Suspicious citation numbers are a red flag for research paper mill slop

• Analysis of 33,159 cancer journal articles discerned tell-tale patterns of fake citation numbers in 4000 paper mill articles. 

• Artificially inflated citation counts distort the evidence base to misinform future research and clinical translation. 

Their findings point to an organised campaign to inflate citations on paper mill-produced journal articles published in established molecular oncology journals. 

The study, Suspected distortion of citations in high-impact cancer journals is published in PNAS (Proceedings of the National Academy of Science). 

Corresponding author and statistician Professor Adrian Barnett (pictured above), from QUT's  Australian Centre for Health Services Innovation and School of Public Health and Social Work, said the researchers had documented a pattern that suggested widespread, coordinated manipulation of citation counts by paper mills. 

"Artificially high citation rates can inflate journal impact factors, amplify unreliable findings, and slow scientific progress,'' Professor Barnett said. 

"Research paper mills are fraudulent organisations that produce and sell manuscripts and publishing services of faked research.  

''Just as likes and shares indicate social media influencers' audience engagement, citations from other researchers are an important academic measure for evaluating scientists' research impact. 

"Because citation counts are now a key measure of academic performance, they can incentivise journals and researchers to maximise citations counts." 

Professor Barnett said paper mill papers were not genuine scientific contributions, and so highly manipulated citation counts reflected artificial amplification rather than genuine scientific engagement. 

"A danger is that artificially inflated citation counts may distort the evidence base by misinforming future research and clinical translation,” he said. 

Co researcher Baptiste Scancar said the team evaluated the impact of paper mills on citation patterns in 33,159 articles published in 20 journals between 2012 and 2023. 

''Applying our previously developed detection model, we flagged a total of 4,085 articles as potentially associated with paper mills, based on their textual similarity to the titles and abstracts of retracted paper mill articles,'' Mr Scancar said. 

''We show that flagged papers had many more citations than non-flagged papers but had fewer readers. 

“Having more citations but fewer readers makes no sense, as citations come from readers.