Just a few years in the past, researchers would discover appreciable hurdles when making an attempt to review quotation patterns to light up traits in a discipline, determine new areas of analysis curiosity or pinpoint questionable practices reminiscent of extreme self-citation.
First, they’d have to request entry to one of many massive scholarly databases containing quotation knowledge, reminiscent of Internet of Science or Scopus. Even when entry was granted, they wouldn’t be capable to make public the proprietary knowledge on which their findings have been based mostly.
That’s now altering. Most on-line papers are recognized with a singular set of characters known as digital object identifiers (DOIs). This method is run by Crossref, a non-profit affiliation based mostly in Lynnfield, Massachusetts, that has round 15,000 publishers, funding businesses and different establishments as members. Final month, Crossref introduced that the quotation knowledge related to the greater than 60 million journal articles in its database have been now brazenly accessible for downloading and use.
That’s largely due to the Initiative for Open Citations (I4OC), a collaboration between educational publishers, researchers and different stakeholders, which since its launch in 2017 has been encouraging publishers to make quotation knowledge open. Uptake in some quarters, together with amongst some massive publishers, was initially gradual. A Nature editorial in 2019 known as for these publishers nonetheless dragging their toes to leap on board (see Nature 573, 163–164; 2019). (Springer Nature, publishers of Nature, joined the initiative in 2018. Nature’s information group is unbiased of its writer.)
The opening up of quotation knowledge is welcome. It means higher transparency and accountability for analysis research designed to tell teachers, funders and governments of their choices about areas of analysis they need to focus power and cash on.
However extra is required. Not all publishers index papers on Crossref, and never all listed papers have quotation knowledge related to them. One research revealed in July discovered that about one-third of papers listed in 2021 are missing such knowledge (N. J. van Eck and L. Waltman. Preprint at https://doi.org/10.31222/osf.io/smxe5; 2022). A few of these articles — notably editorials, letters, corrections and e-book opinions — may not have any references, however this in no way applies to all of them. Importing quotation knowledge shouldn’t be seen as elective.
There is a vital caveat to any quest for bibliometric openness. The 2012 San Francisco Declaration on Analysis Evaluation (DORA) states that metrics ought to by no means be used out of context or in isolation to evaluate researchers and their work. We must be cautious to not place an excessive amount of reliance on quotation knowledge, particularly when evaluating scientists for promotions and job functions. But when used properly, it might probably solely be higher to have such knowledge open to all.
And openness mustn’t finish with quotation knowledge. Crossref additionally permits publishers to submit different sorts of metadata, reminiscent of creator affiliations, funding data, data- and code-availability statements, and ORCID IDs, that are used to determine particular person researchers. Nevertheless, not all publishers do that. In an open letter in June, the Open Analysis Funders Group, a partnership of philanthropic organizations — together with the Invoice & Melinda Gates Basis and the Chan Zuckerberg Initiative — that advocates the open sharing of science rightly argued that such metadata must be made accessible (see go.nature.com/3qvfp3u).
Moreover, the Initiative for Open Abstracts (I4OA), launched in 2020, has been pushing for abstracts of research to be brazenly accessible. This could make it simpler for researchers to find, learn and cite research, and would open up extra potentialities for analyses utilizing machine-learning methods, as an example to determine traits in the usage of phrases. (Such strategies are already offering insights in different areas of science when knowledge are opened up, for instance in assessing the standard of peer-review stories.) In line with the July research, solely 39% of the articles with a Crossref DOI listed in 2021 have open abstracts — though that proportion has nearly doubled since 2018.
Depositing all related metadata on Crossref ought to develop into the norm in scholarly publishing, as ought to producing DOIs for each paper. For these publishers that don’t have the time or sources to do that, I4OC, I4OA and others within the open-science neighborhood have declared themselves prepared to supply help.
Finally, all these strikes should be solely steps in direction of the objective of getting all analysis papers brazenly accessible of their entirety. However till we arrive at that time, they’re key to the transparency and reproducibility of analysis. They need to be supported by all.