You can better control whether same or similar records being imported from different sources overright or duplicate
While the problem with duplicates seems to have been well covered in other peoples ideas, and we also want better systems in place to manage these, we also have evidence that some systems (ResearcherID in this case) will overwrite works that already appear in an ORCID record and that these works are actually NOT the same. At the least it would be good to be able to control if the import overwrites or duplicates in these instances.
Posted on behalf of the University of Otago Subject Librarian team.
ORCID have a process built into the Registry that groups works, funding items, and peer review contributions. However this may be different from how other systems may perform them, as no items are deleted.
As you may imagine, much as we do not disambiguate researchers based on the name alone, we do not disambiguate works (or funding or peer review items) based on the name alone. Rather, grouping on the ORCID records functions is by grouping based on the persistent identifier (PID). I’ll use works as an explanation. When multiple works are added from different sources with the same identifier, they are grouped together on the ORCID record.
We see this as a good thing: Multiple versions mean multiple assertions of the same data, e.g. a work. This builds trust in the data belonging to that user.
An example can be seen on the record of our Executive Director Laure Haak (https://orcid.org/0000-0001-5109-3700):
The article “ORCID: connecting researchers and scholars with their works” published in Insights: the UKSG journal has been added by two sources: Crossref and Scopus. The DOI identifier is the same, so the two items are grouped together as one.
If a work does not include a identifier — for example, if the researcher has added it manually without an identifier, or imported using BibTeX without an identifier — then the work will not group with others. And if a work has been added by two different systems, for example ResearcherID and Scopus, with two different identifiers, then the researcher can group them together by making a copy of one and adding the missing identifier. See more about that here: https://support.orcid.org/knowledgebase/articles/1807567
Of course, for the researcher, the user experience of having to add an identifier themselves, or making a copy and adding an identifier, can be annoying. This is why we encourage researchers to add works and funding using the Search & Link wizards or links with their institutions — when organisations update the records via the API (as Scopus, etc. do), at least one identifier is required.
This is also why it is extremely important that organisations which update researchers’ ORCID records add persistent identifiers (PIDs) such as DOIs to the works when transferring the data to ORCID. We encourage members when building these processes to include as many PIDs as they know. This will provide a great service for researchers, helping the works group easily on their record, so they do not have to do it themselves.
ORCID Community Team
Ha ha spelt Overwrite wrong in the title lol