User talk:SuperHamster/CiteUnseen

Previous discussion

Documentation on Cite Unseen was previously hosted on meta; refer to previous discussion at m:Talk:Cite Unseen. ~SuperHamster Talk Contribs 22:01, 20 December 2020 (UTC)

Media Bias Fact Check is a one-man operation that we have repeatedly dissed and is highlighted as a "Generally unreliable source".

User:SuperHamster, I notice the script uses Media Bias Fact Check, but this is a one-man operation that we have repeatedly dissed and is highlighted as a "Generally unreliable source". The better one is the Media Bias Chart by Ad Fontes Media. They are a large, well-trained, team effort. Please make that switch. -- Valjean (talk) 23:26, 20 December 2020 (UTC)

@Valjean: Completely removing MBFC categories is on the bucket list, though Ad Fontes doesn't really fulfill what we're still using MBFC for. The recent update was our first step to moving away from MBFC: the previous biased source icon, which was based on what MBFC classifies as being very far left or right, was removed in favor of the new advocacy + RSP icons (and what Ad Fontes would also be most useful for). The questionable and conspiracy/pseudoscience categorizations (excluding "mild" ratings) were left in, at least for now—while still subject to the same issues of course, I think they're a bit more grounded, and track numerous unreliable sources that are too uncommon and niche to appear in either RSP or MBC. Let me know if you know any other more reputable sources that could help out here.

That all being said, I'm also not necessarily opposed to removing the remaining MBFC categories right now, or disabling them by default and having them be opt-in. ~SuperHamster Talk Contribs 03:24, 21 December 2020 (UTC)

I'm glad to see you have been thinking about this issue. Maybe just disable that part for now and see if that helps, even if there might still be a need for something. That something might get filled later with a better alternative than MBFC. Even just a step in the right direction would be good. -- Valjean (talk) 03:31, 21 December 2020 (UTC)

After looking into it some more, I've gone ahead and removed both MBFC categories. Most of the most commonly used sources recognized by MBFC are already marked with advocacy or RSP, and I've seen at least one case where a MBFC categorization influenced an action/discussion on Wikipedia, which is what we want to avoid. ~SuperHamster Talk Contribs 01:26, 25 December 2020 (UTC)

They seem to have 8 people. I did a big look at this because it was fun. And because I am trying to work out whether it's possible to give editors (especially newbies) a way of rating a page's truthiness. I live in Australia and am no way related to anyone that does not speak Strine

@Valjean Have you an example where they are incorrect? And does Wikipedia:Reliable sources agree ?

They passed "The Daily Mail" test https://mediabiasfactcheck.com/daily-mail/ and they like wikipedia and the reviewer says they a wikipedia editor and has not corrected the accusations by a competitor on their page. Hats off to them.

The difference in size in companies (see below where I rough counted and tried to exclude advisory boards) seems about countries covered, whether they are corporate, sell detailed information about news sources to advertising companies, whether it's a web protection tool or browser extension as well, and whether they are ranking individual journalists. They all seem to be disliked by the extremes, which is good. The critical articles mentioned on wikipedia are all linked via employment or publication to Poynter which has International Fact-Checking Network and journalism courses

NPR mentions this study which used a combination of Newsguard and Mediabiasfactcheck .

https://mediabiasfactcheck.com/about/ 8 @SuperHamster (A github project has a cleaned csv)

https://adfontesmedia.com/team/ 50 +

https://www.allsides.com/unbiased-balanced-news Allsides had 25ish, but has a Chrome extension partner called https://our.news/about/ which has 7

https://www.newsguardtech.com/about/team/ 50 plus Wakelamp d[@-@]b (talk) 11:53, 15 November 2021 (UTC)

Replacing government icon

We're currently using the icon to the left (classic building with columns) to denote state-run sources. The downside is that it's easy to confuse with the Internet Archive logo. Planning to replace it, if anyone has any suggestions for a good icon. ~SuperHamster Talk Contribs 23:41, 20 December 2020 (UTC)

Good catch. It does remind me of them. -- Valjean (talk) 23:41, 20 December 2020 (UTC)

Maybe one of these can be used. -- Valjean (talk) 23:43, 20 December 2020 (UTC)

