An open-access journal will be removed from a leading bibliographic data provider’s list of trusted outlets after it was alleged to have inflated the citation counts of its papers by sneaking “irrelevant and undue” references into metadata.
The action by Crossref, which supplies reference data used by research databases such as SpringerLink, OpenAlex and Dimensions, comes after research fraud sleuths claimed that the Jaipur-based International Journal of Innovative Science and Technology inserted more than 80,000 citations into its metadata references lists at the time of registration in scientific repositories.
According to a preprint published on arXiv this month, 2,782 papers had their citation counts falsely inflated by “sneaked references” that did not appear on the paper’s bibliography but did feature in electronically submitted metadata.
In one case, a paper received 6,059 undue citation counts, “some of which did make their way into various scientometrics services”, explains the paper.
At its peak each paper was allegedly carrying an average of 73 “sneaked references”, with some referencing papers published only six days earlier. In one case, a paper was reportedly referenced by an article published on the same day.
According to the preprint, authors were unlikely to be responsible for the manipulation, which appeared to benefit solely papers published by the Indian open-access journal.
As a result, Crossref, which collaborated on the preprint, decided there was “an intention to manipulate the citation record” and had “started the process to revoke this organisation’s membership”, the preprint states.
The journal – which claims to be indexed by Google Scholar and Pubmed, among others – blamed a clerical error, claiming it “happened by mistake” and the offending employee’s supervisor had been fired over the incident.
Efforts to explain the erroneous references to Crossref had not been fruitful, it added, claiming the company had “tried all methods to solve [the problem] with Crossref but it does not want to understand”.
“We accept that maybe someone from our team [was] involved in this [but] every company works with humans [so] mistakes are possible. If someone gives an explanation then [you] have to listen and consider that error so that issue can be solved,” it added.
According to research integrity sleuth Guillaume Cabanac, a researcher at the University of Toulouse who co-authored the study, the fraud demonstrates how “fragile bibliometric platforms building on Crossref metadata are” because “Crossref hosts data but does not verify them”.
Unlocking the potential of open access and open research
“All publishers but a fraction, to the best of our knowledge, push reliable data but, when they don’t, the manipulation spreads to multiple places that get ‘contaminated’,” he said, noting how this could spread to bibliometric indicators used for performance assessment and institutional rankings.
The paper also ran an analysis on what it called “duplicated references” – references that were repeated multiple times in a bibliography, contrary to accepted practice in which individual references are usually mentioned in full only once in a single paper.
While these duplicated references may not be intentional manipulations in the same way as sneaked references, certain journals did benefit from an “impressive number of duplications”, explains the paper, which identified 10 journals and 10 individual papers that benefited most.
Overall, more than 1 million papers were identified as having at least five more references in their metadata than their official bibliographies, adds the study on the extent of erroneous referencing.
Another of the study’s co-authors, Lonni Besançon, a researcher at Linköping University, said he wanted to see “sanctions for journals which have engaged in sneaked references”.
“Crossref’s corrective actions are important but not enough,” he said.
Publicly blacklisting journals that engaged in this fraud would, however, be “complicated”, he continued.
“Some of these lists are useful but they are quite difficult and time-consuming to maintain – I would not particularly feel comfortable myself creating a list of blacklisted journals. But the fact that some journals have engaged in any kind of questionable practices should be publicly documented and easily verifiable,” he said.
More broadly, Besançon said he was “in favour of not trusting journals” to provide reliable metadata information. “Publishers have, in general, shown how unreliable they were to maintain a good record of legit academic findings, of efficiently and accurately correcting problematic papers, which may even have ended up costing lives,” he said.
“We should have higher standards for the publishing industry in particular considering the money we are paying them.”
