Wikipedia talk:External links

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

Google scholar / researchgate / etc.

On biographies, are these links suitable external links. Google scholar is just a search result, researchgate merely a form of social networking. —Dirk Beetstra T C 14:19, 25 June 2018 (UTC)

For clarity, do you mean the GScholar profiles (eg)? Nikkimaria (talk) 01:37, 26 June 2018 (UTC)
Yes. What does that tell me about that person except for what Google defined as what they published. —Dirk Beetstra T C 06:12, 26 June 2018 (UTC)
Self published, nothing authorative. I guess you could even leave out lesser publications. —Dirk Beetstra T C 06:18, 26 June 2018 (UTC)
Bumping thread. Dirk Beetstra T C 09:49, 26 July 2018 (UTC)
I think the GScholar profile would tell you the citation metrics (in the panel on the side), which I understand are of importance to select audiences (e.g., tenure committees and other people who believe in magic numbers).
I don't think it would ever occur to me to add such a thing, but perhaps others feel differently. DGG likely knows far, far more about these things than I do. WhatamIdoing (talk) 22:47, 29 August 2018 (UTC)
  • The influence of scientific work can be roughly measured by the amount of amount of attention other scientists in the field pay to the articles published on it in peer-reviewed journals, and this is shown by the extent to which the other scientists cite them in their own published work. (this is analogous in many ways to our basic rule on notability). There are many special cases and caveats, and the data needs to be interpreted by those who understand the subject and its publication patterns. (Again, analogous in many ways to the way we discuss references at AfD) . In academic fields other than the sciences, those that depend upon published books, rather than articles, citations occur in a much more irregular and delayed fashion, and the numbers are usually uninterpretable).
There are 3 databases that record such citations: Google Scholar, Web of Science, and Scopus. It is generally accepted that Google Scholar is not the most accurate, but it is the only one accessible to the general public. The numbers from it are usually about 1 1/2 to 2 times those from WoS or Scopus, because it includes a good deal of less formal material, casting a very wide net, as is typical for Google.This, as well as the open access, makes it in my opinion preferable for WP.
The extent of citations need for notability under WP:PROF depends on the field of science--there is no single value. Interpreting the data requires looking at the pattern--a few heavily cited articles count for much more than any number of low-cited ones. (for example, in experimental biomedicine, there usually needs to be at least one paper with more than 100 citations, but in mathematics the citation density is much lower. ) There are figures that purport to summarize this, of which the best known is h factor, but they are not universally accepted (they tend to over-emphasise the medium-cited papers)
I'm not myself a researcher in the field of developing and validating these metrics, but I have closely followed it since its beginning in the mid 1960s. As you may have gathered, I do have a personal view--I am very strongly convinced in the validity of citation analysis,and I am almost as strongly convinced about the dubious validity of any current single measure--I think the effort to find one is driven by academic bureaucrats.
Therefore, I think it is important to have this information in articles. I usually add it by including a list of the 3 or 4 most cited articles, using figures from Google Scholar.
On the other hand, I think it extremely inappropriate to include it in an infobox. It requires interpretation, and used as a raw number gives an often misleading impression. My advice is that the field should be removed. DGG ( talk ) 00:18, 30 August 2018 (UTC)

@DGG: Let me be clear, I am talking about:

in external links sections, not about references in prose.

The former is 'a freely accessible web search engine that indexes the full text or metadata of scholarly literature across an array of publishing formats and disciplines', the second is 'a social networking site for scientists and researchers', the third is a subscription service from Elsevier. (probably we should also include Web of Science search links in this discussion, we don't seem to have a template for that?? and that one is 'an online subscription-based scientific citation indexing service')

The indices these provide do have a place on Wikipedia, though taken with a care (I would talk about a range, not the number which is too prone to change).  These links have their function in building articles (finding most cited material for a researcher), or reference their indices (where I would go for Scopus and WoS, not Google Scholar), or reference numbers of publications, etc. (again, WoS or Scopus).

