primary and secondary source at equal footing will lead to endless POV forks. If we treat reliable primary sources with the same regard as reliable secondary sources, a single scientific paper (reliable source) finding some weird effect, possibly as the result of a luck hit in data collection (and this happens a lot - see the replication crisis in social psychology) would be allowed for inclusion in an article. Especially in social sciences where different schools, and even different groups following the same school use subtle different paradigms to find slightly different takes on the same phenomenon this would lead to an endless explosion of micro-theories in almost all articles. To avoid this we simply need aggregated knowledge to avoid overly reliance on these spur of the moment findings (i.e. secondary sourcing). Arnoutf (talk) 08:16, 3 November 2017 (UTC)

The argument against scientific papers is lost, given that the majority of our scientific articles cite journal papers and the fraction is increasing as more qualified people edit articles. Your argument doesn't work either, as we have a rule (WP:NPOV) telling us exactly what to do when multiple sources of similar reliability disagree with each other. Zero^talk 11:17, 3 November 2017 (UTC)

A different answer that doesn't conflict with Arnoutf's: It's just not even close to true that "Everything we are not allowed to do with primary sources is also not allowed with secondary sources." WP:AEIS is pretty much the cornerstone of our entire reliability system. These things – analysis, evaluation, interpretation, and synthesis – are not permissible with or from primary sources, but are not only permissible with and from secondary sources they are a substantial portion of what determine that the source is secondary at all. — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 08:54, 3 November 2017 (UTC)

This is interesting. Are you suggesting there are different RS/OR sourcing rules for primary sources than for secondary sources? Huggums537 (talk) 09:24, 3 November 2017 (UTC)

@SMcCandlis. I'm simply staggered by your misreading of WP:AEIS. It refers to the author's "analysis, evaluation, interpretation, or synthesis" of the subject matter, not our "analysis, evaluation, interpretation, or synthesis". We are not permitted to present our own "analysis, evaluation, interpretation, or synthesis" of any source at all. Talk about cornerstones! Zero^talk 11:06, 3 November 2017 (UTC)

I'm staggered you'd think I'm saying otherwise. One thing we are allowed to do with secondary sources is cite them for analytic material. Another is to cite them for evaluative material. And so on through A, E, I, and S. You cannot do this with primary sources. Any apparent AEIS claim in a primary source cannot be used here and cited to that source. (Exception: It can be done with direct quotation/attribution, but only if the primary analysis statement itself is important in the context, and even then it cannot be given undue weight, and it cannot be done in a way that implies it's anything but the opinion of the writer – "not in Wikipedia's voice", as we say. E.g., you may do "According to Siskel and Ebert ..." about a film they reviewed, being world-renowned film reviewers, but you cannot do the same with the film reviewer of your high school newspaper, or cite Siskel and Ebert's opinion about the economics of Tahiti, even if they tried hard to analyze it, because film reviewers' opinion about economics are meaningless, unencyclopedic trivia.) — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 09:03, 4 November 2017 (UTC)

The point of a secondary source is that it was, we hope, carefully written over an extended period with time for reflection and fact checking. The author read the work, we hope, several times to make sure it says what was intended. A secondary source gives an analysis and the author includes anything related which they regard as important. That helps us decide what is WP:DUE. An interview is a primary source—when speaking off-the-cuff, people make mistakes such as getting someone's name wrong—the speaker would be clear in their mind who they are talking about, but the wrong name comes out of their mouth. A polite interviewee may answer the questions provided while privately thinking the questions are rubbish focused on unimportant issues. Using an interview to decide what is WP:DUE may reflect the views of the interviewer, not the expert being interviewed. Johnuniq (talk) 09:50, 3 November 2017 (UTC)

An example I came across was an editor wanting to use the television interview of a noted expert as a source for an article about U.S. foreign policy. The content of the interview could have been presented as a paper in a journal in which case it would have been a reliable secondary source. However because it was an interview, it is not reliable because there was no fact-checking and we do not know if editing to fit the time-slot might have changed the meaning. Furthermore, if the facts mentioned by the interviewee were significant, they would have been mentioned in a reliable source. So we are left with only one possible use - as a primary source for the interviewee's opinions. And here common sense comes into play. Why would anyone use an interview with its limitations as opposed to a published article as a source? In my experience, it is because it contains something not found elsewhere. And the usual reason is that the interviewee misspoke or was misrepresented or that the information the editor has found is trivial. And that is the usual reason for straying from the best sources. TFD (talk) 10:41, 3 November 2017 (UTC)

I agree with the result, but that is because interviews (especially live interviews) are inherently unreliable for details. WP:RS applies to everything. The conclusion of unreliability doesn't change whether we put it into the "primary" column or the "secondary" column. Zero^talk 10:58, 3 November 2017 (UTC)

The problem here is that the interview itself may be a reliable albeit primary report of the spoken opinion of the interviewee (in fact the audio-recording may prove that). The content of the statement of the interviewee may, or may not be reliable - and we have no way to check that. That is why, in my opinion, interviews can generally be regarded as a reliable primary source of what words the interviewee actually uttered (but not much more). Arnoutf (talk) 12:17, 3 November 2017 (UTC)

Of course every source is reliable for its own content. I should have written "inherently unreliable for factual details". We choose to regard live interviews as unreliable for facts because (as Johnuniq said) we know from experience that interviewees often mis-speak or mis-remember on the spur of the moment. However, not being able to check something is irrelevant as there is no rule anywhere in the Wikipedia rule book that we have to be able to check facts ourselves. The closest it gets is the general guideline (alas, frequently violated) to seek out the most reliable sources. Zero^talk 10:47, 4 November 2017 (UTC)

