Wikipedia talk:Identifying reliable sources

From Wikipedia, the free encyclopedia
  (Redirected from Wikipedia talk:RS)
Jump to: navigation, search
Frequently Asked Questions (FAQ)
Where should I ask whether this source supports this statement in an article?
At Wikipedia:Reliable sources/Noticeboard. Don't forget to tell the editors the full name of the source and the exact sentence it is supposed to support.
Do sources have to be free, online and/or conveniently available to me?
No. Sources can be expensive, print-only, or available only in certain places. A source does not stop being reliable simply because you personally aren't able to obtain a copy. See Wikipedia:Reliable sources/cost. If you need help verifying that a source supports the material in the article, ask for help at Wikipedia:WikiProject Resource Exchange or a relevant WikiProject.
Do sources have to be in English?
No. Sources can be written in any language. However, if equally good sources in English exist, they will be more useful to our readers. If you need help verifying that a non-English source supports the material in the article, ask for help at Wikipedia:Translators available.
I personally know that this information is true. Isn't that good enough to include it?
No. Wikipedia includes only what is verifiable, not what someone believes is true. It must be possible to provide a bibliographic citation to a published reliable source that says this. Your personal knowledge or belief is not enough.
I personally know that this information is false. Isn't that good enough to remove it?
Your personal belief or knowledge that the information is false is not sufficient for removal of verifiable and well-sourced material.
Is personal communication from an expert a reliable source?
No. It is not good enough for you to talk to an expert in person or by telephone, or to have a written letter, e-mail message, or text message from a source. Reliable sources must be published.
Are there sources that are "always reliable" or sources that are "always unreliable"?
No. The reliability of a source is entirely dependent on the context of the situation, and the statement it is being used to support. Some sources are generally better than others, but reliability is always contextual.
What if the source is biased?
Depends on how it is used. In all cases, Wikipedia articles are required to be neutral, so Wikipedia cannot simply parrot a biased or non-neutral source's opinion as if it were fact. But, sometimes "non-neutral" sources are the best possible sources for supporting information (with due weight) about the different viewpoints held on a controversial subject.
Does every single sentence need to be followed by an inline citation?
No. Only four broad categories of material need to be supported by inline citations. Editors need not supply citations for perfectly obvious material. However, it must be possible to provide a bibliographic citation to a published reliable source for all material.
Are reliable sources required to name the author?
No. Many reliable sources, such as government and corporate websites, do not name their authors or say only that it was written by staff writers. Although many high-quality sources do name the author, this is not a requirement.
Are reliable sources required to provide a list of references?
No. Wikipedia editors should list any required sources in a references or notes section. However, the sources you are using to write the Wikipedia article do not need to provide a bibliography. Most reliable sources, such as newspaper and magazine articles, do not provide a bibliography.


Regarding the current event templates[edit]

It should be mentioned on WP:RSBREAKING that the current event templates are "not intended to be used to mark an article that merely has recent news articles about the topic; if it were, hundreds of thousands of articles would have this template, with no informational consequence." These points have been discussed and debated extensively on Template talk:Current, Wikipedia talk:Current event templates and elsewhere well before WP:RSBREAKING was added in December 2014.[1] And I see no mention of this specific issue addressed on the original discussion at Wikipedia talk:Identifying reliable sources/Archive 45#Breaking news. So the original consensus on Template talk:Current should still prevail. Thanks. Zzyzx11 (talk) 11:51, 2 July 2015 (UTC)

I don't see any value in adding this here. When and how to use a template is normally documented in the template's /doc page, not in sourcing guidelines. WhatamIdoing (talk) 16:48, 20 July 2015 (UTC)

Proceedings, Festschriften, industry journals, and masters theses[edit]

Three related issues, and suggestion of what to do about them:

  1. We do not have anything at all about citations to conference proceedings. Present practice is to cite these exactly like journals, and I see no evidence that they're being treated as less reliable, but they probably should be, on par with PhD dissertations, or perhaps even masters theses (depending on the prestige and exclusivity of the conference and the panel reviewing papers to be presented), since while they're subject to some degree of peer review (enough to be accepted, and enough, in reaction, to be criticized in later publications if they turn out controversial), it's usually less than would be required for publication in a peer-reviewed journal.
  2. How should we treat non-academic, industry/trade journals? Most major (even most minor) industries have multiple trade-insider publications, and their content is usually written by subject-matter experts, but they're not subject to the peer-review rigors of academic journals. My feeling is that these should be treated with the same care as other primary-or-mostly-primary sources, but like academic journal papers are liable to be high-quality sources when they aren't presenting anything controversial. Where they contain material more in the form of journalistic reporting on their field, they should be treated as news sources as long as they're acting independent of the topic (i.e., not just regurgitating press releases, giving all-favorable product reviews, or otherwise acting like house organs).
  3. For a Festschrift/Gedenkschrift/liber amicorum, I think we should treat them (as paper or online publications) exactly the same as journals when subjected to the same sort of peer-review and publication process, but as primary or mostly-primary sources on par with doctoral dissertations otherwise, as in many cases the peer-review process is less rigorous, and/or they may not be published by a university faculty. A specific case has come up in already-written encyclopedia content, at Vlfberht, where an online Gedenkschrift source has been cited. The paper was written by a published expert, on a site devoted to papers by professional archaeologists in honor of one of their regional mentors who died recently. A Gedenkschrift like this is essentially a virtual conference. This is thus about as much peer review as a masters thesis at least, perhaps as much as a doctoral or journal paper; while we don't know the exact criteria, it is written and published by people with academic reputations to maintain, in a narrow field (Scandinavian archaeology), where errors or controversial claims would be noted by colleagues. The specific work in question is straightforward, mostly a matter of gathering raw data about the inscriptions on Viking swords, and drawing statistically-based inferences from it. I've flagged this material in the WP article as (presently written) being improperly cited to a primary source. It's thus probably subject to removal if not re-sourced some other way (I've only found one journal paper that can source some of it, so some material would still be lost). It seems preferable to me to attribute the work carefully, and state in WP's voice that it's a hypothesis by this specific researcher, not to state the claims in the article as if they're known facts. That should be sufficient. The hypothesis presented is noteworthy and relevant, and appears to have influenced a 2012 PBS Nova documentary on the topic, though they did not cite it by name. [I can't find any other source for the information, anywhere, so it seems highly probable.]
  4. How do we feel about carefully attributed (or directly quoted) use of a masters thesis, e.g. "[Researcher_name], in a 2011 [University_name] masters thesis, drew parallels between the results found by [Peer_reviewed_paper_1] and [Peer_reviewed_paper_2]"? (I.e., secondary work, not the presentation of new data, which would be primary sourcing.) I'm especially thinking not of hard sciences (e.g. claims to have mastered cold fusion :-), but rather of non-controversial work in obscure topics in the social sciences, like archaeological, anthropological, linguistic, and mythographic details about which there is very little peer-reviewed material yet, and very little funding for further research (some subjects in these fields get examined once only in journals). Whether good encyclopedia writing or not, common WP practice right now seems to be to simply say "[Peer_reviewed_paper_1] concluded [X], and [similarly|dissimilarly], [Peer_reviewed_paper_2] concluded [Y]", letting the reader connect the dots. It seems better to cite someone else connecting those dots, even if indicating clearly that the synthesis is only from a completed thesis. This would frankly seem far better sourcing that much of what it done in pop-culture topics, where we cite random journalist connecting the dots between two other random journalists' work, without any peer review of any kind other than whether an editor thinks it's due-diligence enough to publish and will interest their newspaper/magazine readers. We regularly quote/attribute primary sources "with caution". Simple attribution in cases like what I have in mind would seem to be sufficient caution, for material that is not controversial, either in making extraordinary claims or contradicting prevailing scholarly views. As a concrete example, I've found a thesis, Prehal, Brenda (2011). "Freyja’s Cats: Perspectives on Recent Viking Age Finds in !egjandadalur North Iceland" (PDF). New York University.  Among quite an array of material, there's a section suggesting continuity/verisimilitude of use, reported separately in previous peer-reviewed publications, of various Northern and Western European words for "cat" as vulgarities, simultaneously in reference to female genitals and to imply male cowardice. I haven't found a peer-reviewed paper that makes the same connection, which is only barely synthetic/analytic (namely that "puss" and its equivalents have a triple use with a long, multi-language history). I actually need the same paper for another pretty obvious, non-controversial synthesis of prior work, this time about cat demographics in historical Scandinavia (about which very, very little has been published in journals, even in the Scandinavian languages; one of the only two papers I can find on this is a doctoral dissertation, with only one journal paper on the subject).

