Wikipedia talk:No original research

Press releases

I see no mention of press or news releases in the list of primary items in WP:PRIMARY. Are they there and I missed them? Or should I add it? Shawn in Montreal (talk) 15:38, 19 August 2009 (UTC)

These tend to be a grey area... certainly a press release by an individual would be a Primary source, but what about press releases from organizations? Material from organizations have always fallen through the cracks at Wikipedia... they are sometimes considered SPS, and sometimes are not; their webpages are sometimes considered primary sources, but sometimes not. Is a press release from a governmental agency a primary source? Blueboar (talk) 16:29, 19 August 2009 (UTC)

Well, that was certainly my assumption: a press release from any organization -- public or private -- is not vetted by any external group for accuracy. It's delivered to media with the expectation that it be factual, though. Interesting: I just assumed that an organization's internally generated media -- either website or news release -- would automatically be primary. I gather there's nothing automatic about it. Shawn in Montreal (talk) 16:45, 19 August 2009 (UTC)

They can be veted extensively by internal editors, however. It really depends on the organization in question. Which is probably why we don't mention them in the policy itself. People don't like policies that say "It depends". Blueboar (talk) 17:48, 19 August 2009 (UTC)

A press release is SPS by definition. It is certainly a source about what a person or org says about themselves or their opinions about others. They are generally not useful as RS about other people and things - if the press release contains vetted facts by some other RS, use that RS instead.

They can be RS though, for instance a gov't dep't issues a press release that says "This road project was done 15% under budget because the contractor, Joe's Acme Pavement, used a new paving machine from Big Manufacturer." This says several things that would be RS in several places:

1. The road project came in under budget, fact could be used on the road article if it has one.
2. Joe's Acme Pavement, if they have an article, it sources their ownership of the machine, and that they are a gov't contractor.
3. Big Manufacturer, or the machine they make, sources that their machine gets used, etc.

OTOH, if the Lawn Liberation Front issues a press release that they have removed the sod from a local golf course, it would only be RS for the claim, not as fact, even if the sod was removed.

And gov't dep't aren't all that either, if the N Koreans said their new Type Long Dong Missile flew 2500 miles, I'd wait for the Japanese or the Pentagon before quoting it as fact, rather than claim...

SchmuckyTheCat (talk)

A press release isn't de facto self-published. The White House does not "self-publish" its press releases, for example. They are primary sources, but not self-published ones. The latter typically refers to one guy in his basement churning out his latest opinion with no legal or editorial oversight. SlimVirgin ^{talk|contribs} 21:31, 19 August 2009 (UTC)

Can you explain what definition you are using for "self-published" that does not include press releases? I would think that press releases are the epitome of self-published sources, as they are written and published with no external review or control. — Carl (CBM · talk) 21:34, 19 August 2009 (UTC)

A self-published source is one with no editorial or legal oversight. The reason we require reliable sources as a project is that we can't afford to do our own research, pay for an editorial board, hire our own lawyers, buy our own libel insurance. We therefore use sources who have all these things, in the hope that that process has weeded out bad material. That includes primary sources (e.g. press releases from the White House) and secondary sources (e.g. articles in The New York Times about those press releases).

When we use sources that do not have any of the above, we take a big risk, editorially and legally. Therefore, we severely restrict their use. A typical example of a publisher who has none of the above would be one guy who produces a blog from his home. Or me, writing this post. Those are examples of what we mean by "self-published source," as the policy on self-published sources indicates: "Anyone can create a website or pay to have a book published, then claim to be an expert in a certain field." SlimVirgin ^{talk|contribs} 22:02, 19 August 2009 (UTC)

I know about our sourcing policies. I am only trying to understand your comment "The White House does not 'self-publish' its press releases, for example.", in light of the White House page where press releases are published.