Maybe something like , , , or would work. Zoozaz1 talk 00:07, 21 December 2020 (UTC)

The second one looks pretty good. A flag is a universal symbol used by nation states. -- Valjean (talk) 01:04, 21 December 2020 (UTC)

As much as I'm enthused at the idea of using the anarchist black flag to represent government, I think there are other versions of civic building icons that could fit here: (the same icon we use in maps). I think it's the long colonnades in the current version that most resemble the Internet Archive logo. czar 21:46, 17 January 2021 (UTC)

 Done Went ahead with Czar's suggested map icon. My main concern with the flag is that I don't think it's strongly correlated with government, and many may think of it as a source being "flagged" as problematic rather than simply a state-controlled source. ~SuperHamster Talk Contribs 06:16, 17 April 2021 (UTC)

Symbol for TV programs?

What about one for this?

• Ferguson, Sarah (June 6, 2018). "Trump/Russia: Follow the Money". Four Corners.

Valjean (talk) 23:57, 20 December 2020 (UTC)

@Valjean: Nice idea. One question is whether we want to limit it to TV programs, or just have an icon for all videos (including sites like YouTube). In terms of categorizing, looks like most news sites have a path or subdomain just for their video/TV content (bbc.co.uk/programmes/, cnn.com/videos/, video.foxnews.com, abc.com/shows, etc.), so matching most common TV programs by URL should be feasible. If we want to be inclusive of all videos, we can probably string match reliably on a few paths, such as /video/ and /videos/. ~SuperHamster Talk Contribs 05:22, 21 December 2020 (UTC)

I'd keep YouTube separate and reserve it only for official media channels. -- Valjean (talk) 08:02, 21 December 2020 (UTC)

Makes sense. Perhaps sites like YouTube and Vimeo can have their own icons, or just be a part of social media.

Here are some icons we've got on Commons that could work for TV programs and other official media channels:

Test Citation

Test Citation

Test Citation

Test Citation

Test Citation

I'm leaning towards either of the first two, as they're simpler and easy to recognize at a small scale. ~SuperHamster Talk Contribs 01:27, 25 December 2020 (UTC)

@SuperHamster Just wanted to pop this back on the agenda - Draft:Dave Oscillation is a good example of an article having references that at first glance, look 'legitimate' - because the BBC references show the reputable news agency symbol. In reality though, these are links to TV or radio programmes so should either be stripped out of the 'reputable news' tag, or given their own tag. Darren-M talk 11:34, 1 February 2021 (UTC)

@Valjean and Darren-M: Apologies for the delay! Finally got this implemented . Very small list of program URLs documented for now (ABC, BBC, CNN, PBS, Fox News), hoping to expand it soon. Cheers, ~SuperHamster Talk Contribs 06:18, 17 April 2021 (UTC)

Tabloids

Here are some tabloids for the script:

• https://www.intouchweekly.com/
• https://radaronline.com/
• https://www.usmagazine.com/
• https://www.dailystar.co.uk/

In a format for categorized-domains.json:

"https://www.intouchweekly.com/",
"https://radaronline.com/",
"https://www.usmagazine.com/",
"https://www.dailystar.co.uk/"

I would also be happy to compile a list of reliable/unreliable sources from WP:NPPSG if wanted. — Yours, Berrely (🎅 Ho ho ho! 🎄) • ^Talk∕_Contribs 12:36, 21 December 2020 (UTC)

And some more for social:

"pinterest.at/",
"pinterest.be/",
"pinterest.biz/",
"pinterest.ca/",
"pinterest.cl/",
"pinterest.co.uk/",
"pinterest.com.au/",
"pinterest.com.ec/",
"pinterest.com.mx/",
"pinterest.dk/",
"pinterest.es/",
"pinterest.hu/",
"pinterest.in/",
"pinterest.it/",
"pinterest.jp/",
"pinterest.nl/",
"pinterest.nz/",
"pinterest.pe/",
"pinterest.ph/",
"pinterest.tw/",
"youtube.com/watch?v=",
"tiktok.com/",
"linkedin.com/in/",
"linkedin.cn/",
"linkedin.co.id/",
"linkedin.co.nz/",
"linkedin.co.uk/",
"linkedin.com.ar/",
"linkedin.com.br/",
"linkedin.com.gt/",
"linkedin.com.hk/",
"linkedin.com.mx/",
"linkedin.com.ni/",
"linkedin.com.pk/",
"linkedin.com.py/",
"linkedin.com.sv/"