Conclusion: It's important that we include something about academic (and tech, and other industry) conference proceedings and how to approach them as sources, because of the frequency with which they publish material we want to cite. My take: they are primary or mostly primary (unless just summarizing the state of current research, in which case they're tertiary), and should be used with caution, attributed as such presentation, and replaced with secondary, or at least peer-reviewed primary sources in journals, when possible. Tech and consumer conference presentations of new products, technologies, methods, and draft standards should be treated as strictly primary sources. But when inclusion criteria in academic conferences are very stringent, presentations can be treated the same as journals if publicly available in [e-]paper or recorded form. I think what I wrote above about industry journals can easily be directly adapted into guideline wording about them. Finally, the statement that "Masters dissertations and theses are considered reliable only if they can be shown to have had significant scholarly influence" is a bit overbroad. As long as they present nothing controversial, such a paper should be treated like any other primary or semi-primary source, if it"s completed, approved, and published/archived by an accredited university, in publicly available form, or is in the online equivalent of a Gedenkschrift/Festschrift or other scholarly compendium, and it isn't making controversial or extraordinary claims. This would of course not extend to undergraduate papers, unfinished theses/dissertations, and other pure-primary sources that are not from reputable publishers.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:49, 3 July 2015 (UTC)

  • @SMcCandlish: Reliability is related to notability, so for 1 and 2, make sure the proceedings and magazine have their own article under, respectively, Category:Conference proceedings and Category:Magazines. Also, I've noted in conference proceedings that there are different levels of quality control applied in the article selection process, from an accept/reject decision to a fuller peer review. I also sourced how their indexing in bibliographic databases differs from that of journals (e.g., often no impact factors). Fgnievinski (talk) 05:59, 3 July 2015 (UTC)
    • Reliability is tied to reputability of publisher among other factors, including reputation of the authors, and of the reviewers (if any), the target audience, the breadth and depth of the work, the nature of the claims made and their relation to currently-accepted thinking on the topic, and much more. Some of this relates to notability, but does not equate to it. Disreputable publishers can be notable, and reputable ones can surely fall short of notability, since that requires significant coverage in multiple, reliable, independent secondary sources, and not many articles in off-WP publications are written about publishing companies themselves, meanwhile even coverage of universities does not often cover their publishing operations. The notability of an institution as a campus people go to does not automatically make it notable, or reputable, as a publisher; notability of any kind is never transferable. Whether something already has an article on WP doesn't equate to notability either, especially for something like conference proceedings, since we have hardly any articles on any of them at this point (less than 10!). Nothing in WP:V or WP:RS suggests that a source's publisher must have a WP article before we can use that source on WP. The fact that we have a huge category full of articles on mainstream magazines makes really clear how little that relates to reputability; WP:RS categorically considers most magazines to be lower-quality sources by their very nature (though there are exceptions of course, since some publish, e.g., serious investigative journalism, not just mixed secondary-tertiary summary fare.

      Anyway, yes, I myself made the point that quality control differs; that's the essential issue. It even differs for mainstream academic journals, though not as widely. I'm suggesting we account for this, in the guideline, because current WP de facto practice is to cite conference proceedings precisely like journals, other than the template is different. I have yet to find a single case of conference proceedings being challenged on reliability grounds. Admittedly I haven't been searching for it, but in almost 10 years of editing (over 10, counting early anon editing), I should have seen it by now. We have a reliability blind spot here. The impact factors note is interesting (I hadn't noticed that before), but we do precious little with impact factors to begin with, unless one heck of a dispute breaks out. WP relies mostly on secondary sources, but other than literature reviews, most material in academic journals is primary, or a mixture of primary and secondary, that we must "use with caution", because it is presenting novel claims for others to test and develop, and does not yet represent the mainstream, accepted view or even necessarily a noteworthy minority one. Much of it, especially in the hard sciences, is just experimental "noise". The second point of my original post (aside from the blind spot) is the "inverse" blind spot, of treating certain primary sources published as theses by universities as if verboten, when they're really just published primary sources like any others. They're not as useful as fully peer-reviewed ones, but in cases like I've outlined, the idea that they must have had a notable impact is probably too stringent, if we at least directly attribute them, and make it clear what kind of publication they are, in the article text, not just cite them as if authoritative. For some topic like the early history of domestic cats in Scandinavia (one of the above examples), it may well be that there will never be any further research. At bare minimum it should be enough that something in a peer reviewed journal has cited the thesis in question. I think the language we have about graduate theses was written with big, hard-science fields in mind, where new ideas are usually wrong and are very difficult to test. In many of the softer disciplines this isn't really the case, and much of the work in question is rote reporting of data with minimal synthesis. The "caution" needed is lower, and we can probably account for that in the guideline without much effort.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  08:23, 3 July 2015 (UTC)

  • *Groan* This is a huge number of words that ought to just say 'case-by-case basis'. This stuff is so variable from one academic field to another than trying to write general guidelines is hopeless. And anyone who doesn't have a reasonable sense of the landscape in the field they're trying to write about shouldn't be writing about it anyway. How "we" "feel" about individual academic communities' publication norms is irrelevant. Opabinia regalis (talk) 06:32, 3 July 2015 (UTC)
    • I don't think the attitude is necessary, thanks. I never said anything about how we feel "about individual academic communities' publication norms"; you're just making up weird stuff and putting it my mouth. I'm talking about how we decide, internally, to approach particular types of sources in, e.g. anthropology and linguistics (I have one of those degree things in those fields, so you can keep your assumptions to yourself about what I know about their publication norms). I'm fairly confident I can write serviceable guideline text to cover all of this, despite the pessimism. I've written substantial portions of quite a few of our guidelines. I also decline to apologize for writing clearly, and with sufficient detail to cover what needed to be covered to forestall various (though clearly not all) distracting objections that miss the point, and for tying these matters to actual, specific examples (which is almost always required for proposals to go anywhere). Even if we conclude on "case-by-case basis", we don't presently indicate anything to suggest some of these categories of publication be treated on a case by case basis. For some we say nothing (leaving them in limbo, and with conf. proceedings, leading to over-citing), while with others we're being over-inclusive (not distinguishing industry "journals" from peer-reviewed academic ones, which has implications for our tech industry coverage in particular), and for another category we're a bit too exclusive. I'll just draft something I guess, in absence of any constructive suggestions.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  08:23, 3 July 2015 (UTC)
  • Regarding your point #2, could you post a link to an example of the type of journal to which you refer? As things stand, I'd agree that academic journals would be more reliable than trade journals, but even US Weekly is reliable enough. RS is a pass/fail criterion. We only need to establish a hierarchy when sources contradict each other.
As for masters' theses, if they've been reviewed, then why not? Though I concur that "case-by-case basis" is relevant here.
As for establishing rules/guidance/call-it-what-you-like for conferences and to a lesser extent rules for theses, I could support adding specifics to WP:IRS if the Wikieditors are having problems that adding such rules could solve. Are people citing conferences using a confusing mishmash of jury-rigged formats? Are people deleting conference-sourced material because they assume conferences are not RS? Any other problem? These questions are not rhetorical. Darkfrog24 (talk) 21:34, 3 July 2015 (UTC)
How they're formatted isn't relevant; that's a WP:CITE matter, not a WP:V/WP:RS matter. Yes, the problem is that editors will delete citations of conference proceedings or masters theses, or tag them with dispute tags, or remove the entire passage, even when they are used appropriately and cautiously as primary sources, and will do so not because of any controversy about the material but because they believe they're not permissible sources (or that someone should have to argue with them for ten days on the talk page to make a case that there should be an "exception"). This guideline effectively encourages them to do so with regard to theses, and less directly does so (by omitting mention of them at all) with regard to proceedings.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:51, 22 July 2015 (UTC)

