5 ways Google Scholar helps you get access to what you discovered
If there is one academic discovery search that dominates it is Google Scholar.
Much has been said about it's merits , particularly over library discovery systems but even the best discovery service will not be popular if it does not help the user access the full text whether open access or based on the user's own unique circumstances (typically institutional affiliation).
In this blog post, I will list 5 different ways Google Scholar helps a user get to full text. The last two were methods I recently discovered and it seems may not be very well known even by academic librarians.
They are
1. Free full text tagged [PDF] or [HTML]
2. Library Links programme
3. Library search via Open WorldCat Search
4. The print/or non-electronic holdings option...
5. Subscriber links programme
Of course with extensions like EndNote Click, Lean Library you can have even more capabilities to find full text, but in this blog post I am going to focus only on features native to Google Scholar accessible to any user without an extension.
Note: It goes without saying Google Scholar has many issues.
There are technical issues that affect it's effectiveness
e.g Due to limitations in search syntax (e.g no nested boolean, limited character length etc) and results display (max 1,000 results) it's not suitable for systematic reviews particularly alone, possibly poor bibliometric data for citing etc
But there are other reasons not to use Google Scholar that have less to do with it's technical specs.
This includes
Possible privacy issues - As you will see below for method #5, while Campus Activated Subscriber Access (CASA) does not require Google accounts, if you are signed on with your Google account while activating CASA, the CASA token will be tied to your Google account.
1. Free full text tagged [PDF] or [HTML]
As I have written before in the past, before the era of OA detection brought on by Unpaywall, Library systems (discovery, publisher) generally ignored or did not take seriously the task of finding and linking to open access copies. At best they had poor unreliable link collections that caused frustration when you clicked on it and get broken links.
I have not so fond memories of asking to activate OA package of journals (roughly 2012ish to 2016ish) using Proquest 360base and later Alma and seeing lots of complaints due to broken links. The DOAJ collection of titles was particularly tricky then.
Things only improved in the last 5 years or so and I noted that our systems started to get serious about linking users to Open Access copies roughly around 2017 as the volume of OA became too big to ignore , and today almost every system whether discovery system, A&I, browser extension, link resover etc uses some form of OA finding service such as OACore Discovery and particularly Unpaywall
OA aggregators like DOAJ also began to take such issues seriously and in a recent tweet their Twitter accout notes there are two staff devoted to watching the issue of delivery.
But what if you were a user who relied heavily on open access (or free to read sources) before this relatively enlightened period we live in today? There were other OA only collections you could use such as OAIster but they were relatively unknown and if I recall had problems with broken links as well.
The answer was simple, you used Google Scholar! Since 2004, IMHO Google Scholar singlehandedly held the flag for the Open Access movement by making such links discoverable.
Whether it was a open access copy available in a preprint server, institutional repository, or a Gold OA article on a journal page, Google Scholar found it all.
Users learnt quickly to look for the [PDF] whateverdomain.xxx link on the right side of the Google Scholar results.
In those days, Google Scholar was unbeatable in finding such links. Using crawler based technologies, they could tell for sure if a PDF was where it was, leading to very reliable linking with few broken links.
It had unbeatable breadth too, as I understand it, it was basically the Google Crawler except it indexed Scholarly works, and it could find papers sitting in the unconventional places like a random author's home page or University home page (not institutional repository).
In the example above I search for papers found by Google Scholar on my university domain but exclude those found on our institutional repository subdomain
Note : the site operator only finds results where the copy is the main/preferred entry but this is good enough for our purposes here
While some are false hits e.g. the first two appear to be reading lists, the next two surface real scholarly unique papers that are on a University page somewhere but not in our institutional repository.
In fact looking at the cluster of papers in the second last result , there are 6 copies, 5 of which are on University domains (but not the Institutional Repository) and 1 is on Academia.edu! For the very last result, Google Scholar can find only one copy and it's the one on the University page!
Even today, this feat cannot be duplicated by modern services like Unpaywall which only crawl repositories, journals etc and not the whole web.