My argument is that these, in external links sections, are almost always inappropriate. Google, WoS and Scopus search results, ResearchGate a plain social networking site. Scopus and WoS are not accessible to a whole lot of people (and to those who have, generally only accessible while 'at work'). Much of the information is a duplicate from what is in their official website or on their official profile (barring cases where ResearchGate could be the main web-presence of a researcher per WP:ELOFFICIAL). Often these links are on official profiles or official websites already (duplication of links). Lists of publications are typical on official profiles/websites, and often higher profile articles published by the researcher are even already mentioned in the article. And a complete list of all publications by a researcher is generally only of interest on the very very high profile ones (Nobel prize nominees and similar ground-breaking researchers), generally only a couple of high-profile are for most enough (and those are then mentioned/discussed in the article already). --Dirk Beetstra T C 08:10, 30 August 2018 (UTC)

for use as an EL, I agree that ResearchGate is an inappropriate EL unless its the pprincipal social media site for an individual. If it is, it's permissible. . Scopus and WoS do not meet the EL criteria, because they require payment. GS, however, I consider acceptable there when the research has a GS ID, which provides validation that these papers are by this author, not by someone with a confusable name, because it meets the definition of providing additional information. It's not a nonprofit site, but I think it is still usable for an EL it is completely objective cumulation of data. But for use as an EL the site should just be listed without any attempt to summarize the data, as there is no unambiguous way of doing that.
but GS (and Scopus and WoS) are the only acceptable references for citation data for individual papers in the text of an article, and I consider such information essential content when it is available. As I said, summary figures from any of them are in my opinion unacceptable in the text of an article, unless properly qualified.
I will note that deWP permits the use of a general reference to a database as a reference for giving a list of books, journal articles , etc.(in fact, they use such a reference as if it were a list without giving the specifics, on the basis that the information can be found there. However logical it may be to do this, my understanding is that enWP does not allow this. DGG ( talk ) 20:48, 3 September 2018 (UTC)

@DGG: OK, I think we all reasonably agree on WoS, researchgate and scopus.

For Google, lets go through some:

  • Stuart J. Russell
    • Reference, inappropriate, amounts to o.r.
  • John Lott
    • Google contains way more(>100?) than a list linked on his official website (49). Google includesthis, where I do not immediately see the link to the subject.
  • Gary Becker
    • same with second result, not a paper BY subject

To me still, Google scholar is still a search result, not a human vetted list. And seen examples above, it seems to show incorrect results, which makes me wonder what the number from that actually mean. (I may add examples to this list) — Preceding unsigned comment added by Beetstra (talkcontribs) 04:05, 4 September 2018 (UTC)