Off-the-cuff replies:

  1. Conference proceedings are firmly discouraged, but not banned, at Wikipedia:Identifying reliable sources (medicine)#Other sources. They tend not to have any meaningful editorial oversight, fact-checking, or peer-review. They also tend to have WP:DUE problems, since a lot of things that are presented never get mentioned again.
  2. We should treat non-academic, industry/trade journals exactly like we treat other forms of non-academic media, e.g., political newsweeklies. InformationWeek == Newsweek for reliability purposes. Neither of them are stellar, but both of them are okay for supporting non-extraordinary claims.
  3. We should treat a Festschrift exactly like we would treat any other compilation of academic writings, which means ignoring why it was written and focusing on the editorial process.
  4. If a master's thesis is actually WP:Published, and it appears to meet or exceed the same standards that we would apply to any other source, then you can cite it. If it is not published, then do not cite it. If it is self-published, then follow those rules. If it is published by a publisher with a reputation for fact-checking and accuracy, then treat it like any other source that is published by a publisher with a reputation for fact-checking and accuracy. The fact that the publication was written in exchange for getting an academic degree rather than in exchange for money is really unimportant.

Conclusion: Classification as primary, secondary, and tertiary is largely irrelevant, since it usually changes what a source is reliable for, rather than whether it is reliable for anything at all. We don't need WP:CREEPy rules for every separate type of publications. I recommend focusing on the general principles: fact-checking, accuracy, and editorial control are important. Context always matters. A source can be reliable for a given statement without being the best possible source for that statement (and you aren't required to use only the best possible source). WhatamIdoing (talk) 17:38, 20 July 2015 (UTC)

Thanks for the cogent response. The DUE problem of conference proceedings is precisely the same as that of most news reportage on happenings that do not generate long-term attention in the media: They're both ephemeral, even one-off, but published, and do not benefit from the synthesis and analysis that happens in ongoing re-coverage. The only other salient difference between them is the proceedings material is probably primary, while the news article is more likely to be secondary (depending on its exact nature; was it investigative journalism, or a column/editorial?) Agreed on trade journals, for the most part, but as I noted above, they can act sometime quite promotionally, and this matters for WP:N purposes as well as WP:NPOV ones. Agreed on Festscriften. Theses: Well, yes, it would have to be published, of course. I'm not talking about "Jane Smith gave me a copy of her thesis." Mostly they're published, after some level of peer review (usually within the university) in annual volumes by the university under which they were written, if not done as part of some project that generates something published in a journal. (From my experience with some US universities handling of them; I have no idea what is done in Europe or whatever, nor whether different US institutions do it differently, or don't bother, or what). As far as I can tell, you basically have to use the individual university's library system to find them, and then request one through interlibrary loan or see if they'll send you hardcopy for a fee. From my perspective this qualifies as "publication", since the publication can in fact be obtained, even if it's not "type this into Google" or "buy it on Amazon". There's one I want to use, but it'll end up coming from Scandinavia, and I haven't figured out how to get ahold of it yet. Been a long time since I was at a university, and I don't have access to the full resources of a uni. library, so this is more difficult to pursue than it might be otherwise. I get your point about classification as primary, etc., but what it would be reliable for isn't at issue (in this case, it'd be basically be a table of data and basic interpretation of it, and attributed to the author, not cited as unattributed fact). The data is essentially impossible to obtain without duplicating the research, which was digging up prepublished data, not research in the sense of doing experiments. The concern was that WP:RS is generally taken to imply that a these simply cannot be cited, and this appears to be incorrect to me.
Wikipedia:Identifying reliable sources (medicine)#Other sources should not be extended from medicine to other fields of study, as its statements are certainly untrue of the computer-science literature. We should really discern between peer-reviewed and non-peer-reviewed conference proceedings. Fgnievinski (talk) 22:36, 20 July 2015 (UTC)
I agree that different fields have different standards. Also, "conference proceedings" encompasses a lot of sources. There is an obvious gap, for example, between "Here's the poster that I put up at the conference" and "Here is the peer-reviewed article about the conference, published in a respectable academic journal". WhatamIdoing (talk) 07:11, 21 July 2015 (UTC)
I'm with Fgnievinski on that. I mean "conference proceedings" in the formal sense. I'm not talking about posters, or about coverage of a conference in other publications, or a programme. I'm talking about the publication of the conference, the hardcopy of the material presented at it. Were the papers that were presented peer reviewed in detail? Reviewed to an extent by an acceptance committee on the basis of research quality? Just "reviewed" on the basis of whether they seemed like interesting presentations? There's a big difference between presentation a medical professional conference, TED, and a sci-fi or comic convention, in decreasing order of usefulness as a source about independent facts [con highlights can actually be a good source for things like actors' and writers' own statements about their work; that's not what I'm on about here].  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:51, 22 July 2015 (UTC)

Biased or opinionated sources[edit]

Wikipedia articles are required to present a neutral point of view. However, reliable sources are not required to be neutral, unbiased, or objective.

I have a problem with the unqualified declaration made in the second sentence. I believe it should read, "However, reliable sources do not always have to be neutral, unbiased, or objective. (addition in italics)" My fear is that Randy in Boise will read that simplistic sentence and conclude, "Hey, I can just quote what somebody says as fact, because I got it from an RS!" (I have already encountered such a "Randy".)

I appreciate that "no RS is perfectly unbiased", but this is a special case where we must hold our sources accountable to some of the same standards to which we hold ourselves; at least we have to exercise discretion. The guideline must make clear that opinionated sources are only properly used in the context described: i.e., a topic has been recognized to be controversial, and the source is only to be used (ideally with in-line attribution) to verify the existence of the opinion (subject, of course, to WP:WEIGHT), never the alleged truth of the opinion. There is no way we can build a NPOV encyclopedia if we don't expect the sources for the facts we represent to have both a similar NPOV, and a healthy respect for fact checking; else we are building our house on sand. JustinTime55 (talk) 17:12, 15 July 2015 (UTC)