— Yours, Berrely (🎅 Ho ho ho! 🎄) • ^Talk∕_Contribs 13:48, 21 December 2020 (UTC)

Here's some for Press releases:

"newswire.com/newsroom/",
"webwire.com/",
"prunderground.com/",
"icrowdnewswire.com/",
"en.acnnewswire.com/",
"newsvoir.com/",
"marketpressrelease.com/",
"express-press-release.net/",
"eprnews.com/",
"prurgent.com/",
"theopenpress.com/",
"pressat.co.uk/",
"prmac.com/",
"infoxen.com/",
"gov.uk/government/news/",
"home.treasury.gov/news/",

— Yours, Berrely (🎅 Ho ho ho! 🎄) • ^Talk∕_Contribs 14:10, 21 December 2020 (UTC)

@Berrely: Lovely, thank you! I'll review these and add them in soon. ~SuperHamster Talk Contribs 23:59, 21 December 2020 (UTC)

@Berrely: All have been added except for US Magazine - I'd like to look into this one a bit more, as I don't think it is as "gossipy" and unreliable as other publications. Thanks for the contributions! ~SuperHamster Talk Contribs 01:29, 25 December 2020 (UTC)

Changelog

Changes are now being recorded at User:SuperHamster/CiteUnseen/Changelog, to make it easier for users to track and understand changes. ~SuperHamster Talk Contribs 00:41, 25 December 2020 (UTC)

PubMed

Citations to journal articles that link to PubMed are incorrectly marked as "state-owned media", presumably because PubMed has a .gov domain. For example:

Dietrich, Laura; et al. (2019). "Cereal Processing at Early Neolithic Göbekli Tepe, Southeastern Turkey". PLOS ONE. 14 (5): e0215214. Bibcode:2019PLoSO..1415214D. doi:10.1371/journal.pone.0215214. PMC 6493732. PMID 31042741.{{cite journal}}: CS1 maint: unflagged free DOI (link)

– Joe (talk) 09:36, 29 December 2020 (UTC)

@Joe Roe: Thanks for noting this! I'll update soon, will probably just add an exclusion for PubMed. ~SuperHamster Talk Contribs 18:13, 7 January 2021 (UTC)

 Doneish, made a change so that any references using the cite journal template cannot be classified as government media, to prevent false positives like these. Might still get around to implementing an exclusions list for certain categories. ~SuperHamster Talk Contribs 06:19, 17 April 2021 (UTC)

ResearchGate

Journal articles with a link to ResearchGate are flagged as unreliable, because ResearchGate is listed at WP:RSP. E.g. (from Lawrence Shaw (archaeologist)):

Welham, Kate; Manley, Harry Peter; Shaw, Lawrence; Parker Pearson, Mike (2015). "Google Under-the-Earth: Seeing Beneath Stonehenge using Google Earth - a Tool for Public Engagement and the Dissemination of Archaeological Data". Internet Archaeology. 40. Retrieved 2 January 2021.

But with citations like these, ResearchGate isn't actually the publisher, it's just being used to provide an open access link, which is specifically encouraged by WP:RSP.

The common thread in this and the problem with PubMed above is that the tool seems to only be considering the url parameter. Maybe for {{cite journal}} it should instead, or also, look at work? – Joe (talk) 15:37, 4 January 2021 (UTC)

@Joe Roe: That sounds like a good solution, and/or also look at various identifiers. I'll work on this in the coming days, thanks! ~SuperHamster Talk Contribs 18:16, 7 January 2021 (UTC)

A follow-up after many moons: As CiteUnseen doesn't currently distinguish by journal, I've removed ResearchGate, as I believe it being marked as generally unreliable is more confusing than helpful. ~SuperHamster Talk Contribs 06:43, 31 July 2023 (UTC)