I have never seen "legal oversight" as a requirement (WP:SPS doesn't mention it, or define "self-published source" at all). Are you saying that if I publish my own book, but have a lawyer look at it first for libel, the book is no longer self-published? There is certainly no external editorial control over companies' press releases. — Carl (CBM · talk) 22:19, 19 August 2009 (UTC)

No, I am saying that a self-published source is someone who publishes without editorial and legal oversight -- "oversight" is a key term here, meaning that someone can tell you "no, you're not allowed to say that." If you're trying to pin down a strict definition of "self-published," you won't find it, because it's fuzzy round the edges, but the thrust of the term implies no oversight i.e. no-one standing between you and publication. As with this post from me now -- I want to publish it now, and I do publish it now: that's the essence of self-publication. The further a source strays from that framework, the less "self-published" it is. SlimVirgin ^{talk|contribs} 22:27, 19 August 2009 (UTC)

Who is it that tells the White House what they are and are not allowed to say in their press releases? Who stands between them and the publication of press releases on the White House website? — Carl (CBM · talk) 22:36, 19 August 2009 (UTC)

An entire battery of staff—legal, political, and editorial—and a strict editorial process, ensure that no single person or tiny group gets to publish exactly what it wants, whenever it wants, using the name of the White House. The president himself is subject to it. It is the same with every professional organization in the world. SlimVirgin ^{talk|contribs} 22:46, 19 August 2009 (UTC)

So: if I write a book, hire a large staff to vet it, but pay all the press expenses myself on the press I own, the book is no longer self published? That's a stretch. We cannot count internal editorial policies when deciding if something is self published; they count as part of the "self". What matters is external editorial control. For example, the author of a journal article does not control the editor of the journal, much less the referee. The White House controls itself and its staff, without external control, and so even if there is internal approval for a press release, it is still self-published. It would be hard to think of a reasonable situation where a self-published work does not have the internal support of the organization that publishes it. — Carl (CBM · talk) 22:54, 19 August 2009 (UTC)

Press releases are no different than any other kind of source. There is no OR as long as the source is regurgitated with fidelity / is not misrepresented. But not violating NOR is not a carte blanche. A statement could still violate RS. -- Fullstop (talk) 00:27, 20 August 2009 (UTC)

(to CBM) According to your argument, every article published by The New York Times, and by every other news organization functioning independently in a democracy, is "self-published," because not controlled "externally." SlimVirgin ^{talk|contribs} 23:33, 20 August 2009 (UTC)

I thought about that some yesterday, and came to the conclusion that the problem is that our policy page is not clear that some self-published sources (e.g. New York Times articles by their own staff) are OK while others (books put out on vanity presses) are not. But it seems to matter what the topic of the source is. A NYTimes article on the NYTimes itself is again the sort of self-published source that we have to be skeptical about, because we cannot presume that the NYTimes is an objective source about themselves. — Carl (CBM · talk) 23:41, 20 August 2009 (UTC)

Which is another reason why I have long thought that NOR goes off track in discussing source typing (in terms of Primary, Secondary and Teriary) instead of appropriate source usage and misuse. There are so many examples of where the distinction isn't clear. But I know that is a dead horse, so never mind. Blueboar (talk) 01:38, 21 August 2009 (UTC)

The distinction is pretty clear.

Self-published means there is no one standing between you and the page.there is no one standing between you and the page. You want to publish something, and that's what you do. You don't have to argue with a copy editor, have the lawyer say he refuses to sign off on it, worry that you won't be covered by your libel insurance, have the boss say, "god, no, you can't say that." Self-published material means you want to publish X and you do publish X, either by yourself (as with this post here by me) or by paying a vanity company.

The closer a source is to that model, the more likely it is to be a self-published source. It's a fuzzy concept, so there are going to be hard cases, but in the almost five years I've been editing Wikipedia, I've not encountered a single example where I couldn't quickly and easily decide whether something was self-published.

A primary source, on the other hand, is an entirely different concept, which we've discussed many times, so I won't rehash it here in detail. But the two ideas — self-published and primary — have nothing to do with one another. If I'm directly involved in an incident and write a book about it that is published by Random House, the book is a primary source about the things I have direct knowledge of, but is not self-published. If I've studied an incident I have no direct knowledge of, and write a book about it that no one will publish, and have to publish it myself, it's a self-published secondary source. Wikipedia would use the first, but not the second, except in very limited circumstances: we do prefer secondary sources to primary ones, but we strongly dislike self-published sources, primary or secondary.