If you do not already, you need to realize that this Identifying reliable sources guideline is really a somewhat simplified interpretation or guide to policy, not the source which creates the policy. The actual policies involved here are (at least) Wikipedia:Verifiability#Neutrality and Wikipedia:Neutral point of view#Bias in sources. If you'll look at those, I think you'll find that policy allows — indeed, requires — non-neutral or biased sources to be used more broadly that you suggest in your second paragraph, above, for purposes of "fairly representing all majority and significant-minority viewpoints." Per the Verifiability policy, establishing that a source is a RS is a threshold for inclusion of the material in the source, not a guarantee of inclusion, and Neutral point of view and other policies such as BLP and NOT determine or help to determine what should and should not be included, decided ultimately by consensus (in which context you might want to take a look at WP:CONSENSUS#No consensus). Regards, TransporterMan (TALK) 19:08, 15 July 2015 (UTC)
  • The key to understanding NPOV is this: it applies to Wikipeida, not to our sources. We (Wikipedia's editors) must be neutral in our writing... Our sources don't have to be neutral in their writing. Sources can state opinions on the topics they are writing about (in fact, we rely on them to do so). Sources can challenge, dismiss or ignore viewpoints that they disagree with. We (Wikipedia's editors) can not. We must report all significant views... even if we disagree with those views, or think they are wrong. Blueboar (talk) 22:20, 15 July 2015 (UTC)
  • I think that the fundamental problem with this question is that it conflates "neutral" with "accurate". The key characteristic of a reliable source is its reputation for fact-checking and accuracy. They should also have other indications of reliability (see WP:NOTGOODSOURCE for a handy list). If WP:RANDY can actually find a source that has a reputation for fact-checking and accuracy, professional editors, etc., then yes, even Randy is allowed to use it (usually) to present something as a fact. "Bias" isn't on the list: an academic journal may be strongly "pro" on the question of whether humans affect Earth's climate without becoming unreliable.
    On your example, a book about how astronauts' wives worried and had to play the devoted, patriotic wife in public probably does not have a reputation for accuracy about how NASA made politically significant decisions. A book can be reliable for one thing and unreliable for other things. WhatamIdoing (talk) 17:00, 20 July 2015 (UTC)
  • This exactly: 'The key to understanding NPOV is this: it applies to Wikipedia, not to our sources.'  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:14, 22 July 2015 (UTC)
    • Well, as with most bromides it only goes so far - it takes skill and wide reading of sources (generally in the tertiary voice) to recognize what neutral presentation is - you certainly won't learn it by following the partisan source. Alanscottwalker (talk) 00:25, 22 July 2015 (UTC)
Actually, you can learn neutrality by reading partisan sources... you just have to read lots of them, and (as a group) they have to express a wide range' - and ' of partisan viewpoints. Blueboar (talk) 14:08, 24 July 2015 (UTC)
Most unlikely, and not actually well recommended for the anyone can do it encyclopedia. (see, Tertiary Sources section). Original misplaced emphasis, and lack of context is the bane of being neutral and of presenting the unoriginal - so the partisan source which is all about persuading one to place-emphasis-here and buy into a partisan context (or pretext) in ways that may in a solid tertiary presentation be unacceptable, are poor models - we actually have to exclude and not emphasize, that which should not be emphasized, and appropriately contextualize that which must be contextualized. -- Alanscottwalker (talk) 19:21, 25 July 2015 (UTC)

Distinguishing primary, secondary, tertiary material in op-eds, editorials, official blogs, documentaries[edit]

I've drafted some material on how to analyze these sorts of sources. It may be a good supplementary essay topic, if worked up properly, but many of its key points should, in distilled form, be worked into the guideline, to explain how we address these types of sources. I find disturbing the multiple statements in the thread at Wikipedia talk:Verifiability#Are opinion pieces primary or secondary? that come down to a belief that an editorial, or even op-ed by someone notable, appearing in any major newspaper means that the entire piece is a secondary source. That's a failure to acknowledge the difference between fact and assertion, verifiable research and subjective opinion, and amounts to a belief that "fame makes truth". Primary, secondary, and tertiary are not categoric, permanent descriptions of publications themselves (especially multi-department serial ones), nor of publishers, publication types, or media, but of specific content in a specific context. (This confusion stems from errors in WP:RS itself, where it misuses these words to imply, e.g., that newspapers "are" secondary or that journal articles "are" primary, when either may be primary or secondary depending on what content in what part of it is used to source what facts in an article here. We should address these problems separately, and soon. The confusion level across WP about these matters is thick and pervasive.)

  • Op-eds ("opinion editorials" by persons not on the publication's staff) are primary sources with regard to most, often all, of what they contain. They may be treated as secondary for clearly sourced secondary material they contain that is unlaced with opinion, hypothesis, or advocacy. If the writer cannot keep from inserting value judgements or other opinions into every step of the writing, the factual material cannot be extracted from the spin, even if sources are cited for some fact. If they do not cite their own sources, the claims in them, even if presented in a just-the-facts style, cannot be verified. Neither notability nor reputability are transferable. The publisher's own reputability as a source does not infuse by osmosis the analysis or claims made by unedited op-ed writers (this also means that they do not automatically help establish notability for anything they mention. [Selection for publication can help establish notability for the writer, but that's not germane here.]

    An op-ed in the form "[Intro opinional statement] [Bunch of facts and their sources, with any opinional material clearly separated from the factual material] [Advocacy-laden conclusion]" can (not "must") be used (cautiously) as a secondary source for the factual material, and a primary one for the writer's opinion. However, its citation quality as a secondary source is low, even if the writer is a notable expert, because it's not subject to the kind of editorial process that a normal article would be even in the same publication, including bias revision, fact checking, etc. It may also be a low-quality primary source, depending on what it's being used for, even if it's necessarily a high-quality primary source for what the author's own opinion is. Op-eds date quickly, and should be treated as entirely primary sources if things have significantly changed in the topic area to which they pertain, most often a socio-political issue.

    It may be encyclopedically useless in some cases. Example: Some op-ed piece from a non-notable person whose credentials cannot be checked, who is writing about the "likelihood" that hostilities between North Korea and the West will lead to a third world war, cannot be cited on Wikipedia, no matter how reputable the publisher, as evidence for something like "...but others are concerned that these hostilities will lead to a third world war", per undue weight policy. The writer's random-schmoe opinion, which may have been published only for its excellent, poignant wording, is not encyclopedically relevant. An obvious demonstration of this is that if a news outlet publishes some obscure terrorist group's manifesto, they are in fact publishing an op-ed, but that does not confer one iota of reliability upon the claims made by the group, nor any notability upon something they rant about.

    When judging the reputability of the author, consider the reputability or identity of an organization the views of which they are officially representing in the piece. Do not equate the reputability or identity of the organization with those of the writer: A junior policy analyst may produce an inferior piece for a thinktank that otherwise has a high reputation for reliability; a departmental spokesperson may not accurately reflect the views of an organization's leaders; the chapters of a widespread organization may operate with a high degree of autonomy and not have any official position that all branches adopt uniformly. Never attribute to an organization an op-ed or similar statement by someone associated in the source with the organization only for identification or credential-establishment purposes. Misattribution of individuals' writing to organizations is another form of undue weight. Op-eds should be attributed to the author, not the publisher, unless the publication itself is what is relevant in the context.

  • Editorials (by publication staff) are similar to op-eds but written in-house and subject to more editorial control (though less than that of the publication's typical articles). They are primary sources in general (their entire point is to express an opinion). They may be treated as secondary for genuinely secondary material they include, and their need to be specific as to their own sources decreases with the reliability of the publisher and the writer(s), as well as with the level of editorial control they were subject to. The more akin they are to a regular news piece, the more they may shade into a secondary source.

    Caution is still required, because the level of editorial control is usually indeterminate, and it varies a lot. Whether an editorial has a high, middling, or low citation quality depends on the editorial (and sometimes authorial) reputability, and the context, but it's rarely high-secondary because of the nature of the piece, and the lower editorial control than is exercised over regular articles. They may be high-primary on social issues, but never on highly technical matters in any field.

    Three problems that op-eds often have usually do not apply to reputable publishers' editorials: 1) indistinguishable commingling of facts and opinions/allegations/spin; 2) opining on topics a non-notable writer (or notable one in a different field) didn't adequately research; 2) insertion of conflict-of-interest material or irrelevancies. One feature that many op-eds have that editorials usually do not is authorship by subject-matter experts. (Three obvious exceptions are when the topic of the piece is news journalism; when the editorial material is topical, in a section written by a staff member who is a subject-matter expert (e.g. a veterinary department written by a prominent DVM); and when the editorial is by a guest-writer expert and is about their field.)

    Beware treating an editorial as statement of the position of the publisher as an entity. If an editorial represents such a statement it will usually say so explicitly, or the publication history of the publisher will indicate that when it publishes editorials they are intended to be taken this way. It is always safer to write "An editorial in the Madagascar Times stated ...", not "The Madagascar Times stated ...". Editorials are attributed to the publication in most cases, unless the authorship is relevant.

  • Official blogs: Much of the above also applies to news sites' official blogs, which most major newspapers, magazines, and news programs now operate as an outlet for additional staff writing and for the provision of background, followup, and other supplementary material to the main stories they run. Some of them are subjected to a lot of editorial control, but this seems to be rare. Usually there's some (often retroactive), but it may be by a completely different editorial staff with different criteria and less reputability, if any, as an editorial body. The authorship may also be in question for some material (e.g. it may really have been written by interns, but published under the official byline of the staffer for whom they work). Either way, news outlets' official blogs are always subject to editorial policies that the writers are expected to adhere to, so they are not equivalent to self-published blogs. Being non-SPS doesn't confer reliability, it just escapes one particular reliability disqualification.