The Guardian

According to WP:RSP, The Guardian is flagged as "generally reliable" (except for opinion pieces). However CiteUnseen marks it as "the status of this source depends on one or more..." as CiteUnseen already marks opinion pieces from The Guardian as opinion pieces, wouldn't it make more sent to mark The Guardian as reliable? — Berrely • ^Talk∕_Contribs 09:37, 18 April 2021 (UTC)

It is generally reliable. -- Valjean (talk) 17:19, 31 July 2021 (UTC)

Tags at the wrong place

{{better source needed}} (and other tags regarding sources) usually go after the end of the reference tag, i.e. <ref>Something</ref>{{better source needed}}. This is also what the template documentation says. Not inside the tags, since this defeats the point of them by hiding them away in the citation... Easy fix, I'd hope. Cheers, RandomCanadian (talk / contribs) 23:32, 1 May 2021 (UTC)

Hi @RandomCanadian: Cite Unseen doesn't modify articles at all or provide any sort of editing automation, GenQuest is doing their own thing there. I'm guessing their mention of Cite Unseen is simply to say that they evaluated the source using Cite Unseen (albeit not accurately since Cite Unseen does not call the removed sources unreliable). ~SuperHamster Talk Contribs 08:01, 2 May 2021 (UTC)

self-publishers

Hey, SH! I'm wondering if it might be worth adding the self-publishers who pop up in a warning template, like Xlibris. They get caught by Headbomb's, but not by CU, so anyone using only CU to scan an article's references won't see it. (I noticed because I have both installed.) —valereee (talk) 17:35, 5 June 2021 (UTC)

Another self-pub is Lulu.com —valereee (talk) 14:34, 21 June 2021 (UTC)

References

Unionpedia.org is editable

@SuperHamster: Unionpedia is used more and more often as source on Wikipedia (and also wikidata). It is a wiki, and therefore should be categorized "Editable" by CiteUnseen. Often it causes circular references. Tomastvivlaren (talk) 20:49, 30 December 2021 (UTC)

@Tomastvivlaren: Added - thank you! ~SuperHamster Talk Contribs 21:01, 30 December 2021 (UTC)

Charleston Daily Mail

Hello. I was wondering if it was possible to make links to the Charleston Daily Mail not flagged as deprecated. The news site source used to be dailymail .com, which is now the Daily Mail per this RSN discussion. As the Daily Mail is deprecated, any matching URLs are flagged even if they are by the Charleston Daily Mail. For example, please see You Get What You Give (album). Thanks! MrLinkinPark333 (talk) 00:55, 21 September 2022 (UTC)

 Done dailymail.com has been removed from CiteUnseen. ~SuperHamster Talk Contribs 06:44, 31 July 2023 (UTC)

Sources you can maybe add

@SuperHamster:

• vk.com (social media)
• filmaffinity.com (social media, listed here as unreliable)
• 9gag.com (social media)
• boardgamegeek.com (forums count as social media, right?)
• nairaland.com (forum)
• habr.com (blog)
• helpdeskgeek.com (blog)
• livejournal.com (blog/social media)
• blog.naver.com (blog, see WP:KO/RS)
• highstakesdb.com (blog)
• nickiswift.com (blog)
• patribotics.blog (url says it all)
• Wikipedia:NPPSG#Nigeria (stuff in "unreliable" section are almost all blogs)
• strongtowns.org (advocacy)
• dogsbite.org, nationalpitbullvictimawareness.org and others (advocacy, also self published)
• crooksandliars.com (advocacy/blog)
• heartland.org (advocacy)
• redice.tv (white nationalist advocacy)
• expose-news.com (anti-vax advocacy)
• heraldweekly.com (front page says "gossip", so likely a tabloid)
• wikihow (editable by anyone)
• rigvedawiki.net and namu.wiki (editable by anyone)
• localwiki.org (self-explanatory)
• planetmath.org (wiki)
• orthodoxwiki.org

Add a new section for sponsored content:

• wired.com/insights
• thestar.com/sponsored_sections
• lamag.com/sponsored
• ctvnews.ca/sponsored-content
• nationalpost.com/sponsored
• amny.com/sponsored
• seattletimes.com/sponsored
• lfpress.com/sponsored
• gq.com/sponsored
• newsweek.com/sponsored (check newsweek.com/insights as well)
• vancouverisawesome.com/sponsored