The example mentioned earlier, a White House press release, is a primary source, when writing about material it's directly involved in, or where it represents those directly involved, but clearly isn't self-published, because the White House employs ranks of professional people answering to the chief press secretary to make sure it doesn't publish nonsense. SlimVirgin ^{talk|contribs} 00:14, 22 August 2009 (UTC)

Well... what about the Legal Notice section of a newspaper... this is a regular section of a newspaper that contains what are effectively press releases from the court or an attorney (announcing, say, a forclosure sale). Is that primary or secondary? Or what about the Obituaries... some are written by staffers, but others are submitted by the public... so is the obit pabe a primary or secondary source? Blueboar (talk) 00:40, 22 August 2009 (UTC)

The author of a White House press release is the White House itself, and there is nothing that stands between the White House and publication – if they want to publish something, that's what they do. There is no other entity beyond the author (which is the White House itself) that prevents them from publishing whatever they want. If you want to say "self-published sources by non-professional authors" or "self-published sources that have not been vetted by a lawyer", then that's another thing. But an IBM press release, a General Motors press release, a White House press release, are all "self-published" based on the plain meaning of those words. They are written by, attributed to, and published by the same entity. — Carl (CBM · talk) 00:35, 22 August 2009 (UTC)

There is no such thing as the White House qua author; there are professional people who work for the White House who write things. As I said earlier, your argument means that everything published by the New York Times is self-published. It's not an argument that's sustainable. SlimVirgin ^{talk|contribs} 00:40, 22 August 2009 (UTC)

If I cited a White House press release, the author I would list is the White House (same for a Ford press release). This is not a system where an independent author writes something and then submits it for review; rather it is a system where the entity itself (White House, Ford) decides what it wants to publish, writes it, and then publishes it. You are right that this makes many NYTimes article self-published as well; if our policy page might suggest not to use these articles because they are self-published, it's just an error in the policy page. But there is a difference with some NYTimes stories: they have a byline, unlike press releases, which are authored collectively by the organization that releases them. — Carl (CBM · talk) 00:47, 22 August 2009 (UTC)

A New York Times editorial with no byline is not self-published, not when it's about other issues, and not when it's about itself. Not ever. That's simply not part of the definition of "self-published" by any reasonable standard. SlimVirgin ^{talk|contribs} 00:55, 22 August 2009 (UTC)

"Published by the same entity that produced the content" is a perfectly reasonable definition of "sef-published". Moreover, the definitions you have proposed – such as "there is no one standing between you and the page" – also apply, because there is nobody else standing between when the NYTimes publishes something they have produced themselves. But the situation is even more clear for press releases than news stories. Actually, a newspaper editorial with no byline is a perfect example of a self-published source, since the content by convention represents the opinion of the publication itself, and there is nobody else who controls what the publication itself puts on its editorial page. In terms of "self-publication" alone, this is no different than if I write up my own opinion and photocopy it. — Carl (CBM · talk) 01:00, 22 August 2009 (UTC)

Let's look at an example of something that is not self published. When I write a journal paper, I have no idea if it will ever be accepted. There is no expectation that editors or referees will look upon if favorably, or that they will work with me to get it into a publishable state. They may reject it with no explanation at all. Thus, when a paper does get accepted, readers have some confidence that there has been a neutral, genuine review of the suitability of the paper. This cannot be assumed for things like press releases, SEC filings, advertisements, editorials, etc. — Carl (CBM · talk) 01:10, 22 August 2009 (UTC)

(ec) Okay, if you really are arguing that a New York Times editorial is a self-published source, then you've produced your own reductio ad absurdum. (Plus, you don't understand the process involved in publishing editorials.) By your standards, pretty well everything is a self-published source.