One argument made by some is these extra copies found by Google Scholar are not really open access but are merely free to read because they lack actual OA licenses. This may be true but I would add that this argument does not prevent Unpaywall from detecting "Bronze OA" articles (generally articles found on publisher platforms, free to read but have no open licenses) which would not be considered OA under the same argument.
Another similar debate is whether OA finding services should surface copies found on ResearchGate or Academia.edu. Google Scholar generally takes a liberal view and includes them all.
I have not seen a recent head to head test of Google Scholar vs modern OA finding services like Unpaywall, Core discovery, OAbutton which fuel OA detection behind the scenes for almost all library and research systems, though some of the earliest I have seen gave Google Scholar the nod for higher recall and precision.
The gap probably has narrowed by now, and in a sense even if Google Scholar still proved better, it's lack of openness, APIs etc means it cannot be used by discovery and related services the way Unpaywall, Core discovery are today.
2. Google Scholar Library Links Programme
This is probably something most academic librarians and experienced researchers are familiar with.
After all, this is a feature available from the early days of Google Scholar in 2005!
If you know a little bit of electronic resource management and linking you will know this is Google Scholar's integration with library link resolvers (typically OpenURL).
You will also likely know that Google Scholar has their own little twist on this.
From the user point of view, you will need to go into settings->Library link and add your library. As an aside you can add more than one institution (up to 5).
Depending on your institution's settings to Google Scholar, you may not even need to do this if you are within the IP range (typically Campus IP) specified by your library. I believe if you happen to be signed-in with your Google account this setting will follow you on all devices you have signed-in (this might not be a good thing if you are consicious of privacy)!
Adding your institution to Google Scholar in Library Links
Once this is done, when you search in Google Scholar you will see a link on the right (if you are lucky) with a label like Findit @ XXXX or similar. (Labels defined by your library)
Below shows our it looks in my institution.
Library Links appearing in Google Scholar for my institution
You no doubt know that this means you can access the full text of this article via these links and for those entries without such links? You probably have no access via your institution.
From the backend, your institution/library has probably setup a system where the holdings or entitlements of your institution is sent to Google Scholar periodically (ours is weekly). Google Scholar then uses these information (holdings file) to know when to display these library links (we label ours findit@smulibraries) and when you click on it, it transfers you to your link resolver.
For my institution, alma our library system creates a "publishing profile" that sends holdings weekly to Google Scholar
You might notice this is slightly different from the way the link resolvers work in databases like Scopus , Web of Science. In those platforms, you do not know in advance if you will have access and all entries will have the findit@xxx button or link. This is because such platforms do not ask for or get your holdings files, so it does not know if you have access until you try clicking on the link.
Most platforms like Web of Science (unlike Google Scholar) will display a link next to ALL results, you will only know if you have access after clicking on it.
Today some publishers are pushing for GetFTR which superficially looks like Google Scholar, in that they dynamically show links depending on whether you have access. But as I explain here, the main difference is that while Google Scholar gets the holding information from the library, GetFTR gets it directly from the content owner/publisher. (See this post for a longer discussion)
In this sense, Google Scholar's innovation has made it easier for users to know at a glance what they have access without "clicking and praying" like on most platforms.
Note: There might be other platforms that work similarly to Google Scholar (not counting GetFTR implementations), but I don't know of any currently. Up to Feb 2020, Pubmed supported link resolvers with the Linkout local option which works like Google Scholar library links. Libraries would supply Pubmed with their holdings and this would be used to selectively show link badges on Pubmed platform. However as of writing the only option supported going forward is Linkout "outside tool" which provides a link next to all records, similar to most platforms
When I was a new librarian at my former place of work, we lacked a link resolver and we constantly got complaints about why we couldn't do Google Scholar links (and no Proxy bookmarklet doesn't cut it for APA journals), when I finally managed to get the link resolver working properly and integrate with Google Scholar, it was one of the most satisfying moments in my professional working life because it clearly made a difference in the lives of our users.
Even today, when Google Scholar occasionally rebels and "forgets" part of our holdings and the library links disappear next to articles, I see emails coming from faculty (who have rarely contacted the library) asking why.