Sites that have contributor pieces that are considered self-published, but also other articles that may be usable:

• The Hill
• HuffPost
• Forbes
• Entrepreneur
• The Next Web
• Rolling Stone (Culture Council, same thing)

Satire news websites:

• newyorker.com/humor
• theonion.com
• thebeaverton.com
• babylonbee.com
• thedailymash.co.uk
• private-eye.co.uk
• burrardstreetjournal.com
• mousetrapnews.com
• spacexmania.com
• onlysky.media
• (List of satirical news websites)

The Mail on Sunday, Royal Central and New Eastern Outlook has been deprecated

Maybe add stuff from WP:VGRS with red/yellow/green video game icons?

More lists of sources:

• Wikipedia:Record charts (record chart icon with check for reliable ones, X for unreliable ones)
• WP:A/S (maybe a green CD with a check for reliable sources, red CD with an X for unreliable ones)
• WP:CM/S
• wikipedia:WikiProject Dogs/Reliable sources
• WP:KO/RS
• wikipedia:WikiProject Board and table games/Sources (green/yellow/red board game pieces depending on reliability)
• WP:VENRS
• Wikipedia:WikiProject Professional wrestling/Sources
• WP:HORROR/S
• wp:A&M/RS
• WP:WikiProject Eurovision/Sources
• Category:WikiProject reference libraries
• Category:WikiProject lists of reliable sources

137a (talk • edits) 16:17, 22 February 2023 (UTC)

Feel free to strike sources that you added.

More sources:

Techdirt (blog)

Unrevealedfiles.com (blog)

Ratingsryan.com (blog, see Special:PermaLink/1141094265#Ratings Ryan)

Anarchist Federation UK (advocacy)

Healthliberationnow.com (advocacy) 137a (talk • edits) 16:07, 24 February 2023 (UTC)

@137a: Great list + ideas for new categories, thank you! I'll start working on implementing these over the next few days. ~SuperHamster Talk Contribs 16:52, 21 March 2023 (UTC)

@137a: Quite late, but I've added a bunch of your suggestions above, and have also introduced the new sponsored and satire categories. Great suggestions. I haven't trawled through the WikiProject sources lists you linked; that's still on the to-do. Thanks, ~SuperHamster Talk Contribs 06:46, 31 July 2023 (UTC)

Why the stopsign on this source?

At User:Doug Weller/Archaeology and racism (going live today) it shows this journal [1] with a stop sign. I'm confused. I see I can still use it, but of course the stop sign is a bit unfortunate. I don't mind it for YouTube which I'm also using as a source but where I know there's no problem with the particular videos. Thanks. Doug Weller talk 14:58, 13 March 2023 (UTC)

@Doug Weller, it's not the journal it's the link to the copy on ResearchGate (which is on WP:RSP). I reported this before at #ResearchGate. – Joe (talk) 15:39, 13 March 2023 (UTC)

Thanks. Doug Weller talk 16:02, 13 March 2023 (UTC)

@Joe Roe and Doug Weller: I know this is a very late follow-up, but I've removed ResearchGate from CiteUnseen (it'll no longer be marked as generally unreliable). Until CiteUnseen has a way to distinguish journals, it sounds like marking it as generally unreliable is more confusing and inaccurate than it is helpful. Please let me know if this is change isn't desirable. Thanks, ~SuperHamster Talk Contribs 06:48, 31 July 2023 (UTC)

I guess so, but on the other hand it spurs me at least to check the source. Doug Weller talk 08:48, 31 July 2023 (UTC)

@Doug Weller: I've also been thinking about letting users whitelist sources in their configurations e.g. ResearchGate could be marked unreliable by default, but if a user doesn't like that, they can turn off the marker for ResearchGate. That might be a better compromise. ~SuperHamster Talk Contribs 14:39, 31 July 2023 (UTC)

I'd prefer that. Doug Weller talk 15:32, 31 July 2023 (UTC)