Re-ping User:DGG. WhatamIdoing (talk) 23:11, 4 September 2018 (UTC)
Nothing from GS Scopus or ISI is a human vetted list. Any human vetted list involves research, as I will explain a little further below. In the sciences, there are no responsible reliable lists available except in the rare instances where there is a critical scholarly biography or a history of science study. Most of the secondary sources available are uncritical. The most reliable extensive source besides the rare monographs is the Dictionary of Scientific Biography and its supplements. In the humanities there are more likely to be such monographs. In the social sciences I have little knowledge. But there is much less truly reliable biography available for scholars in any field than for literary or artistic figures, or political or military people .
We use what there is, using a reasonable not a perfect standard of RSourcing. We are after all, not producing a scholarly work here, and there is no intetion to claim WP to be a reliable source. There is no point in aiming for truly high academic standards here--even where there is material, the length of time and amount of effort for one always makes a WP biography is a very brief summary. We're just a popular encyclopedia, a convenient ready reference.
GS and Scopus and ISI are prepared by algorithms, with minor human intervention for ISI. But that does not make-them substantially less reliable than manually prepared sources. We've usually assumed that human intervention is needed for accuracy, but this is not the case here--unless the human intervention is exceptionally skilled,(see below for everythingt hat needs to be checked) The errors that you point out are inevitable. In fact, there are others. Your search for Lott on GS used his GS "profile" profile, in addition to the oddity you mention. you will notice that a number of the publications are listed multiple times. I do not know how to account for this, but errors in the profile pages are common. The proper search is for "JR Lott " (including the quotation marks in the search--this is in fact the GS-recommended technique for complete recovery.) [1] and then rely on manually sorting out the ones in different fields. The practicality of this depends on how common the name is. In cases of multiple people with similar names working in the same field, it often requires checking to verify the institutional affiliation. It will inevitably miss some with alternate forms of the name or names transcribed wrong--this is an especially significant problem for the early years of ISI.
ISI uses a relatively restrictive definition of what is indexable. It has broadened over the years. but in earlier years it eliminated many items that academics usually do count as formal publications. In the other direction , GS uses a very broad definition, including many that academics do not count as formal publications. Scopus has usually tried to be broader than ISI, as a selling point. There is no absolute standard that is field-independent. For example, in some fields, ARXIV does count as a formal publication regardless of not being peer-reviewed. When I do a list, I limit it to peer-reviewed formal papers, but selecting which they are is a matter of judgment, and I try to take field into account. For example, in some areas of humanities, almost everything is published in chapters of a collected volume. I usually evade it by listing the 3 or 4 most cited papers and their citations from GS because its most accessible to our readers. , but that's inaccurate by de-emphasising current work. One wrong way to proceed is no tto list the total number of papers-- another the the total no. of citations;; I remove both when I see them
In considering how to count papers, there is also the problem of primary responsibility and first authorship, which is different in different fields, and different periods. You can not go by first or last position, or corresponding author. Many advisors put their name on everything from their group; at the other extreme, Watson was noteworthy for deliberately omitting his name from almost everything and just listing his students' names. Most advisors make sure that each student has one paper as primary author, regardless of who did what. My advisor was particularly generous, and made sure every student had one sole-authored paper. A true distinction of how many papers and which ones a person is responsible for is history of science research, and is unavailable except for famous people, and when it is, is often contested.
In terms of a source for describing a person's field, there is no RS at all. Their web page may be written in a uniform style by the dept of university PR office, or by the person himself, or a mixture. No matter who wrote them, they tend to emphasize current interests---these websites are designed to attract prospective students. A full CV, which sometimes can be hard to find, typically includes everything--the current practice is to make it as long as possible., A citation for an award will emphasize whatever work the award is for, usually npot his current interests. Afull 3rd party biography , even if available, will discuss everything, and may or may not try for a concise summary phrase.
in summary, there's no pooint to try for precision. DGG ( talk ) 18:32, 5 September 2018 (UTC)
That is still no reason to have them in an external links section. Itis exactly the reason why we are not supposed to link to search results. Moreover, we avoid links to material that isobviously wrong.
I stand with the point that all of these, including GS, should NOT be used in external links sections. --Dirk Beetstra T C 18:45, 5 September 2018 (UTC)


I recently used Template:ASIN in a citation for a book that had no ISBN. I was surprised to find that the template expands to an EL that goes to the Amazon web page where I can buy the book. I would have thought this is prohibited by WP:ELNO. Should I be using this template? This was at List of IBA official cocktails, and it's a Further reading, not a ref. Kendall-K1 (talk) 23:17, 2 August 2018 (UTC)