3. Library Search via Open WorldCat Search
Library Search option in Google Scholar
As you search in Google Scholar, you may sometimes see the link option "Library Search" appear in the row of links below a search result.
If you click on it, it goes straight to OCLC.
I don't fully understand how this option works but it appears to be a way to work around the fact that the library links work only for journal articles and conference papers but not for books (I think not even ebooks but I'm not 100% sure) , journals or other content.
Google Scholar's help file on this says
Library search offers links to book catalogs where the patron can check local availability of a book, or request a library loan. This program works best for print resources, such as textbooks and monographs. To sign up, you'll need to participate in a union catalog, such as OCLC's Open WorldCat. We also work with many other national and regional union catalogs. Participating union catalogs make their bibliographic records available to our automatic indexing system. We index these records and link to the union catalog pages, which, in turn, normally link to the catalog of your library. You don't need to register with us for library search. We link to union catalogs, and the union catalogs link to individual library collections.
In theory if your institution is one of those that upload your local holdings/local catalog to OCLC (see Alma procedure etc), this could allow you to be quickly redirected to an OCLC page with your institution holdings showing.
If you find this option useless you can remove it from the usual library links page.
Turning on/off Library Search in Google Scholar
4. The print/or non-electronic holdings option...
Arguably the library links programme implementation is superior to the normal link resolver implementations because users immediately can tell at a glance what they have access to and what they don't. While for the typical setup, the users won't know what is available until they click and see the result.
However there's a drawback with only showing a link if full text is available.
What can users do if Google Scholar tells them they don't have access?
Under the typical implementation, every entry in the search page will have a button and if they click on the button, they get redirected to the link resolver which might offer options like Interlibrary loan or Document delivery even though there is no full text.
Clicking through a link which the link resolver determines there is no access to full text, ILL option offered
But if you are on Google Scholar, you are stuck since no link appears if it determines you have no access to full text. Or are you?
In fact, Google Scholar library links programme does offer an alternative option for such scenarios, though it seems not many/all academic libraries take this option.
It seems for each institution that setups library links with Google Scholar, you can set an option to still provide a OpenURL link with a separate label when it detects you do not have full text. For us, the option is available in Alma Publishing Profile, and the same is likely to be similar for other link resolvers/Knowledge base.
Adding the label "Try SMU Libraries" where there is no electronic holdings available in Google Scholar
In my institution, we have chosen "Try SMU Libraries" as a label to differentiate it from the usual "Findit@SMU Libraries" for library links.
"Try SMU Libraries" appears in Google Scholar when no Library Links (electronic copy) is available via institution
As you can see above this option appears in a different spot, right below each entry. It is often next to the "Library Search" option if you have turned it on.
This option is mutually exclusive with the normal Library Links (Find it @ SMU libraries), so either one or the other will appear. Sometimes, both won't appear, I suspect this occurs in cases where there is no electronic full text AND the metadata in Google Scholar isn't sufficient to generate a OpenURL link.
Another nice effect of this is that you can use this for checking availability of books (both physical and ebooks) and for most setups including mine it will send you to the book/ebook (record if physical) or offer the usual ILL/DDS services if no full text or physical item is available.
As such you may not even need the OCLC Library search setting at least for institutions like mine where we don't allow users to use ILL directly via OCLC, so there is no advantage sending them to OCLC.
As a sidenote, I was slightly disappointed to note that Google Scholar button does not support this feature.
It can show you library links when your institution has access (method #2) as well as free full text (method #1) but it does not appear to show method #4....
In the example below, I am on a article on IEEE platform. My institution does not have access to it. If I did the search on Google Scholar, I would be able to use the "Try SMU Libraries" option to look for other options like ILL
This does not work when I use Google Scholar button, where I am stuck with no options.
5. Subscriber links
This is I think the newest feature Google Scholar has done and I suspect it is the feature least likely to be known to both researchers and even librarians. I myself didn't fully understand how this works and only got an inkling after reading about Google's Campus Activated Subscriber Access (CASA).
This feature is called subscriber links.
Part of the reason why it has so flown under the radar of many including librarians like me is that it's use is pretty much transparent.
