Wikipedia talk:Applying reliability guidelines

Seeking feedback
This essay was motivated by some recurrent questions and misunderstandings at RSN. The intent is to supplement policy pages (which are very bullet-point oriented) with guidance on identifying which policies are at the crux of an issue in a given case. I could particularly use feedback on two questions: Rhoark (talk) 00:35, 7 April 2015 (UTC)
 * First, is what I've written an accurate reflection of policy? Some of the things I say are, by design, divergent from instances I've seen of policy applied in actual practice; however, they should adhere to policies as written and as applied in best practice.
 * Secondly, is it helpful? The hope is that someone having read it would be less likely to (innocently) engage in policy shopping or forum shopping out of ignorance, and instead cut to the heart of the matter.


 * ".Content that underwent an editorial process is not self-published, regardless of where it is hosted"? A little to emphatic for my taste - there are vanity-publishing mills that claim to 'edit' submissions, and some may indeed do so - thought the 'editing' may consist of nothing beyond a quick run with a spell-checker, and some minor formatting tweaks. To my mind, the distinguishing characteristic of a self-published work is that the substantive content is entirely under the control of the author, and not produced on behalf of the publisher - not subject to 'gatekeeping' as you put it. It is this part of the editorial process that matters here. Maybe the wording can be tweaked? AndyTheGrump (talk) 17:24, 7 April 2015 (UTC)
 * That's a good point. I was not thinking of that sort of editing in-name-only. Perhaps, "Self-publishing denotes a complete lack of editorial gatekeeping, not only a lack of gatekeeping by a web hosting platform." Rhoark (talk) 17:51, 7 April 2015 (UTC)

First, a bit long. I also question this part of the Reliability section: "editorial oversight ... constitutes the principal driver for a source's reliability." That is not true. For example, statements by the Church of Scientology, or the government of North Korea have a great deal of editorial oversight, but rather little reliability. On the other extreme, a self published statement by a Nobel prize winner in their area of specialty could well be considered highly reliable. Reliability is more a matter of "reputation for fact-checking and accuracy" than of process. Then I'm not sure why you called out the Columbia Journalism Review specifically, but not many other sources. Then the statement "Including verifiable material is never original research, even if the effect is to confirm or detract from another claim." is way too strong, as in many cases this is a fine example of WP:SYNTH. Then ... but let me get back to my first point. It is too long to really read, so there are probably plenty of parts to it that could be disputed. --GRuban (talk) 21:26, 9 April 2015 (UTC)
 * I'd make it shorter if I could without leaving out something important. I'll need to add something else about conflict of interest, which (alongside reputation) is what makes North Korea or Scientology unreliable for most claims. I don't know of any meta-source as respected as CJR. If someone can suggest some, they could go in too. SYNTH requires OR. OR requires a novel claim that is not verifiable. If there's no such claim, there's no OR or SYNTH. That is one of the main things that comes up so much it motivated me to write something. Rhoark (talk) 00:10, 10 April 2015 (UTC)
 * You have a section on conflict of interest or bias already; I guess you could say that the only reason NK or CoS are not reliable is that they have a conflict of interest with 90% of the world, but that seems to be stretching the definition of conflict of interest rather wide. I humbly propose thet "poor reputation" is the correct terminology, especially as it is actually used in the text of our guideline.
 * Reading the article on Columbia Journalism Review it seems to be an American magazine with a staff of 8. This is our ultimate source for all meta-reviews? Surely not. Do they even cover non-American sources? This is an outright recommendation for systemic bias: "we only care about the US".
 * The WP:SYNTH link UN/war examples are fine examples of synthesis without particularly novel claims. --GRuban (talk) 18:34, 10 April 2015 (UTC)
 * Reputation is part of it, as I said. Having a financial or legal interest in the claims is also part of it. CJR is very trusted, but I don't mind dropping it. The novel claim in the UN example is the nature of the relationship between claims A and B. Rhoark (talk) 22:20, 10 April 2015 (UTC)

Press releases
Some past discussions on reliability of press releases: Rhoark (talk) 17:54, 11 April 2015 (UTC)
 * To put my own spin on it, press releases are usually not reliable, but for standard reasons such as self-publishing or COI rather than special reasons unique to press releases. Rhoark (talk) 18:02, 11 April 2015 (UTC)
 * I'm not sure that the links you've provided really prove much one way or another - press releases are certainly primary sources, and accordingly have to be treated with care, but we don't categorically rule out ever using them for non-controversial content, and the general reliability of the issuer is certainly a matter to take into consideration. And describing them as 'self-published' is problematic: is a statement from say the BBC any more 'self-published' if it has 'press release' on the top, than if it doesn't, but is placed on their own website instead? Statements cited to press releases should often be attributed, where they are used (which is often a WP:UNDUE issue), but as far as reliability goes, we shouldn't be making definitive statements - it clearly depends on the specifics, and may come down to a matter of judgement. AndyTheGrump (talk) 18:27, 11 April 2015 (UTC)
 * That's pretty much where I'm coming from as well, which is why I objected to stating that press releases are never reliable. It looks like Fuh agrees and has pulled it out. Rhoark (talk) 18:50, 11 April 2015 (UTC)
 * It would clearly depend on context. For example, if the video game company Capcom issued a press release announcing a localized version of Dai Gyakuten Saiban: Naruhodō Ryūnosuke no Bōken (a spinoff of the Ace Attorney series) for a North American release we can use that to confirm that the game was announced to be released in that region since that would be an appropriate use of the primary source.--67.68.209.200 (talk) 04:26, 17 April 2015 (UTC)