How so? Would you rather have it like {{ISBN}}? Walter Görlitz (talk) 06:08, 11 August 2018 (UTC)
Kendall-K1, the key words in your question are "in a citation". This guideline doesn't apply to citations. See Wikipedia:Citing sources for instructions on how to write a citation.
Or, if you don't want to wade through another l-o-n-g guideline, the short answer is: We don't care. If you find that it's useful to use the Amazon Standard Identification Number, then you should expect it to lead you to Amazon's website, but that's actually okay. The ISBN page does the same thing, only requiring one extra click to get there. WhatamIdoing (talk) 22:32, 29 August 2018 (UTC)
When I say "citation" I meant Template:Citation, not Wikipedia:Citing sources. As I said, this is in a "Further reading" section and is not used as a source citation. It sounds to me like I should not be using either {{ASIN}} or {{ISBN}} this way. But apparently I'm not making myself clear, or am asking the wrong question. Kendall-K1 (talk) 00:14, 31 August 2018 (UTC)
Why shouldn't you use ISBNs and ASINs, as legitimate parts of a citation? The point of a bibliographic citation is to identify the thing you're talking about. Widely used, unique identifiers are an excellent way to do that. WhatamIdoing (talk) 23:06, 4 September 2018 (UTC)
It seems to me they violate ELNO point 5. But as I said, I'm obviously not communicating, so let's drop it. Kendall-K1 (talk) 00:23, 5 September 2018 (UTC)
You are correct that they wouldn't be appropriate for the ==External links== section. But they're not, so those rules don't apply. WhatamIdoing (talk) 17:41, 5 September 2018 (UTC)
The guideline has a statement in the lead about linking to webpages otherwise failing our criteria at ELNO: Besides those kinds of links listed in § Restrictions on linking [ELNEVER], these external-link guidelines do not apply to citations to reliable sources within the body of the article'. (ELNEVER note mine; emphasis original). If you are using something like {{ASIN}} in the context of a reference, you may continue to use ASIN in that reference. There is a preference to use a more universal identifier such as a DOI or an ISBN, but if a particular work does not have one of those, then an ASIN is Good Enough.
If your concern is for a "further reading" section, as you say above, I think you can defend that use in the spirit of that statement there in ELNO (+IAR if necessary). The intent of providing identifiers for works in the further reading section is the same as in a specific citation--to assist in finding the work in question. (Same caveat as above about preferring other identifiers if available.) --Izno (talk) 18:07, 5 September 2018 (UTC)
Finally, someone who actually read and answered my question. Thank you. Kendall-K1 (talk) 01:46, 6 September 2018 (UTC)

Simple stupid EL check

Over at Free Bird, a user provided a YouTube link to a live performance posted by the band's official account at YouTube (which does acknowledge the various other copyrights involved), so I do not it fails ELNO. This seems like a straight up acceptable EL to include, so I wrapped it up in an {{external media}} template in the body. I feel this is all appropriate under this, but the only reason I ask is that seems to provide a way to for editors on song/album/artist articles to use officially-posted performances or music videos within the body of the article. eg: for a song that has an associated music video that is discussed, and the video was posted appropriately by the artist or the label to their YouTube or Vevo channel, there seems no reason not to include that as an external media block within the music video section.

We still have to be 100% clear that the video posted is not a copyright violation, the account owner is whom they claim to be, and so forth, obviously. --Masem (t) 18:27, 22 August 2018 (UTC)

You can also post simple questions about individual articles at Wikipedia:External links/Noticeboard.
There is no absolute prohibition on this. If there were, the template presumably wouldn't have survived its two trips to WP:TFD.
That said, the template makes the link more prominent than a regular entry. This might be reasonable if, e.g., you thought you could make a reasonable case for fair use of that media in a particular section, but you didn't want to actually upload a copy (e.g., due to excessive file size or file format problems). If the case for inclusion was less compelling, then putting that kind of link in the ==External links== section might be more appropriate. WhatamIdoing (talk) 22:39, 29 August 2018 (UTC)
It might make sense to have some inclusion-in-body guidelines similar to NFC allowances, but as this aren't copyvio concerns, this would more guidance than policy that NFC is. For example, I have wrote up many of OK Go's videos which gain attention for secondary coverage (how they did it), so a video link next to this section makes 100% sense. But many popular songs get videos that simply are typical music video fare, so linking these next to the body wouldn't help much. I might draft up some guidelines for this to consider. --Masem (t) 23:14, 29 August 2018 (UTC)
I would recommend that you write that up on the template's /doc page, rather than trying to make a formal guideline for it. WhatamIdoing (talk) 23:04, 4 September 2018 (UTC)