    As with the above two categories, these are rarely of high quality as citation sources, and are primary sources except where they contain, and clearly distinguish, proper secondary material from primary material. The nature of the content has to be examined on a case by case basis, as it may vary greatly, from personal anecdotes, to fully-developed and -reviewed stories that there wasn't room for in the main publication, to supplementary educational material written for children (always a tertiary source even if well-researched), to just forum-style post-and-discussion material between the writers and the readership. If you're lucky, sometimes it's "here are the sources for my article on the main site". Attribution for official blog material is to the credited author, like a regular article, unless posted by an editorial role account, in which case it may be attributed to the publisher, like an unsigned editorial.

  • Analogous publications: Not only news organizations publish op-eds, editorials, and in-house blogs. Most of the above rubrics apply pretty evenly across other types of publication, though reliability level can vary by reputability of publisher, narrowness and neutrality of the publisher (an academic journal is more reliable in this particular sense than an advocacy organization), topic (a government department may be reliable for census or motoring fatality statistics, if not for the accountability of its work in gathering them, but may be unreliable on a social issue that reliable independent sources indicate it may have addressed inadequately), and editorial process (an op-ed in a peer-reviewed science journal may be subject to little review at all, and "peer-reviewed" applies at the specific article level, not "rubbing off" on every word in the publication; a political or literary magazine may offer op-ed space only to prominent figures, and subject their submissions to a high degree of quality control).

    Non-print media like television and news radio usually do not label editorial and op-ed material as such. Consequently, content has to be examined carefully to determine what sorts of claims it is making, on what authority. Frequently, many "documentaries" are in fact editorials, or are a mixture of editorial (primary) and genuine documentary (tertiary) material. Documentaries that come to novel, extraordinary, or polemic conclusions are primary sources for them; those that come to, or give undue weight to, fringe science hypotheses are not reliable sources (except for very narrow things, e.g. identifying advocates of the fringe science). Quasi-documentaries with a strong reality TV influence, having a behind-the-scenes or follow-our-journey character, are primary sources, like live news reporting. Those that focus on the producer-narrator's views or that of previously published works on which they are based, are video op-eds (or editorials if they represent a production organization's views). Where these materials are not primary, they are almost uniformly tertiary, and should be used with caution, because tertiary video sources, exactly like school textbooks, may gloss over important details, overrepresent a particular view as more widely accepted than it is, have unknown inclusion criteria for what information will be encompassed, and date quickly. An exception is when a video production co-documents something, comprehensively for the medium, along with a published volume, and the combined works are primarily analytical, interpretative, and/or synthetic, not just summarizing, of previous research; in this case, the video version is a secondary source, albeit of lower quality as a citation than the book edition.

Hopefully this analysis can be recycled in some way, after whatever massaging and squeezing it needs.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  23:31, 21 July 2015 (UTC)

I've just started reading this, and I'm already concerned about this sentence: "They may be treated as secondary for clearly sourced secondary material they contain that is unlaced with opinion, hypothesis, or advocacy."
What is this supposed to mean? Imagine that I've got an op-ed from a highly respected general-news source that says, "America should stop wasting money on annual screening mammograms for low-risk women with no symptoms because a systematic review by the highly respected Cochrane Collaboration proved that these tests aren't saving lives, probably because treatment for moderate stages of breast cancer have improved during the last half-century." Imagine that I want to write "Screening mammograms no longer save lives" (if you're curious, early detection only saves lives if you are able to cure early-stage cancer but not middle-stage cancer; when you can cure both, then early detection stops saving lives). The bit about mammograms not saving lives is clearly sourced secondary material, but it is also "laced with advocacy". Is that usable? What if the sentence merely repeated the conclusion without advocacy, e.g., "A systematic review by the highly respected Cochrane Collaboration proved that screening mammograms do not save the lives of low-risk women with no symptoms, probably because treatment for moderate stages of breast cancer have improved during the last half-century, and the "opinion, hypothesis, or advocacy" was in a different paragraph? Would the op-ed now be okay? Or must the entire op-ed contain no opinion (in which case, it's not really an op-ed any more, is it?)? WhatamIdoing (talk) 15:39, 25 July 2015 (UTC)
@WhatamIdoing: Sorry this is long; you encapsulated a lot of stuff to essay-think through in two rather superficially but really pretty different questions. In the first case it would be usable as an attributed opinion to the writer: "Screening mammograms no longer save lives, according to [whoever]". It's a bold claim that glosses over a lot of details, and seems to actually misrepresent the actual research, so WP can't advance it as a statement of fact. Depending on the context, it might not be encyclopedic to include it at all (maybe yes in a medical policy article or section, probably no in a facts-about-cancer-treatment article/section)

In the second, it would be usable attributed to the Cochrane Collaboration, but you probably wouldn't say it that way, but phrase it more like the source did, e.g. "According to the Cochrane Collaboration, screening mammograms have little impact when it comes to low-risk women with no symptoms", preserving the gist without overgeneralizing. You'd still want to attribute it, to C.C. (though citing the op-ed, per SAYWHEREYOUGOTIT), because it's a specific claim based on recent research and [at least with the one source at hand] there's no evidence anyone thinks this research is correct other than the unusually neutral editorial writer, who is passing it on to us like a human signal repeater. The material is technically secondary, but of very low quality as a secondary source, due to lack of editorial control. The problem with op-eds is we cannot ever be certain the writer is independent of the subject, and the publisher doesn't rewrite them much or at all. The respectability of the news publisher doesn't matter much with an op-ed; they're running it because they think it's interesting and will sell newspapers, not because they've had a peer-review committee go over the claims the editorialist is making, to fine-tooth-comb them. Many op-eds contain outrageous statements, and they're run on purpose to generate debate. I was just today going over one at Talk:Race (human classification) and the various counter-editorial reactions to it; rather illustrative case of this sourcing issue. It would be totally wrong for WP to cite that op-ed's conclusion/premise as unattributed fact, though various bits of secondary material in it (e.g. whose prior research influenced this argument) could be treated that way.

With editorial material, it all basically comes back to the "extraordinary claims require extraordinary sources" maxim - the overall reliability of the source (on all the criteria that add up to reliability) is expected to rise commensurate with the weight of the claim. We use editorials and op-eds all the time for common-sense claims, but attribute or directly quote them if we use them at all for more dubious ones. Same goes for recent research that hasn't been reproduced and verified by others. The disappointing disproof this year of the "discovery" of gravity waves in background radiation in the universe, from the first moments of the Big Bang, is a huge case in point. But lots of secondary material in the same paper isn't in question; even some of the primary material isn't dubious (e.g. description of the methodology used) only the data and the conclusion.

Anyway, given a new book on breast cancer research from an academic publisher, or a literature review in a major journal, which factors in the Cochrane work with other research, and comes to the conclusion that the consensus view in the field is that what Cochrane Collaboration said is correct, with that we can probably just state the Cochrane conclusion in our article as a fact and cite the book/lit.rev. and move on (I'd probably also cite the original C.C. paper, but not the op-ed, which wouldn't add anything to the reliability of the claim). Medical articles tend to be stricter about this sort of thing, per WP:MEDRS, than many other topics. If it were an article section on the production of a movie, or whatever, the stakes for being wrong about some detail, like how many days it took to film some scene, are much lower than passing on wrong information about life-threatening illnesses. Technically, we should treat all primary sources evenly regardless of topic, but in actual practice there are too few editors patrolling source usage to ever ensure this is done well in every case. So we set out a "!rule" (this is a guideline after all) that sets a high bar, and it gets "enforced" mostly when the stakes are high. It also matters, because many aspects of primary vs. secondary source usage are matters of actual policy (real rules) when it gets to WP:NOR matters: Only reliable (i.e. not low-quality) secondary sources can be used for WP:AEIS claims. That's why we wouldn't want to use an op-ed at all to source something like this medical claim without attributing it (facts to the research team, politicized conclusions to the writer). By getting at the "advocacy" aspect of op-eds – a form of AEIS, especially evaluation and interpretation, often short on meaningful analysis or synthesis – I'm showing how the AEIS part of the policy applies to such material in more concrete terms. Not everyone is sure what "evaluation and interpretation" are but most of us know politicizing, polemics, persuasion, and activism when we read or hear them.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  13:40, 1 August 2015 (UTC)

I think you need to spend a lot longer thinking about this. Here are a few points that you will need to think about.

