Looking for a ruling on the quality of references which can be used on detail articles, i.e. most of the articles linked to in this template. Placed here because it applies to all.

Question revolves around use of social media sites as reliable sources. It has been previously established that social media can be a reliable source, depending on the context: Wikipedia_talk:Verifiability/Archive_65#Reliability_of_social_media_sources_representing_companies

Further, Wikipedia:Verifiability#Access_to_sources says:

   "Some reliable sources may not be easily accessible. For example, an online source may require payment, and a print-only source may be available only in university libraries. Do not reject reliable sources just because they are difficult or costly to access. If you have trouble accessing a source, others may be able to do so on your behalf (see WikiProject Resource Exchange)."

So therefore, a locked facebook group (and posts within) is no more or less reliable than an open group, and the group being closed is not in and of itself a reason to reject a source.

Item 1[edit]

If "Verified Expert X" uses their verified Facebook account to discuss the topic they are an expert in, then that specific post should be considered a reliable source? How is it any different to the above case study with the Yarra Valley Railway's facebook page, which can be manually verified by checking previous posts for legitimate content?

Item 2[edit]

Where should I list the personal facebook accounts which are controlled by people who are experts in this field as established by previous published works etc? For example, Daryl Gregory, who (in conjunction with others) published the books used as definitive sources for most of these articles. I'd posit that posts made by his facebook account should qualify as expert sources, and therefore be permitted.

So for example, a list could be:

Group A[edit]

URL to Group A

• Person1 & URL to profile page
• Person2 & URL to profile page

Group B[edit]

URL to Group B

• Person3 & URL to profile page
• Person4 & URL to profile page

Group C[edit]

URL to Group C

• Person5 & URL to profile page
• Person6 & URL to profile page

The list needs to be dynamic, with some sort of name submission system if reliable people create their first accounts, or if a person becomes reliable after getting promoted and gaining access to information.

Anothersignalman (talk) 15:49, 19 June 2017 (UTC)[reply]

Discussion[edit]

• "So therefore, a locked facebook group (and posts within) is no more or less reliable than an open group, and the group being closed is not in and of itself a reason to reject a source." I don't have time right now to address everything in full but wanted to address this comment specifically. There is a difference between something being behind a paywall of a reliable source, like Wall Street Journal, for example and a closed facebook group. Whether it is open or closed is mostly irrelevant other than the fact that it is virtually inaccessible to someone IF the administrator decided they don't want that person to have access, whereas a paywall is technically accessible to *everyone*. Second, closed/open aside, it is a Facebook group. Period. It is not a published source. Subject matter experts may very well post there but it is not published material. Period. It also appears to me that you're asking in this RfC to vet individuals on Facebook for the purposes of being able to use them as a reliable source - if there is not published content to support something, perhaps it is not appropriate to be in an article but we as editors do not get to decide who is and is not an SME based on Facebook.

We aren't talking a press release posted on a companies website. We're talking the equivalent of a forum, where any individual can claim to be someone and there is no tight affiliation "officially" with the subject matter.

Lastly, by associating someone as an individual as either an SME or an employee of xyz is a grey area and it forces an assumption of legal authority that you cannot make for them (think where an individual is employed by a company but not a representative and speaks for them on social media - which cannot be verified and has no editorial oversight.)CHRISSYMAD ❯❯❯¯\_(ツ)_/¯ 18:27, 19 June 2017 (UTC)[reply]

A few items to respond to:

1. Access - how is the access or not aspect any different to, say, a language barrier or region-lock barrier? Also, repeated from above:

    "If you have trouble accessing a source, others may be able to do so on your behalf (see WikiProject Resource Exchange)."

2. Published material must form the basis for an article, but I think smaller details shouldn't be as stringent. As long as a source is provided, the reader can decide for themselves whether or not to trust that reference. Also, it's not like this stuff is life-or-death. Context is key, and that's why this sort of discussion DOES NOT set a precedent for, say, anti-vaccination idiots to copy their posts across to here.

3. Yes, I'm asking for a process to vet individual facebook (and forum in general) posters as experts within specific contexts.

4. In this case, the published content does not exist yet, but it will at some unidentified point in the future. Many of the articles refer primarily to the "brief history" series of books by Vincent, Gregory & Bray, but those books are a few years out of date so do not include changes between publication and current. Facebook is the only place that information is currently available, but hunting for individual posts by reputable authors (including Gregory, Leslie et al.) is awful; so when I find one I transfer the details here and include the URL. That's why the URLs are so complex, rather than just the group URL - they each link to the specific comment that makes the claim.