I had tended historically to agree, but I'm getting a directly contradictory answer at Wikipedia:Reliable sources/Noticeboard#Citation of TV broadcast. When long-term, smart, committed, and source-serious Wikipedians directly contradict each other on something this basic, then we clearly need a policy clarification. — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 13:36, 5 November 2017 (UTC)

No, there's no contradiction. The question at RSN is about whether citing an old broadcast where there is no clear access to the source per WP:V is met, which is a separate question from here, where, presuming WP:V is met, if that source is reliable. They are two different questions. --MASEM (t) 17:53, 5 November 2017 (UTC)

Agree with Masem. I see no contradictions either. Huggums537 (talk) 22:29, 5 November 2017 (UTC)

Neither do I for much the same reasons as Masem mentions. Arnoutf (talk) 07:18, 6 November 2017 (UTC)

Reading the above discussion, I get the feeling that some of the comments assume that Primary sources are considered Unreliable. That is not the case. A primary source can be reliable... indeed in some cases a primary source is the most reliable source possible. Of course, the reliability of a primary source depends on the specific statement that the primary source is supporting. But that is true for secondary sources as well. The caution in this policy (and it's preference for secondary sources) is focused purely on the fact that it is easier to engage in Original Research when using primary sources. It is not talking about the appropriateness of the source... but the appropriateness of what WE say, based on that source. Blueboar (talk) 14:51, 6 November 2017 (UTC)

Blueboar Thanks for that important element. Sadly, in many article discussions the nuance you voice so well in your comment is lost. I fully agree Primary is not bad. In fact an interview would be a brilliant (primary) source for the claim: "In recent ABC news interview person Y claimed "X is true"." (or a more realistic (albeit made up) example e.g. In an North Korean TV interview, Kim Jong-Un claimed that North Korea has ballistic missiles capable of delivering the H-bomb to California) ; but less brilliant for the outright claim: "X is true" (North Korea has ballistic missiles capable of delivering the H-bomb to California) where a secondary source would be needed. Arnoutf (talk) 10:07, 7 November 2017 (UTC)

Opinions needed at Surrealism

Your opinion is requested at Talk:Surrealism#Creeping SYNTH and OR. Thanks, Mathglot (talk) 22:11, 6 November 2017 (UTC)

Some kind of Dalî joke belongs here, but I haven't had enough coffee to come up with one. — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 23:21, 6 November 2017 (UTC)

Well, as long as you're having another coffee, here's one for you:

Q: What did Salvador Dali eat in the morning?

A: Breakfast surreal.

Mathglot (talk) 11:26, 7 November 2017 (UTC)

Ceci n'est pas une réponse. — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ< 23:37, 20 November 2017 (UTC)

Cite error: There are <ref group=note> tags on this page, but the references will not show without a {{reflist|group=note}} template (see the help page).

[1] ~~For example, talk page discussions evaluating sources or identifying appropriate weight to be given to a sources are not original research.~~

[EestiSõna-2] Rüütlirist SS-Hauptsturmführer Paul Maitlale anti üle (Knight's Cross to SS-Hauptsturmführer Paul Maitla awarded; in Estonian). Eesti Sõna 11 October 1944

[Eesti-3] Rüütlirist SS-Hauptsturmführer Paul Maitlale anti üle (Knight's Cross to SS-Hauptsturmführer Paul Maitla awarded; in Estonian). Eesti Sõna 11 October 1944

[Eesti-4] Rüütlirist SS-Hauptsturmführer Paul Maitlale anti üle (Knight's Cross to SS-Hauptsturmführer Paul Maitla awarded; in Estonian). Eesti Sõna 11 October 1944

[Sainis_KB_[Author]-5] "Sainis KB [Author]". US National Library of Medicine. 2017.

[Atta-ur-Rahman2017-7] Atta-ur-Rahman (17 February 2017). Studies in Natural Products Chemistry. Elsevier Science. pp. 486–. ISBN 978-0-444-63937-0.

[Radiation2009-8] United Nations. Scientific Committee on the Effects of Atomic Radiation (2009). Effects of Ionizing Radiation: UNSCEAR 2006 Report to the General Assembly, with Scientific Annexes. United Nations Publications. pp. 193–. ISBN 978-92-1-142270-2.

[BechtoldMussak2009-9] Thomas Bechtold; Rita Mussak (6 April 2009). Handbook of Natural Colorants. John Wiley & Sons. pp. 253–. ISBN 978-0-470-74496-3.

[Yusuf-10] Dr. Mohd Yusuf. GREEN DYES AND PIGMENTS: CLASSES AND APPLICATIONS. Lulu.com. pp. 163–. ISBN 978-1-329-93291-3.

[Upadhyay1999-11] Shakti N. Upadhyay (1999). Immunopharmacology: Strategies for Immunotherapy. CRC Press. pp. 10–. ISBN 978-0-8493-0951-9.

[RaniYadav2014-12] Vibha Rani; Umesh Chand Singh Yadav (14 November 2014). Free Radicals in Human Health and Disease. Springer. pp. 349–. ISBN 978-81-322-2035-0.

[Immunomodulation_-_Immunomodulatory_properties_of_stem_extracts_of_Tinospora_cordifolia:_cell_targets_and_active_principles-13] K. B. Sainis, Sumariwalla, P. F., Goel, A., Chintalwar, G. J., Sipahimalani, A. T., and Banerji, A. (1997). S.N. Upadhyay (ed.). Immunomodulation - Immunomodulatory properties of stem extracts of Tinospora cordifolia: cell targets and active principles. Narosa Publishing House. p. 155.{{cite book}}: CS1 maint: multiple names: authors list (link)

[1]

[1]

[1]

[1]

[1]

[note 1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[note 2]