  • "Screening mammograms no longer save lives" is not an opinion. It is a statement of fact. The statement may be wrong, but it is still a statement about a fact. It is not a statement of values or opinion. (The opinion part is the claim that a values-based public policy should change as a result of the alleged fact.)
  • The fact that the statement concerns is more than "technically" secondary; it is the result of a systematic review of the scientific literature, which is unquestionably secondary.
  • Almost half of your response is about issues of DUE weight, which are irrelevant to the question of whether Source X is sufficiently reliable to support Statement Y. (DUE is about whether Statement Y should be included at all, not about whether Source X is good enough to support it.)
  • The author of an op-ed is always identified (else you could not know that it is an op-ed, i.e., written by someone who is not on the paper’s editorial board). It is very easy to determine that Chris Conservative is independent of anything about mammogram promotion. In fact, I would hard pressed to think of anyone who would truly benefit from fewer low-risk women undergoing screening mammograms: the net savings probably amount to a couple of dollars per taxpayer per year in the U.S. (a savings that would be seen as slightly lower public debt, not as a few bucks in someone's wallet), and everyone else loses revenue (fewer medical device sales, fewer jobs, fewer profitable surgeries, fewer needless chemotherapy sales, fewer patients to recruit for clinical trials, fewer volunteers to sign up for charity fundraisers, etc.).
  • I disagree with your assertion that "We use editorials and op-eds all the time for common-sense claims". We actually discourage opinion pieces for statements of fact, regardless of whether those statements are about where the Sun appears to rise each morning, or how how many days it took to film a scene, or whether George Bush’s tax policy caused the Clinton boom years, or whether mammograms save the lives of average-risk women. Unless the author is a bona fide expert in the subject, then opinion pieces are poor choices to support statements of fact. WP:RS says that opinion pieces are "rarely reliable for statements of fact". WhatamIdoing (talk) 23:35, 1 August 2015 (UTC)

When to fact tag vs. just delete[edit]

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 ¢ ≽ʌⱷ҅ʌ≼  23:40, 22 July 2015 (UTC)

Need to fix confused wording[edit]

WP doesn't actually care about "a reliable publication process"; that means a process that reliably produces a publication. We need to reword this to make some kind of sense that is the same kind of sense to anyone who reads it. I would strongly suggest we work the word "reputable" in, with regard to the publisher as an entity, and distinguish this from the editorial process.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  23:44, 22 July 2015 (UTC)

Perhaps I am misunderstanding... but I think we do care about whether the publication process is reliable (certainly we don't want to cite a source that has an unreliable publication process)... but your point about "reptutable" is well taken. Perhaps we care about both? Blueboar (talk) 13:58, 24 July 2015 (UTC)
WP doesn't care as a policy matter about the reliability of publishing output processes in the production sense (e.g. if Elsevier or the Wall Street Journal or whoever, due to severe budget cuts, start putting out publications behind schedule, or on crappy paper, that's not of any concern to WP, as long as the quality of the material doesn't weaken. My point was that we're consistently using "reputable" everywhere else to refer to the publisher, and there's been noteworthy confusion in the past with people mixing up reliable source and "reliable" publisher. I think what this passage is trying, so vaguely, to say is something to the effect of "a [something] editorial process at a reputable publisher", or however we'd work that into the sentence, and whatever quality that [something] is ("trustworthy", "respected"?). I'm rather sure we shouldn't use "reliable" there. The problem with doing so is that it produces a circular, meaningless pseudo-definition, "a reliable source is one from a reliable editorial process", when the whole point of the exercise is to nail down what we mean by "reliable", by defining it in other terms relating to the reputation of the author(s)', the quality of the editorial process, the publisher's reputation, and the currency of the material. It just doesn't do anything useful to say a reliable source has a reliable author, a reliable publisher, with a reliable editorial process, putting out material that is reliably not obsolete. :-)  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  12:46, 1 August 2015 (UTC)
Wikipedia very much cares about a "reliable publication process", meaning "a process than produces a reliable publication". It is also not circular, because we do provide other definitions.
If you're finding it confusing (or perhaps only irritating), then the current text: Reliable sources may be published materials with a reliable publication process, authors who are regarded as authoritative in relation to the subject, or both. could be re-written to say something like this: Reliable sources may be published a publisher whose editorial process has led to a reputation for fact-checking and accuracy, by authors who are regarded as authoritative in relation to the subject, or both. WhatamIdoing (talk) 20:52, 1 August 2015 (UTC)

Unpublished/SPS/UGC sources and Template:Cite arXiv[edit]

If what is published at arXiv are preprints and there's no review process but "moderation" to categorize them correctly, that would seem to make them low-quality, primary sources, maybe too low for science/medicine citations at least (maybe okay in the humanities, as directly attributed primary sources used with caution for non-controversial claims and only when necessary. They would seem to have the reliability level of, e.g., a thesis/dissertation, maybe less.

An argument can be made that they're self-published sources, since arXiv is acting as a self-publishing house for academics (in WP:RS terms, if not in intent; I gather their intent is more like that of WikiSource and Project Gutenberg and Archive.org, but narrowly tailored).

Another argument can be made that it's WP:UGC; people all over the world uploading what they've written to a website (with an administrative user class) where it is categorized for public, free consumption ... that's a description of a Wiki.

Exception made for material that is actually published in a journal by the time we need to cite it; the arXiv URL might be the only one we have, if the journal is behind a paywall.

(I asked this initially at Template:arXiv and got nothing in response but a trolling comment, so asking it here might be more productive.)  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  10:02, 24 July 2015 (UTC)

And I'll give you the same answer here as on Template talk:Cite arxiv. We'd want to cite something on the arxiv for the same reason why we might want to cite anything. To establish a claim, a fact, support a thing, or whatever. This is no different than using {{cite web}}. Headbomb {talk / contribs / physics / books} 14:28, 24 July 2015 (UTC)
[plonk]  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:42, 24 July 2015 (UTC)
You could say that for any source, no matter how bad. Are you arguing that articles on arXiv meets the criteria "Articles should be based on reliable, third-party, published sources with a reputation for fact-checking and accuracy."? Because I don't think they do. If they are actually preprints that will be reliably published, then we just have to wait until that happens. The process of reliably publishing them might even produce significant changes. If they aren't going to be published, they're just a paper on a website. Doug Weller (talk) 18:23, 24 July 2015 (UTC)
Indeed. And like anything, they can be cited just as any other webpage on the internet can be cited, subject to our usual restrictions on WP:RS/WP:V. Headbomb {talk / contribs / physics / books} 19:46, 24 July 2015 (UTC)
Arxiv is a simply a repository for pre-prints and does not by itself accord any reliability-points to a publication. As far as I can see, there are only two scenarios underwhich we could cite arxiv prints, :
  1. When the paper has been peer reviewed and published by a reputable journal and the arxiv print is essentially a convenience link
  2. As a self-published source, when the author, subject, secondary citations to the work, claim it is being cited for, etc justify such use under WP:SPS
Btw, WP:RSN may be a better venue for such discussion if there is dispute over providing an arxiv link in particular instances. Abecedare (talk) 18:40, 24 July 2015 (UTC)
The arxiv also has many postprints as well as preprints, btw. Altough as far as I understand SMcCandlish's questions, they just can't fathom why anyone would cite something from the arxiv at all. Headbomb {talk / contribs / physics / books} 19:46, 24 July 2015 (UTC)
@Abecedare: I agree with your two scenarios, and #2 is unlikely to arise much, if ever, given the nature of the material. It's a general concern, not an article-by-article one to me. The entire function of the site is self-publication of preprints in a "document-wiki" manner. I think it's a reasonable question whether something needs to be done with/about this template. The fact that some of what is posted there eventually becomes reliably published (but probably often not in exactly the same form) may not be sufficient if the template is likely to be used to try to cite essentially self-published material, which is pretty much guaranteed with this thing.