 Done! ResearchGate has been restored, and domains can now be ignored per category (see documentation). ~SuperHamster Talk Contribs 17:24, 31 July 2023 (UTC)

Thanks. Doug Weller talk 06:59, 1 August 2023 (UTC)

Radio Free Asia

Why does WP:RADIOFREEASIA appear as an advocacy organization when it's actually a WP:GREL news source? Great script, by the way. Amigao (talk) 23:16, 1 August 2023 (UTC)

@Amigao: Thanks, glad you like the script! Radio Free Asia wasn't being tagged as generally reliable simply because Cite Unseen had fallen out of sync with WP:RSP. I've updated the script so you should now see Radio Free Asia (and many other sources) correctly tagged per WP:RSP. Please let me know if you spot any more discrepancies.

As for being tagged as advocacy: most of the advocacy domains list was actually generated from Wikipedia categories (e.g. I pulled articles under Category:Advocacy groups and similar, extracted their official external links, filtered out links that weren't being used as citations, and then did some manual filtering and checking). Given that RFA is a reputable news source, I think it's a bit of a gray area, but given its mission ("promoting democratic values and human rights") and the note at RSP about attributing its point of view and funding by the U.S. government, I think it's reasonable to keep it marked as an advocacy organization. Happy to consider arguments otherwise. Thanks, ~SuperHamster Talk Contribs 07:11, 2 August 2023 (UTC)

Looks like the script might not also be entirely in sync with everything listed as deprecated at WP:DEPS. Again, it's a very helpful script overall. Great work! Amigao (talk) 00:42, 24 September 2023 (UTC)

CCP newspapers

Not all news sources under Category:Chinese Communist Party newspapers have been classified in CiteUnseen as state media/government but some have been. Would recommend that they all be classified as such. Thanks again for the fantastic and very useful script. Amigao (talk) 12:28, 2 August 2023 (UTC)

Disabling "Categorized References" section

Hello SuperHamster! Thank you for your hard work on CiteUnseen; it is a *very* useful script. I use it as a way to triage sources individually, not holistically, if that makes sense. Is there a way to disable the (useful for many people!) "Categorized References" banner? Best, HouseBlaster^talk 20:16, 12 November 2023 (UTC)

@HouseBlaster: Thank you! Yes, I was about to add an opt-out flag, but just decided to make it opt-in, at least for now. You should no longer see the dashboard, unless you add cite_unseen_dashboard = true; to your CiteUnseen-Rules.js. Cheers, ~SuperHamster Talk Contribs 20:20, 12 November 2023 (UTC)

@SuperHamster: I did not realize you were updating the script as I was writing... sorry about that! Anyways, thank you so much! HouseBlaster^talk 20:21, 12 November 2023 (UTC)

Hah, no worries! I was impressed with how quickly you noticed and reached out. Thanks for using the script. ~SuperHamster Talk Contribs 20:22, 12 November 2023 (UTC)

Deprecated sources

Hello, SuperHamster! It looks like the script is not picking up all listed instances of deprecated sources such as WP:SPUTNIK, WP:RT.COM, and WP:ALMAYADEEN. Could all those be added? Thanks again for the amazing script. Amigao (talk) 21:20, 20 November 2023 (UTC)

China News Service

China News Service has two main URLs:

• chinanews.com: Linksearch en (insource) - meta - de - fr - simple - wikt:en - wikt:fr • Spamcheck • MER-C X-wiki • gs • Reports: Links on en - COIBot - COIBot-Local • Discussions: tracked - advanced - RSN • COIBot-Link, Local, & XWiki Reports - Wikipedia: en - fr - de • Google: search • meta • Domain: domaintools • AboutUs.com
• chinanews.com.cn: Linksearch en (insource) - meta - de - fr - simple - wikt:en - wikt:fr • Spamcheck • MER-C X-wiki • gs • Reports: Links on en - COIBot - COIBot-Local • Discussions: tracked - advanced - RSN • COIBot-Link, Local, & XWiki Reports - Wikipedia: en - fr - de • Google: search • meta • Domain: domaintools • AboutUs.com

It looks like your script accurately classifies the first but not the second. Could the second be added in? Thanks. Amigao (talk) 00:54, 18 December 2023 (UTC)