Keeping track of metric manipulators on ResearchGate will be difficult without better moderation, says librarian Savina Kirilova.Credit: Henrik Sorensen/Getty
Hosting an estimated 25 million users, ResearchGate is one of the largest academic networking platforms in the world. But as the site approaches its twentieth anniversary, it is facing growing criticism from some scientists who worry that some users are manipulating its metrics to give their papers more visibility and clout.
Savina Kirilova, a research librarian at Sofia University in Bulgaria, co-authored a paper in February1 claiming that some users are allegedly working together to inflate metrics tracked by ResearchGate, such as the number of reads and recommendations that a paper gets.
Kirilova spoke to Nature Index about what the site can do to combat misuse of its platform and improve its reputation among the research community. Nature Index approached ResearchGate for comment about her points, but did not receive a response.
What concerns you about ResearchGate?
Academic social-networking sites, such as ResearchGate, Academia.edu, Mendeley and Google Scholar, play a key part in the way academics share, collaborate and communicate their research. ResearchGate sets itself apart by offering features such as user recommendations, which can be a valuable tool for driving discussions about papers in the scientific community.
But there are many issues with ResearchGate that can undermine research integrity. For example, the platform — which hosts both peer-reviewed and preprint articles — does not have a formal vetting process for most of the content it hosts. Its open upload policy — in which users are free to upload their own research-related work, as long as it does not break any copyright stipulations — increases the risk of users submitting plagiarized or low-quality research.
ResearchGate uses metrics such as the number of reads, recommendations and citations that a paper gets, which should reflect its quality. But these metrics can be easily manipulated, and in some cases inflated to the point of absurdity.
For example, in our paper, we cite the case of Larry, a cat that belonged to the grandmother of Reese Richardson, a graduate student in metascience and computational biology at Northwestern University, in Evanston, Illinois. Last year, Larry the Cat got many citations from Richardson posting a fake profile and some fake article titles on ResearchGate, with Larry as the author. Larry went on to garner more than 100 fake citations on Google Scholar. (ResearchGate has since deleted the profile and Google Scholar has deleted the citations.)
These metrics feed into other bibliometric indicators, such as ResearchGate’s own Research Interest score (RI score), which is intended to measure the impact of a person’s research.
What are the risks in not regulating uploads and user activity?
The concern is that researchers might artificially inflate their indicators or share their publications in ways that boost their visibility unfairly. Unethical researchers could use generative artificial intelligence (AI) programs to write hundreds of articles, post them on ResearchGate, and create fake metrics for them, for instance.
In our paper, we describe a case of small groups of researchers that are deliberately inflating each other’s metrics to enhance their reputation. Allowing these systems to be abused risks putting reputable researchers and fraudsters at the same level on the platform, because a peer-reviewed paper and an obscure pseudoscientific rant could be presented in the same manner, with equivalent credibility.
