Dutch universities are discussing the pros and cons of the various author identifiers schemes. All these universities have been collaborating with the National Library on the Dutch Author Thesaurus (NTA). The unique identifiers of the Dutch Author Thesaurus, the so called Digital Author Identifiers (DAI), are being used in most of the university libraries’ catalogues as well as the (Open Access) repositories of those universities and Narcis, the overarching open access repository of The Netherlands. But the DAI’s will soon become obsolete.
Most university libraries are going to stop cataloguing in de Dutch General Catalogue (GGC) and will start cataloguing with the OCLC Worldshare platform. One of the problems this creates is that the Worldshare platform doesn’t support the Dutch Author Thesaurus, i.e. we lose the benefit from having the Dutch Author Thesaurus (NTA). When the support of the Dutch Author Thesaurus stops, we also create a problem for the repositories who are making use of the NTA to identify their authors as well.
One of the unknown advantages of the Dutch Author Thesaurus was that it was run on software developed and maintained by OCLC Leiden (NL) office. OCLC has created some time ago already a completely unknown route of unique identifiers from the Dutch Author Thesaurus to the VIAF and ISNI. So most scientific authors, researchers, from the Netherlands have on top of their DAI a VIAF and ISNI as well. Look for instance at Professor Martin van Ittersum, who has a DAI, a VIAF and subsequently and ISNI. We can praise ourselves lucky with the foresight and initiatives by OCLC Leiden. The DAI was ahead of the time, but too local to be practical in a continuously more international research landscape.
In the meantime the National Library of the Netherlands is taking up its role as manager of the ISNI for Dutch authors in collaboration with the university libraries. The ISNI will get support from OCLC Leiden to be integrated as an authority file with the Worldshare platform. So projects are underway to phase out the DAI and replace that with support for ISNI directly. So all problems solved. You would think.
What about the repositories?
Repositories also make use of the DAI to identify authors. Interestingly, most often this was done through the Current Research Information System (CRIS) called Metis. Metis was a Dutch CRIS systems, currently being phased out at the majority of universities and replaced by more modern systems such as Pure and Converis. None of the Dutch universities have selected Symplectic as a solution for their CRIS.
The first installations of CRIS’ses probably still integrate the DAI, but these need to be replaced. OCLC solution is to replace DAIs’ by ISNI’s. That works fine in library catalogues, but is less functional in CRIS/repository systems and secondly authors and researchers are not aware of the fact that they have a range of unique identifiers already.
Interestingly, researchers are applying for privately managed identifiers such as ResearcherID (from Thomson Reuters) Scopus author ID (managed by Elsevier) and Google Scholar Citations (GSC) identifiers. The reason why they do this, is to disambiguate their name from fellow researchers with a similar surname and initials or to repair their sloppy track of abusing their own name and initials over the course of their research career. But probably the most important reason is that they want to have a comprehensive publication list available on the web to which they can refer with a single, persistent URL. For this last service DAI, VIAF and ISNI have failed so far. To be honest they have been created for a different purpose, only to disambiguate author names. Not to collect a comprehensive publication list for researchers. But explain that to those researchers. That is a long story when you have to explain the finesses of identifiers that are not to collect publication lists.
The advantage of ORCiD is that it aims to be both. A unique identifier to disambiguate names, as well as a tool to publish and maintain your comprehensive publication list. And more recently your education, funds and projects as well. And probably most important: the researchers is in control what to add or leave out from their publication track, funding, education etc. ORCiD is actually already an identifier switchboard as you can see for Martin van Ittersum whose ORCiD links out to his Scopus Author ID, ResearcherID and ISNI. However, uptake of ORCiD, although quite spectacular with over 1.5 million registrations, is far from comprehensive for Dutch Researchers. In some countries around us we have seen national initiatives to speed up ORCiD uptake for their researchers, eg. Italy and Denmark. As mentioned at the beginning of this post, we, in the Netherlands, are still debating whether we need to provide support for ORCiD, since we seem to have settled for ISNI (in the cataloguing environment) already.
The advantages of ORCiD above ISNI are many fold in an academic library setting. ORCiD is making solid inroads in the metadata exchange world. Major publishers are supporting ORCiD, and it is not really difficult to predict that in a couple of years ORCiD’s are as well known as DOI’s at the moment. With the ORCiD of your researchers at hand, you can harvest metadata on publications and funding from publishers directly or more conveniently from services as CrossRef and FundRef. Also commercial bibliographic services like Web of Science and Scopus are supporting ORCiD. Metadata from all these kind of services are exchanged with the help of API’s. The development and exchange are only just starting. But typically they are centred around ORCiD’s rather than the related ISNI’s. It only seems likely that exchange of information with WorldCat will be based on ISNI’s rather than ORCiD. So for the humanities, where books are the mainstay of publications, ISNI’s are required as well. I foresee a more important role for ORCID’s rather than ISNI’s in an academic library world. If it was only that articles, and exchange for metadata for these more refined granular items of work are more intensive than the number of scholarly monographs. Actually up until recently most libraries were not involved in metadata creation and collection of items at this level of granularity, and this should change in my opinion.