We don't need a citation template of this sort. Scenario #1 is already covered by the |arXiv= parameter of {{Cite web}}. This method is "idiot-proof", because the |journal= parameter contains the name of the actual journal it was reputably published in; it can't be used (except in an obviously detectible way, e.g. |journal=arXiv) to improperly cite unpublished [in the WP:V definition of "published"] research. Meanwhile, #2 can be done with {{cite web|title=Nutritional Analysis of Cat Hair|first=A. U.|last=Thor|site=arXiv.org|publisher=self-published|...}}, if it's ever needed. Unless the site itself recategorizes papers after they have been published, and this difference is detectible in the URL so tools can distinguish between published and unpublished and do something (warning? cleanup category?) with citations to unpublished papers, I think this template should be TfDed. It's two legitimate functions are redundant with existing templates, and its potential for abuse is high.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  21:16, 24 July 2015 (UTC)

Here's a scenario under #2 (and there are many more of those scenario where one would cite the arxiv). Grigori Perelman published his Poincaré conjecture in three papers published on the arxiv. arXiv:math.DG/0211159, arXiv:math.DG/0303109 and arXiv:math.DG/0307245. Headbomb {talk / contribs / physics / books} 21:25, 24 July 2015 (UTC)
Already addressed under "Meanwhile, #2 can be done with ... if it's ever needed."  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  22:31, 24 July 2015 (UTC)
As for scenario 1, the solution would be to use {{cite journal}}, not {{cite arxiv}}. Headbomb {talk / contribs / physics / books} 21:27, 24 July 2015 (UTC)
That's exactly what I said. If even you agree that the only purpose for {{cite arXiv}} is citing the unpublished version of something, then we don't need a separate template for that. {{cite web}} exists for a reason.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  22:31, 24 July 2015 (UTC)
I agree with Headbomb here: although appearing on arXiv is not sufficient by itself to turn something into a reliable source, there are nevertheless many reliable sources that appear on arXiv that we would want to cite (scenario #2), and the catchall {{cite web}} doesn't do a good job of it. I would add that there are many instances of both {{cite arxiv}} and {{cite web}} where the reference in question has been properly been published in a journal or a book as well as having an arxiv or web version (indeed I have frequently seen {{cite web}} with a url pointing to the official publication page for a journal paper or to a Google books entry), and these should generally be converted to {{cite journal}} or {{cite book}} with |arxiv= or |url= or |contribution-url= when they are discovered. —David Eppstein (talk) 21:36, 24 July 2015 (UTC)
Please re-examine the argument. You're stating why someone could cite something at arXiv, when the question is why would we do it with this special template instead of the existing ones. Even Headbomb agrees (see above) that {{cite journal}} is the one to use for citation of something at arXiv that has been reputably published. If the only use-case for {{cite arXiv}} is citing unpublished research, we have {{cite web}} for that. The {{Cite arXiv}} template is redundant in either scenario, and provides as easily-abused avenue for source misuse. Due to rampant misinterpretation WP:CITEVAR as a license to WP:OWN every detail of citation formatting, it's extremely unlikely that checking uses of {{Cite arXiv}} on a case-by-case and replacing them with the equivalent, long-standing, general templates to indicate "this has been checked out and is a valid use of an arXiv citation" (using the appropriate replacement template for whichever of the two scenario types the use qualifies as), would not be met with revertwarring. There is thus no clear avenue for patrolling abuse of this template, which serves no purpose anyway. The obvious solution is to deprecate it, convert all of its in situ uses, and then TfD it.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  22:31, 24 July 2015 (UTC)
Please re-read my comment. As I thought I clearly stated, there are some citations that should go directly to arXiv rather than to journals (they are reliable, even though arXiv is their only publication) and {{cite arxiv}} formats them better than the generic {{cite web}}. None of your hyperbolic attempts at justifying the removal of this template would be any different for {{cite web}}; both templates can be abused, but that's not a good reason to remove them. —David Eppstein (talk) 00:14, 25 July 2015 (UTC)
As I've pointed out I think 3 times, we already have a template for precisely the situation you identify: {{cite journal |title=Title_of_paper|journal=The_Real_Journal|arxiv=arXiv_ID_to_generate_the_arXiv_URL|...}}. This even has the benefit of following WP:SAYWHEREYOUGOTIT more closely; you can provide the URL to the official published version (if such a URL is available) while also providing a convenience link to the arXiv copy, which also help guard against linkrot. It's not like |arXiv= wasn't added to the extant templates for a reason. The good reason to remove {{cite arxiv}} is that there is no use case for it that is not already covered by extant templates (among other reasons, already covered). The {{arXiv}} typo: {{cite arXiv}} template doesn't do anything special. It's just another identifier-specific template, which were all deprecated as a class, not because of exactly how they were coded, as you go on about above, but because they're redundant and their purpose was merged into the main citation templates. Same story.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  04:52, 27 July 2015 (UTC)
A further rationale, for at least deprecating this, is prior deprecation of all identifier-based (DOI, JSTOR, PMID, etc.) templates. See Template:Citation Style 1: "Identifier-based templates. ... All of these are deprecated." I'm sure I can dig up the actual deprecation discussion if someone thinks that's necessary. We're already handling arXiv identifiers in Template:Cite journal, making the arXiv template redundant, as well as use of a citation template approach that's already been consensus-rejected.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  22:31, 24 July 2015 (UTC)
That's a completely unrelated issue. {{cite arxiv}} is part of the cite series of templates, something to use within an article to provide a citation. Despite the name, {{cite doi}} is not. Instead, it is a way to tell a bot to set up the real citation elsewhere (usually using {{cite journal}}) and transclude it. That has multiple problems, among them making it difficult to detect vandalism (because the separate citation page is unlikely to be watchlisted) and making the citations hard to edit (especially when the |noedit= option has been used). Additionally, because {{cite doi}} is only used for citations that have dois, which in practice means journal articles, it's redundant with {{cite journal}}. The same goes for the other ones you list. —David Eppstein (talk) 00:20, 25 July 2015 (UTC)
You're kinda making my point for me without realizing it: "Because {{Cite arXiv}} is only used for citations that have arXiv eprint IDs, which is already supported by both {{Cite journal}} and {{Cite web}}, it's redundant with them."  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  04:52, 27 July 2015 (UTC)
So basically, you object to {{cite arxiv}}'s existence because you could use a different template to cite the same page? That's as valid a rationale as to delete {{cite encyclopedia}} because you could use {{cite book}}, {{cite web}}, or {{cite work}}. Headbomb {talk / contribs / physics / books} 03:33, 25 July 2015 (UTC)
It's standard operating procedure to merge-and-delete redundant templates; that's around half of what WP:TFD does (the other half being a mixture of subst-and-deleting extremely-low-use templates, deleting occasional totally inappropriate templates, and a few non-deletion/non-merge discussions). The ones you name, like {{Cite web}}, etc., do particular special formatting and variances of that formatting for particular source types. This one does not (it does do one special linking thing that it shouldn't, which we'll get to). There is no difference at all between these two:
  • {{cite arXiv |last=Sparling |first=George A. J. |eprint=gr-qc/0610068v1 |title=Spacetime is spinorial |date=2006 }}Sparling, George A. J. (2006). "Spacetime is spinorial; new dimensions are timelike". arXiv:gr-qc/0610068v1. 
  • {{cite web |last=Sparling |first=George A. J. |arxiv=gr-qc/0610068v1 |title=Spacetime is spinorial |date=2006 }}Sparling, George A. J. (2006). "Spacetime is spinorial; new dimensions are timelike". arXiv:gr-qc/0610068v1. 
And you can't even do this, when the piece is finally published:
  • {{cite arXiv |last=Sparling |first=George A. J. |eprint=gr-qc/0610068v1 |title=Spacetime is spinorial |date=July 2006 |journal=Nature Physics |volume=XIX |issue=2}}Sparling, George A. J. (July 2006). "Spacetime is spinorial" XIX (2). arXiv:gr-qc/0610068v1.  Unsupported parameter(s) in cite arXiv (help)
because this template doesn't support |journal=, etc. But you can do the following, just leaving |journal=, etc., blank (or omitting them):
  • {{cite journal |last=Sparling |first=George A. J. |eprint=gr-qc/0610068v1 |title=Spacetime is spinorial |date=2006 |journal= |volume= |issue=}}Sparling, George A. J. (2006). "Spacetime is spinorial". arXiv:gr-qc/0610068v1. 
which insta-upgrades to the following when you add the journal info after publication:
  • {{cite journal |last=Sparling |first=George A. J. |eprint=gr-qc/0610068v1 |title=Spacetime is spinorial |date=July 2015 |journal=Nature Physics |volume=XIX |issue=2}}Sparling, George A. J. (July 2015). "Spacetime is spinorial". Nature Physics XIX (2). arXiv:gr-qc/0610068v1. 
(plus other parameters that could be added, like |doi=, that {{Cite arXiv}} can't handle).
So, this is even more redundant than I thought: There is never any need to even resort to {{Cite web}} in any case at all. And it's worse than redundant, because you can't add any new information to it, only totally replace the template. It's just pointless.