Sciencedirect subscriber links in Google Scholar
If you look at the links at the side, you might think those links to Sciencedirect are there because those are Open Access (Gold, Hybrid) or just free to read (Bronze OA) and indeed they often are.
But some of them are actually articles behind a paywall! But why are they showing?
Unfortunately explanations of Subscriber links (even Google's own) are often hard to understand. The best I found was in the form of an FAQ here.
The way I understand it is as follows, subscriber links are a mirror to library links (Method #2 above), except the holdings.entitlements are from publishers directly not libraries. (They work independently though, so it is possible for one GS entry to have both Library Links and Subscriber links showing).
To recap for the library links programme, we the library sends Google Scholar our holdings/entitlements (at the journal vol/issue level). For the subscriber links, the same thing happens but the Publishers/Content owners send the holdings!
The clever thing about this is that Google Scholar normally only shows those subscriber links when you are on campus (but see later about CASA).
I'm not sure if the IP range is given by the publisher or they using the one given by libraries in the library link programme,
That's why for most users even librarians they did not notice this feature, because as long as they are on campus, these links to Sciencedirect will work seamless even past paywalls (because they are on campus).
I personally noticed these links cos of a quirk in the way we do IP authentication, aka we require proxies even "in-campus". I initially thought these links [HTML] publisherdomain.com that brought me to paywall articles were errors by Google Scholar in Open access detection until I realized that was not the main reason.
But how does this help when you are off-campus? This is where Campus Activated Subscriber Access (CASA) which is described as "the remote access extension of Google Scholar’s Subscriber Links program." comes in.
Elsever describes it as such
When a user visits Google Scholar from an on-campus IP, Scholar looks up the list of subscriptions accessible to them from the Subscriber Links data it received from publisher sites. This lookup enables Subscriber Links (access links to the subscribed articles) in Scholar interface for on-campus users. CASA extends this by recording the corresponding subscription IDs for the user on Scholar servers. Each subscription ID is an opaque identifier chosen by the publisher and included in the Subscriber Links file to identify a subscription (e.g. “323278”). The recorded subscription IDs are maintained for 120 days and automatically expire after this period is over.
When the user comes back to Google Scholar but is off-campus
Scholar looks up the recorded subscription IDs and appends an encrypted token (CASA token) to URLs from participating publishers. The CASA token includes the subscription ID for the institutional subscription that the article is a part of, the timestamp when the token was generated (to limit the period of validity), and an IP range that the token is valid for.
After 120 days these tokens will expire unless you go back on-campus, but until then this will allow you access past paywalls via your institution.
If this confuses you, the basic idea is when you are on campus IP accessing Google Scholar, it records information about your affiliation and its subscriptions. However the next time you go to Google Scholar, even off campus, the token created will give you access to full-text of papers even behind paywalls.
Privacy wise, while you don't need a Google account to use this, if you are already signed in to your Google account, the CASA token will be tied to your Google account. The advantage is this means all your devices (including mobile) will have the same benefits! The disadvantage is obvious if you worry about being tracked.
Finally to complete the story, some publishers including Elsever support "Universal CASA". If you have a CASA token in your browser and you land on a Sciencedirect page or any other publisher that supports Universal CASA (even if not via Google Scholar), you will see a small grey box on the right side of the page that says "PDF" or "HTML" that will give you access when you click on it.
It has been noted that use of "Universal CASA" will lead to information exchange with Google, you can turn this off in Google Scholar settings , under Account and uncheck "Signed-in off-campus access links".
Turning off CASA links
Backed by Google since 2017, an impressive number of content owners including Heinonline, Gale, JSTOR, Ingenta Connect, Highwire hosted journals, Wiley, Project Muse, APA, Ebscohost ,Emerald and more have signed up to support CASA, so you may find your library is already doing this without you noticing.
Part of it is some content providers do this on a opt-out basis, so you may not have noticed, while others like Elsevier require you explictly opt-in.
Conclusion
There you have it! All 5 different ways Google Scholar help you the reader gain access to full-text or at least provide additional options.
As a bonus below is a table of how some major linking to full text differs. This includes traditional link resolvers vs Library Links programme, Libkey Link , Subscriber links and GetFTR