5. Identity theft on Facebook is harder than you seem to think. The posters are identified and you can check previous posts by the person to clarify that they are, in fact, who they say they are. Look for things like consistent posting formats or writing style. When someone has, say, five to ten years of reliable, accurate writings (regardless of medium), chances are they know what they're talking about. There is also an edit trail - if a post is edited, you can click it and see previous versions.

6. Re legal stuff, you're taking this too seriously. If it was say a medical trial or a safety-related issue then sure, you want a source that can stand up to legal scrutiny. When we're talking about the date a set of carriages was assembled, I don't think it matters nearly as much (unless a person was injured during the process, and then you'd wait for official reports to be publicly released).

Anothersignalman (talk) 23:20, 19 June 2017 (UTC)[reply]

You have missed the point entirely, Anothersignalman. Especially 1, 3 and pretty much everything else. We do not vet individual people as SME's based on Facebook profiles, ever. Period. This is not about access of physicality or payment. This is a closed group that is run by someone who chooses who gets to see it. It is not a published source. It cannot be used reliably. CHRISSYMAD ❯❯❯¯\_(ツ)_/¯ 00:05, 20 June 2017 (UTC)[reply]

Two things from there - a) are you using the dictionary or journal definition of "reliable", and b) why does it matter? Anothersignalman (talk) 06:22, 20 June 2017 (UTC)[reply]

You said at some point that published information about x would not be available until 2040 (something). If that is the case, it does not belong in an article, Facebook post or not. CHRISSYMAD ❯❯❯¯\_(ツ)_/¯ 00:05, 20 June 2017 (UTC)[reply]

That specific case has three pages in a published work plus a handful of references in magazines (with editors) from the era. The rest of the documentation is due to be released publicly c.2041. Anothersignalman (talk) 06:22, 20 June 2017 (UTC)[reply]

The key question I'm asking here is, if the information is not currently easily accessible, needs to be, is demonstrably accurate and isn't published in what you deem to be "a reliable source" as yet, where are people supposed to go to find it? Anothersignalman (talk) 06:22, 20 June 2017 (UTC)[reply]

• I would consider Facebook as more reliable than Wikipedia Trainsofvictoria (talk) 06:58, 20 June 2017 (UTC)[reply]

EG: Girgarre railway line page (now deleted). It consisted of 95% made up info, and all the references were to unrelated lines surrounding it. Trainsofvictoria (talk) 07:09, 20 June 2017 (UTC)[reply]

• Nothing stated in a Facebook group can be considered a reliable source, as WP:SELFPUB (and the above WT:Verifiability post) refers only to official Facebook pages of a company. Unfortunately, even if someone is considered a content expert, by posting to either an open or closed Facebook group we would not consider it a reliable source, as those posting are not posting on behalf of a company as an official representative as required by WP:SELFPUB. Allowing the referencing of a Facebook group would be detrimental to one of the project's key policies -- There'sNoTime ^{(to explain)} 12:15, 20 June 2017 (UTC)[reply]

Is it possible that I'm confusing "reliable source" and "official source"? If we only permitted official sources these rules would make a lot more sense; using the term "reliable" indicates that individuals can qualify on their own merits.

As for WP:SELFPUB, scroll up a paragraph and it says:

    "Anyone can create a personal web page or publish their own book, and also claim to be an expert in a certain field. For that reason, self-published media, such as ... social media postings, are largely not acceptable as sources. Self-published expert sources may be considered reliable when produced by an established expert on the subject matter, whose work in the relevant field has previously been published by reliable third-party publications. Exercise caution when using such sources: if the information in question is really worth reporting, someone else will probably have published it in independent reliable sources." - Wikipedia:Verifiability#Self-published_sources 

A) The linked accounts can be (fairly easily) verified as legitimately belonging to the experts.

B) The work of those experts has previously been published, by groups like the Australian Railway Historical Society and Newsrail.

C) The reliable source (Daryl Gregory et al) has not yet had a chance to publish the content in a formal book. When that does happen the references can be changed, but until then the information needs to be collated and publicly accessible somewhere.

The restriction against self-published sources mainly seems concerned with people talking about themselves, and that's perfectly reasonable. I don't really see how it covers events.

Anothersignalman (talk) 14:38, 20 June 2017 (UTC)[reply]

Anothersignalman It appears that you misunderstand WP:V. "The reliable source (Daryl Gregory et al) has not yet had a chance to publish the content in a formal book." Books are not the only reliable source however if something is not yet available as a source, it does not belong on Wikipedia. Wikipedia is an encyclopedia, not a crystalball.

Regarding "The work of those experts has previously been published, by groups like the Australian Railway Historical Society and Newsrail." If that is the case, use that as a source, not Facebook.