I can only repeat: "self-published" = no professional editorial oversight. SlimVirgin ^{talk|contribs} 01:14, 22 August 2009 (UTC)

I take "self-published" to mean what it says: something published by the same entity that authored it, with no other editorial oversight. Nothing in the term "self-published" refers to internal editorial oversight (since that's part of the "self"). So a large company can self-publish things just as well as an individual person. Yes, this means that many news stories are self-published; I don't see why that actually makes much difference. Your argument seems to suggest that a company's SEC filings are not "self-published", leading to the question of who else is doing the original publication. — Carl (CBM · talk) 01:20, 22 August 2009 (UTC)

(outdent)

I too think that press releases are self-published. A press release is merely the written variant of announcements from a press officer or a publicist.

This does not imply that self-published sources (like press releases) are necessarily unreliable. For example, a press release that says a company was founded on such-and-such date is (probably) a reliable source for that snippet of information. But it would not be a reliable source on (say) a claim that it was the inventor of the wheel.

I suppose one could say that a degree of scepticism and background knowledge is necessary when dealing with truth claims (in any kind of source), but this is probably especially applicable when an organization/person is talking about itself (as in a press release). -- Fullstop (talk) 18:22, 23 August 2009 (UTC)

Blue sky

This may sound like I'm deliberately creating a ridiculous question and thereby disrupting Wikipedia, but I'm just playing devil's advocate here. Would looking up at the sky to see if it was blue and then saying it was blue on Wikipedia be original research? Someone the Person (talk) 01:49, 21 August 2009 (UTC)

No... but that is more because there are so many reliable sources that say the sky is blue that it is accepted common knowledge. In most other cases personal observation is considered OR. Blueboar (talk) 03:47, 21 August 2009 (UTC)

Picking up the gauntlet: What if a source says the sky is black with specks of glittery thingies? Or that a "red sky at night is a sailor's delight"? How do these jive with the "accepted common knowledge" that the sky is blue?

As we know, if challenged, the proponent of "the sky is blue" has to provide a source that says so in context.

Thus, the answer to "Would looking up at the sky to see if it was blue and then saying it was blue on Wikipedia be original research?" would be: Yes, WP:NOR policy applies to that too.

My two cents. -- Fullstop (talk) 18:35, 23 August 2009 (UTC)

There once had been something sometimes referred to as the "apple pie exception"; see e.g. Wikipedia talk:No original research/Primary v. secondary sources discussion/Archive 4. Шизомби (talk) 23:37, 23 August 2009 (UTC)

Whose argument is it?

I came here after discussion with SLR about the definition of antisemitism and whether it is being applied consistently. Through this discussion I am coming to realize that it would be a synthesis violation for me - as a contributor - to conclude that a double standard was being applied if one man criticized all Jews for something (but was not called anti-semitic) while another man was called anti-semitic for making precisely the same criticism. I gather that what seems like putting 2 and 2 together is not permitted by our NPOV and OR policies.

However, if a contributor discovers a reliable published source which puts 2 and 2 together like that, then their argument may be included (quoted or summarized) in the relevant article. Suppose we find something like this:

• Professor Bea Al Fayer of the Vanderbilt Theological Institute said: "Amos and Hosea criticized the entire ancient nation of Israel for being faithless and said that the ancient Hebrews deserved to have their nation destroyed as divine punishment. Yet, when Sun Myung Moon wrote the same thing in his Divine Principle, Rabbi Rudin said that Moon's remarks prove that Moon is anti-Semitic. Clearly, this is a double standard.

Assuming Al Fayer is prominent enough to have his views mentioned, then we could cite his reasoning, right?

The reason I am asking this is partly because I want to be sure I am not making a conflict-of-interest mistake by inadvertently making our article biased in favor of Sun Myung Moon, in response to his opponents. (I've been a follower of his for 32 years.)

The other reason is that if I can get a clear understanding of when and where arguments must be attributed, it will help a lot of us write more clearly and with less bias on politically charged controversies in the environmental and life sciences. --Uncle Ed (talk) 16:05, 21 August 2009 (UTC)

Why the emphasis on 'events' for primary and secondary sources?