This generally is quite OK, though often these official videos are already available from the official page of the song/album on the artist's official website. --Dirk Beetstra T C 14:03, 2 September 2018 (UTC)

Adding crowdfunding sites to ELNO 4

WP:ELNO#EL4 addresses the problem of GoFundMe sites in the generic, but they weren't really 'a thing' the last time we went through ELNO with a fine-toothed comb. I propose expanding it this way:

The target here is cyberbegging sites whose main purpose is to ask people for money. I do not intend this to cover typical non-profit websites, even though they doubtless contain opportunities for people to make donations, nor typical social media pages, even though someone could post a request for money there. Do you think that this is sufficiently clear? Do you support this expansion? WhatamIdoing (talk) 23:16, 4 September 2018 (UTC)

I support this addition. Did we not already blacklist some of these? --Dirk Beetstra T C 03:09, 5 September 2018 (UTC)
I like the idea, but how do you craft the wording to make an exception for non-profits without making it too wordy or too subject to wiki-lawyering? - Donald Albury 12:20, 5 September 2018 (UTC)
Why should non-profits be exempt? --Izno (talk) 12:39, 5 September 2018 (UTC)
We want links to charity websites such as (even though that site probably contains a page for making donations) but we don't want links to (even if we're certain that it's the same charity). The point is "no GoFundMe" links, rather than "no sites that ask for donations on any page, in any way, shape, or form". WhatamIdoing (talk) 17:59, 5 September 2018 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── How about, with the exception of a single link in the article about an organization to the official website of an that organization, be it for-profit or non-profit, allow no links to any webpage that offers advertising or solicits donations? - Donald Albury 19:57, 5 September 2018 (UTC)