"The linked accounts can be (fairly easily) verified as legitimately belonging to the experts." How? CHRISSYMAD ❯❯❯¯\_(ツ)_/¯ 14:48, 20 June 2017 (UTC)[reply]

Agree Wikipedia should not try to predict the future. But that doesn't address events that have happened after the last edition of Book X (in this case, Gregory et al's series on Victorian Railways rolling stock) was published - in this case there could be a time gap of ten years (and that also deals with WP:Recentism). Anothersignalman (talk) 19:26, 20 June 2017 (UTC)[reply]

Misunderstanding here. Gregory et al have had their works published previously, which makes them recognised experts in the field. According to the quote above, that means their more recent (post-publication) work can also be regarded as a reliable source on this specific topic. Say I published fifty articles in various reputable journals, on the use of surgical tape in lieu of masking tape for painting jobs. That would make me an expert in the field. Then, according to the above rule, if I set up my own personal website and kept posting there on the same topic (and the site was verified as mine), then that website would also count as a reliable source. Anothersignalman (talk) 19:26, 20 June 2017 (UTC)[reply]

That's why I linked back to the company example. You can verify to a reasonable degree of accuracy that social media account X is owned by company X, and the same methods allow you to verify that social media account Y is owned by person Y. The verification method is different for individual sites - for example, I'd expect that a blue-tick Twitter account would pass the verified-ownership test. In the case of Facebook, the easiest way is to check the posts of an account over an extended period of time; say a year or two minimum. If they all have a similar writing style, chances are it's one person writing. If they match the style of writing used in previous Published works by the same person, then chances are it is that person. I say the test should look at the last few years of relevant posts because a spam or fake account will most likely be caught within a few days, or a few months in worst-case scenario (on Facebook; different standards would have to apply for other websites). Of course, checking all that is a lot of work which is why I think a register of social media accounts that pass that test would be useful. Anothersignalman (talk) 19:26, 20 June 2017 (UTC)[reply]

• In order:
  1. A locked Facebook group is by definition not a published source. This goes beyond "difficult to access" (and bypasses whether or not it's reliable); it's simply not public. Such locked groups should not be used as references, ever. (If you disagree, I'll start citing experts' notebooks that I store under my bed. If you ask nicely, I'll let you see them when you visit me.)
  2. The short answer is, you shouldn't. If you'd have to maintain a list of "established experts", those experts aren't established enough.

That said, this RfC is both strangely formatted and in a strange place. The template has nothing to do with the questions being discussed here. Huon (talk) 00:10, 21 June 2017 (UTC)[reply]

A) Nearly the entire audience for the pages covered by this template will already have access to the respective facebook groups, and/or can get it. If you have an account, you could request access. It's easier than becoming a paid subscriber to every news outlet with a paywall. B) I'd ask you to provide scans of the expert notebooks, then as a reader of whatever page I'd judge for myself whether those references are legitimately from the people you claim them to be from. C) The experts are established, as seen in their works published in Newsrail and by the ARHS. This is about linking their social media presence to their existing published works, to fill the gap between the last edition of their works and today. D) I placed the RfC thread here because it applies to all the pages contained in the template. Easier than having it repeated across each page. Anothersignalman (talk) 07:50, 21 June 2017 (UTC)[reply]

Please forgive if I am trampling over old ground, but the problem with social media, forums and fan websites is that often there is no vetting process and thus contravene WP:OR policy.

I could set up a Facebook account under the name of someone who is considered a knowledgeable source, spend a couple of months building up my credibility, and then post something which is complete rubbish. If the person who I am pretending to be does not become aware of my actions and seek to have the account closed, then I will probably get away with it and the post stand unchallenged. Granted a bit far fetched and not overly likely in these articles given that we are not dealing with a particularly emotive issue, but it certainly could happen in articles where WP:POV is more likely to occur.

With published sources, while errors and mistruths can still creep in, there is a process whereby the editor / publisher is going to run his eye over the content and if he feels unsure about it, do some digging to check the validity. He's unlikely to take information from an anonymous source as fact. So while not an ironclad guarantee of accuracy, it does make it more likely.

I'm sure we all have extensive knowledge on subjects which would enhance articles, but without being able to back it up with reliable sources, we can't add it. While the no WP:OR policy does hold this back, without it Wikipedia would have no credibility. Beachbo (talk) 04:09, 25 June 2017 (UTC)[reply]

Beachbo, no problem with going over old content, especially since we now have it all in a central location. (Also, I enjoy a discussion that forces me to think, so thanks for that :) )

I think WP:OR doesn't apply here because I'm only talking about using previously-established Expert accounts. The rules already allow an expert's work to be used if their earlier work has been published. I'm not suggesting a free-for-all permit on using social media as a reference.