The one and only thing unusual that this template does is create links to arXiv "class" categories, optionally, with the |class= parameter. But these do not help the reader find the source, they're just linkspam that fails WP:EL, and they're hardly ever used anyway, not being among the default, recommended parameters. If there were an odd case where this link was actually important, it can be done with |at= in {{cite journal}}.

Conclusion: I still do not see any reason this shouldn't be TFD'd, just WP:AADD arguments. I'm looking for reasons to keep it. I thought "Surely, |class= does something important?" Nope.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  04:52, 27 July 2015 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── The reason {{cite arxiv}} doesn't support |journal= is that it shouldn't. If you're citing the journal article, you use {{cite journal}}. If you're citing the arxiv, you use {{cite arxiv}}, much like you wouldn't ask {{cite book}} to support |journal=. This is a clearly different case than the {{cite doi}} (and {{cite pmid}}/{{cite jstor}}) you keep comparing them to. I don't know what's so damned hard to understand about this. Send it to TFD if you hate it so much, and enjoy seen it being snow kept. Headbomb {talk / contribs / physics / books} 04:59, 27 July 2015 (UTC)

I'm not demonstrating any understanding difficulties here. Please do not assume bad faith "hate" motivations on my part; it's uncivil (has have been several other of your posts on two editions of this discussion, starting with the very first one), and seems to indicate that it's you who are not understanding the conversation (a point I made in my response to your first post, too): I clearly specified what the multiple rationales against this template are, and they have nothing to do with feelings or emotions. I've not taken it to TfD yet because I'm doing due diligence to see whether there are genuine use cases for this, that outweigh WP:RS / WP:V concerns like thwarting of the patrolling of source misuse. Citation templates have been merged without resorting to TfD, anyway. I don't see any WP:SNOWBALL here, just a lot of heated, repeated excuses from a handful of supporters of the template, whose invisible and unexpected toes I seem to have stepped on in some way, but whose objections are easy to counter.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  05:45, 27 July 2015 (UTC)
And yes |class= does something important. It indicates the classification of the paper, and is how arxiv papers should be cited generally. E.g. LHCb collaboration (2015). "Observation of...". arXiv:1507.03414 [hep-ex].  tells you this paper is from the High-Energy Physics Experiments repository. Headbomb {talk / contribs / physics / books} 05:12, 27 July 2015 (UTC)
That's the closest I've seen to a rationale yet, extracted by dentistry. But the entire template could be replaced with a call to {{cite web}} that puts the same repository class link in |at=. This would have the benefit of auto-accepting later additions of |doi=, |journal=, etc., after something has been published (as long as the parameters were passed, of course).  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  05:45, 27 July 2015 (UTC)
Putting |journal= or |doi= on a {{cite web}} citation rather than fixing it to be {{cite journal}} as it should be is ignorant and wrong and if that's the best argument you can come up with for turning other kinds of citations into {{cite web}} then I think we should just close this thread, because it's not going anywhere useful. If you want a one-template-fits-all-citations template, use CS2 and {{citation}} instead of CS1 and the cite template family. —David Eppstein (talk) 06:44, 27 July 2015 (UTC)
Not a coherent argument (even aside from the tone and the handwaving away from the actual point, which is the real reason the discussion has been mired). {{cite web}} supports |journal=, so using that parameter is permissible. These template names are a shorthand, just mnemonics, and all those templates are simply wrappers for what's done with the CS1 meta-templating module and its submodules. There is no policy requiring that a particular template with a certain name be used to format a reference because of how you want to classify the publication. They're just tools to get the job done. It's standard operating procedure to merge them when what they do is redundant. I'm fine if the discussion here stops, though, since the case for merging these is clear, and this is no longer an WP:RS question.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:04, 31 July 2015 (UTC)

The above discussion just makes me glad that I don't use citation templates in the first place... I manually type my citations the "old fashioned" way, using the "<ref>citation information</ref>" format. Then I don't have to worry about parameters, null fields, or any of that sort of stuff. Just saying. Blueboar (talk) 11:19, 27 July 2015 (UTC)

I see the draw, but unless you are tightly managing an article, the citations will end up in radically divergent formatting if the templates aren't used. That's the only reason I bother with them. It's annoying to have one citation begin "Onie-Maus, Ann (2005). "Paper 1". Journal ...", and the next "Ann Onie-Maus, PAPER 2, Journal, 2005 ...". Hopefully in 5 or 10 years (or better yet, next week, ha ha), we'll just have some expert-system bot cleaning them up automatically.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  12:14, 1 August 2015 (UTC)
Even if you use citation templates, that's going to happen over time. WhatamIdoing (talk) 21:00, 1 August 2015 (UTC)

RS/AC[edit]

Earlier today, a discussion I was in reminded me of an edit I'd intended to make for some time. It seems like it should be fairly straightforward as a clarification - if something applies to "most," it should also apply to "many" and any similar qualifier. However, it was then suggested that (because of my participation in that same discussion) it looked like I was trying to influence that discussion. I suppose that's fair enough, so I won't pursue this myself and would like to present it for the editors here to evaluate. Thanks, Sunrise (talk) 05:30, 1 August 2015 (UTC)

Seems more like a WP:WEASEL matter. "Many" has no objective definition, unlike "most" (>50%) or "all". If we mean "multiple", "at least 17 {{as of|lc=y|2015|08}}", or whatever, then we should just say without hyperbole what we mean. — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  12:16, 1 August 2015 (UTC)
′if something applies to "most," it should also apply to "many" and any similar qualifier′ does not make sense when the "most" concerns consensus, i.e. general agreement amongst a group. "Many" is absolutely vague, it can be interpreted (and argued) as any number that is "more than a few," and may be used to indicate a significant minority, as in, "while most agree, many do not." A statement of consensus requires a clearly predominant majority - "most agree," not "many agree." It is troubling that such an edit was attempted; as a participant in the discussion mentioned by Sunrise, it is hard to see how making less specific the most directly relevant, and frequently cited, guideline in that ongoing dispute can be viewed as incidental or helpful.--Tsavage (talk) 06:24, 2 August 2015 (UTC)