Are you here talking about the WP:ELOFFICIAL exemption? --Dirk Beetstra T C 00:19, 6 September 2018 (UTC)
Yeah, I guess I am. - Donald Albury 01:10, 6 September 2018 (UTC)
Because that's a major change to the guideline, and editors won't stand for it.
  • "No links to any webpage that offers advertising" means that you'd be banning any page that has any advertising on it, e.g., newspaper articles – not just pages with "objectionable" amounts of advertising.
  • "No links to any webpage that solicits donations" means that you'd ban every content-heavy non-profit webpage if they add a tiny little note that says "Click here to donate". We don't want to say that otherwise excellent external links, such as should get invalidated if they have a discreet link to a donation page, or if they add a banner that says you can donate. The point isn't to exclude anything on any normal website. The point is to exclude all pages on cyberbegging websites.
WhatamIdoing (talk) 02:18, 6 September 2018 (UTC)
Which, to me, means that except for the official domain link on the subject page of the crowdfunding Wikipage and some related primary references there, NO other links on Wikipedia should appear, not even as primary references (with very limited exceptions). That is similar to the situation with petition sites (which are blacklisted), which cannot be used anywhere, even not as primary references. --Dirk Beetstra T C 03:54, 6 September 2018 (UTC)
So, do we make a distinction based on identity, restricting access to web sites for organizations identified as crowd-sourced funding and petition sites, or do we base it on behavior, restricting access to web sites based on the presence of solicitations for sales, donations or signatures (i.e., some form of support)? Whatever we choose to do, the policy/guidance needs to be clear and easy to apply consistently. On the other hand, this is Wikipedia. There will be RfCs. - Donald Albury 12:27, 6 September 2018 (UTC)
We restrict links based upon what you call "identity" when a given site has become a problem. (And by "we", I mean "Dirk and the few other admins who maintain the spam blacklist").
Otherwise, we're looking at the content of the exact page that gets linked. For example, there is a critical difference between linking to "" and linking to "". We have thousands of the first type, but we should have none (or almost none) of the second type. WhatamIdoing (talk) 05:41, 8 September 2018 (UTC)
To go back to petition sites (that is, sites that host petitions, it was seen that a vast majority of the additions were of the 'Sign our petition NOW [here], to push (x) to do (y)', and the rest of them 'a petiotion was opened to ...(ref to active petition)' None of that is what Wikipedia is for. Hardly ever the remark is neutral and encyclopedic when direct links to petitions are involved. And that is still the case if you look at attempted additions (as many are blacklisted). Most of these are, even when added in good faith, abuse and unneccesary. However, most of the organisations behind it are decent, notable organisations (who, themselves, do not spam!).
Here, with crowdfunding, I feel it is similar. I don't think that there are many exceptions where linking to a page that solicits money is warranted, and therefore I would consider to enforce not linking to them, like blanket disallowing sites that are designed for carrying such pages. --Dirk Beetstra T C 09:06, 8 September 2018 (UTC)
I agree with you that it's similar in nature to the online petition sites.
I think we are generally agreed that this is the overall desirable goal for now:
  • Ban external links like and (except in truly extraordinary circumstances, no plausible examples of which have occurred to me yet)
  • Very strongly discourage links like and, just as we have for years (although that's technically a matter for a different ELNO criterion).
  • Encourage links like and, just as we have for years, even if those pages contain a small/non-objectionable amount of donation requests, offers to sell something, etc.
  • Still accept, in the articles about Respectable Organization and New Widget, Inc., the official links to and (standard, well-established practice), just as we have for years, even if those pages say "Donate now", "Buy now", "Contribute to us on GoFundMe", or anything else.
Since there have been no objections (and some desire for an even stronger statement), I'll make this change now, and we can adapt it later as needed. WhatamIdoing (talk) 04:15, 15 September 2018 (UTC)
There have been objections. I have not commented because it’s obvious it shouldn’t be added because it’s too broad. It should not be included. Walter Görlitz (talk) 06:17, 15 September 2018 (UTC)
What is too broad, the permitted sites, or the exclusions? If the problem is that the 4th pt is too permissive, I think it can be improved by changing it to
  • Still accept, in the articles about Respectable Organization and New Widget, Inc., the official links to and (standard, well-established practice), just as we have for years, even if those pages incidentally say "Donate now", "Buy now", "Contribute to us on GoFundMe", or anything else DGG ( talk ) 20:39, 16 September 2018 (UTC)
    • The proposal was to add only these three words, not to add the four long bullet points. Walter, it would be helpful if you would explain exactly what you think is "too broad" about those three words. WhatamIdoing (talk) 16:47, 18 September 2018 (UTC)
      • As was stated earlier, we do not want to preclude websites where they appear. If a valid EL is removed because somewhere on the front page is a link to a crowdfunding campaign, the changed guideline has failed. I understand it's direct links to such pages, but I've seen it argued that references are actually external links so not everyone gets it. Walter Görlitz (talk) 17:46, 18 September 2018 (UTC)
        • Sure, people vary in both their competence and their motivation to willfully misinterpret any rule. But I don't think we're likely to end up with a consensus of editors agreeing that "a link to a crowdfunding site, located somewhere on an otherwise valid site" is the same as "a crowdfunding page" itself – especially if they read this discussion, which will be available in the archives. It is clear from this discussion that everyone accepts links to even if that webpage (at any given point in time) happens to contain a link to a Go Fund Me crowdfunding page. It is only the direct link to that is prohibited. WhatamIdoing (talk) 06:04, 22 September 2018 (UTC)
Retrieved from ""
This content was retrieved from Wikipedia :
This page is based on the copyrighted Wikipedia article "Wikipedia talk:External links"; it is used under the Creative Commons Attribution-ShareAlike 3.0 Unported License (CC-BY-SA). You may redistribute it, verbatim or modified, providing that you comply with the terms of the CC-BY-SA