In the CRIS we are able, and should take our responsibility as librarians, to make it a switchboard for metadata exchange on behalf of our researchers. So we should implement support for ORCiD as well as ISNI and ideally some other researcher identifiers as well, e.g. those of Scopus, Web of Science and Google Scholar. And we shouldn’t underestimate the growth of Researchgate or Academia.edu. The question is though, do we sit and wait till our researchers register themselves and register these registrations in our systems, or are we acting proactive, and do it as much as possible for them? I am in favour of a proactive approach and well for the following reason
Google Scholar Citations uptake at Dutch universities
Google Scholar launched the possibility to make your own profile page some 4 years ago in November 2011. You would think that all researchers would have a public GSC profile by now. At the beginning of July 2015 I have checked the number of researchers for each university based on their e-mail affiliation in GSC, by manually clicking through the list of results (only 10 results per page ). In the case of university hospitals I have included these in the e-mail affiliations as well. To judge the penetration of GSC I downloaded the latest estimates for researchers at Dutch universities from the website of the Association of Dutch Universities (VSNU) In the following table the summary of results is presented:
|University||search string||GS profiles||Faculty (FTE)||Percentage GS profiles|
|EUR||eur.nl OR rsm.nl OR iss.nl OR erasmusmc.nl||670||1166||57%|
|LEI||leidenuniv.nl OR lumc.nl||645||2077||31%|
|RU||ru.nl OR radboudumc.nl OR umcn.nl||754||1878||40%|
|RUG||rug.nl OR umcg.nl||965||2230||43%|
|TiU||uvt.nl OR tilburguniversity.edu||249||849||29%|
|UM||maastrichtuniversity.nl OR mumc.nl OR unimaas.nl||415||1804||23%|
|UT||utwente.nl OR itc.nl||833||1526||55%|
|UU||uu.nl OR umcutrecht.nl||1149||2733||42%|
|UvA||uva.nl OR amc.nl OR acta.nl||1025||2619||39%|
|VU||vu.nl OR vumc.nl OR acta.nl||933||2187||43%|
Technical University Delft (TUD) has by far the most researchers with 1366 public GSC profiles, followed by Utrecht University (UU) and Wageningen UR (WUR). The Open University (OU), Tilburg University (TiU) and University of Maastricht (UM) are the smallest in terms of public GSC profiles. The number of public GSC profiles are subsequently compared to the Faculty per university as reported by the VSNU for the end of 2014 (table 8). The number used by VSNU are FTE, rather than actual persons. On top of that, the academic hospitals are not included in the VSNU count, whereas whey were included in the GSC counts. This should have a positive influence on the share of GSC profiles for universities with academic hospitals. Nevertheless Wageningen (WUR) has the highest share of GSC profiles (69%), followed by Erasmus University Rotterdam (EUR) and University of Twente (UT), at the bottom end we find University of Maastricht (UM) with only 23%, Tilburg University (TiU) and Leiden University (LEI). Overall only 43% of Dutch researchers have claimed a public GSC profile after nearly 4 years after the introduction.
I know for sure that Wageningen University Library has taken an active role in the promotion of GSC profiles, but still only 69% of the researchers have actually claimed their profile after nearly 4 years and that is a gross overestimation since the VSNU figures are reported in FTE, and GSC profiles are whole persons. For Wageningen, researchers at the institutes are not counted in the VSNU figures. At Erasmus University the library also had pages on claiming your GSC profile, but these are currently not available. I can’t really vouch for the activity in this area at the other university libraries. At Erasmus with 57% of the researchers have claimed their GSC profile, and these include the profiles from medical staff which are not counted in the VSNU figures.
If we want to adopt ORCiD in the CRIS systems of Dutch Universities but leave it to a voluntary approach, as chosen in Australia, we can learn from Google Scholar Citation profiles, that after 4 years you have success rates around 60 to 70% at the max. But if we select the ISNI as the only persistent national author identifier system, we start to lose out on the benefits of the many services that are being developed on the ORCiD metadata infrastructure rather than the ISNI infrastructure.
So my conclusion is that we would benefit most from a national implementation of ORCiD in all Dutch CRISses and institutional repositories, next to the ISNI, and use those identifiers in the exchange of metadata with systems such as Narcis, CrossRef, Scopus, Web of Science, WorldCat etc. Research libraries should take the lead in such an exercise.
What I predict though is that a few universities will start to implement ORCiD institution wide, and the rest will follow within a few years. Losing the lead we had when we adopted the DAI as single author identifier.
2015/08/07 Update to include unimaas.nl also as identifier for Maastricht University, this yielded two more profiles. Thanks to Egon Willighagen
Full data set is available at: http://dx.doi.org/10.6084/m9.figshare.1528155