Yes, you could set up a false account on social media, but that's the same as setting up a false company/volunteer group account. The method we agreed on for the latter was to check the history of previous posts by the same account, and I don't see why that's a problem here. After all, if you're able to maintain a facade of accuracy for "a few months", chances are you've learned enough to hold your own in a discussion on the expert's topic, which covers accidental wrong information. You raise a good point about intentionally supplying false information. To handle that, I'd suggest that "Expert" social media accounts are kept in a register and given benefit of doubt. Then, if one of those accounts is used as a source and can be demonstrably proven false, that opens up the door to removing that account from the register.

I can see your point re WP:POV. I think the easiest solution there, at least for now, would be to ban use of this sub-clause on topics where WP:POV applies. Anothersignalman (talk) 14:01, 27 June 2017 (UTC)[reply]

Anothersignalman No amount of consensus on this or related talk pages overrides policy. No one agreed to anything about vetting nor does any individual involved in this discussion have the authority to do so. Facebook is not a reliable source and a closed Facebook group is not a source. We, as editors, cannot and do not vet individuals on the basis of their Facebook account when it comes to use as a source. End of story. CHRISSYMAD ❯❯❯¯\_(ツ)_/¯ 16:07, 27 June 2017 (UTC)[reply]

Then the obvious answer is to fix the policy, as has happened previously under Wikipedia_talk:Verifiability/Archive_65#Reliability_of_social_media_sources_representing_companies. The result of that discussion was "Conclusion: if social media are reliably associated with official spokespersons for the company, then see WP:SELFPUB", as enacted by Staszek Lem. I'm suggesting extending that rule to include "if social media are reliably associated with recognised experts who have previously had reliable works published in the relevant field, then see WP:SELFPUB". What's wrong with that? Anothersignalman (talk) 18:50, 30 June 2017 (UTC)[reply]

No, the answer is to understand and accept why the policy exists. Think it has been explained fairly extensively above that to be considered reliable, a source needs to be published. A discussion on Facebook, Railpage or any other similar platform fails the verifiability test. Likewise the drawing of conclusions from Vicsig and Flickr images that don’t actually back up what is being stated.

That many of these articles are relying on cites from these sources just makes them vunerable to having whole sections of text deleted or being deleted entirely. Best to have a smaller articles with text that can be backed up by reliable sources than one dominated by sources that are not. An article with 10 reliable cites is far more likely so survive a deletion challenge than one with 10 reliable and 40 unreliable.

The policy isn't going to change on the basis of the discussion on this page, so if you still feel a change should be made, then a consensus to change will need to be reached on the Verifiability page. Beachbo (talk) 02:23, 3 July 2017 (UTC)[reply]

Re photos on Flickr, Vicsig and non-locked Facebook pages - check the links. The photos establish that a vehicle existed at a certain time in a certain livery and/or consist. Photoshop, incorrect captions and/or timestamps all possible but unlikely given context. That's why most of those references in the tables say "before date X" or "after date X" - they establish limits but not exact details. Obviously the ideal scenario would be to find exact details and replace the sources.

Is there a way to segregate claims and associated references into major and minor? Agree it doesn't look good to have the majority of the reference list dominated by facebook links; but it's sort of like how, say, the S type page is largely based on two ISBNs listed 30 times for different page numbers. It dominates the list and gives undue emphasis to those sources.

I think I'll take your advice and start a thread on the Verifiability talk page. Will post a link here when that's done. Anothersignalman (talk) 07:37, 3 July 2017 (UTC)[reply]

Wikipedia_talk:Verifiability#Permitting_use_of_established_experts_as_references.2C_if_posted_on_social_media.3F Anothersignalman (talk) 08:42, 3 July 2017 (UTC)[reply]

• My generic outsider view ... Published is print media, having some assurance of editorial review, and definite durable wp:v. Online articles by commercial sites have editorial checks and better access but may not be durable ... and if behind a paywall then wp:v issue. Facebook... is down at the level of blog in no editorial controls. Individual cases might be wonderful, but I would say avoid or use as last resort. If this is all there is, then maybe the info is too low in detail to have coverage so might be better to say nothing. Markbassett (talk) 00:47, 4 July 2017 (UTC)[reply]

That's all there is, for now. The content will be published properly at some unidentified point in the future, but until then posts on Facebook by previously-established experts in the field are the only option. Agree it's a last resort, but it's still an option. Anothersignalman (talk) 00:55, 4 July 2017 (UTC)[reply]