"Secondary sources are at least one step removed from an event" and "primary sources are sources very close to an event", we are told here. Is there meant to be an implication that this distinction does not apply to sources used in articles about theories, movements, populations, techniques, inventions, patents, formats etc etc? I wouldn't bother to ask except that the use of that word in this policy does not seem to be helping in a current discussion at Talk:Office Open XML#Potential sources. --Nigelj (talk) 12:56, 23 August 2009 (UTC)

You are right that the present wording is aimed more at historical articles and not so much at technical articles. Regarding the specific issue of OOXML, it will probably be more productive to evaluate sources in terms of potential for bias and self-interest rather than in terms of primariness and secondariness. If you would like someone uninvolved to help mediate that discussion, let me know. — Carl (CBM · talk) 13:34, 23 August 2009 (UTC)

Thanks Carl. If there's anything you can think of to help move things along at OOXML, please do join in. I wonder if we can get a more general word or phrase into this policy too. --Nigelj (talk) 15:12, 23 August 2009 (UTC)

I'll look through the OOXML page and leave a comment. Changing the wording here would be difficult, and is not really worth the effort, since we are always free to evaluate sources for an article in whatever way is most appropriate for that article. The advice here will be relevant to some articles more than to others. — Carl (CBM · talk) 15:25, 23 August 2009 (UTC)

it will probably be more productive to evaluate sources in terms of potential for bias and self-interest rather than in terms of primariness and secondariness ... well said. We should add something link it to the policy... suggest:

• When it comes to original research, it will probably be more productive to evaluate sources in terms of appropriate and inappropriate use than in terms of primariness or secondariness. Blueboar (talk) 15:41, 23 August 2009 (UTC)

Recent edits

If people want to make substantive changes, could they please propose them here first? SlimVirgin ^{talk|contribs} 02:16, 27 August 2009 (UTC)

What about that change makes it substantive? It's a great effort to remove a giant block of needless text that has no substantive content. Frankly, it's rather confusing, and belongs in an essay. "Here comes the original synthesis" indeed. If you don't have an objection beyond "you can't change anything until you check if someone can think of an objection at the talk page", then just don't revert.    M   02:36, 27 August 2009 (UTC)

It's an excellent example of WP:SYNTH, you'll need to find consensus to either remove it or replace it. Dreadstar ☥ 02:41, 27 August 2009 (UTC)

Rich Farmbrough's points ("Remove the Smith Jones example per WP:POL because one example is enough. Also it is an example discussing sourcing which can confuse") are perfectly valid. What exactly does it illustrate that isn't already clear, and which justifies 4 extra paragraphs and 300 extra words? I can't see anything, but if you do please let us know.    M   02:51, 27 August 2009 (UTC)

These are excellent and longstanding examples. I can't imagine the policy would be better off by removing them; quite the opposite, in fact. Also, I see no consensus for making these major changes to longstanding policies. Jayjg ^(talk) 23:43, 27 August 2009 (UTC)

Agree with Jayjg. JN466 00:20, 28 August 2009 (UTC)

The Smith-Jones example is an example of bad logic. It doesn't serve to show that synthesis to prove a point is improper in Wikipedia (I think such synthesis is improper), it just shows how bad logic can be misused to "prove" something that it doesn't prove. We moderns are at a disadvantage when it comes to logic: we aren't educated in the various forms of syllogism nor in the various logical fallacies. It seems to me to be a grave error to use an example of bad logic as a claimed example of why synthesis in general is improper. If the use of logic is appropriate for a topic then I suggest that any proper synthesis that helps further understanding within that topic is not just valid but is desirable. "Proper" synthesis would be synthesis that follows valid logic and which is based on well-accepted premises. I can appreciate that some would think this can go to far. I'd welcome examples that show "going too far" (with the provision that the example not really be an example of invalid logic masquerading as valid logic.) Minasbeede (talk) 14:39, 28 August 2009 (UTC)