Journal papers ‘largely indistinguishable’ from preprints

A study that found that most preprint scientific papers are “largely indistinguishable” from the versions that appear in academic journals has triggered fresh debate about the role of commercial publishers in scholarly communication.

Researchers from the University of California, Los Angeles and Los Alamos National Laboratory compared the text of more than 10,000 papers on the ArXiv and BioRxiv servers, which cover fields such as physics, mathematics and biology, with their final published versions.

After using text-comparison algorithms to measure insertions, deletions and substitutions, the researchers concluded that “the contents of the vast majority of final published papers are largely indistinguishable from their preprint versions”.

Papers published on ArXiv reveal “no significant differences”, the study says, while there is “slightly larger divergence” between preprint and final versions of BioRxiv papers.

Most of the final published versions were placed in Elsevier journals (96 per cent), but papers from Plos One and eLife were also measured, among others.

The authors argue that their findings, published in the International Journal on Digital Libraries, “should inform discussions about commercial publishers’ value propositions in scholarly communication”.

US universities spent $1.7 billion (£1.2 billion) on periodical subscriptions annually as long ago as 2008, while the UK figure is estimated at £180 million.

Authors Martin Klein, Peter Broadwell, Sharon Farb and Todd Grappone argue that significant differences between preprints and journal papers should be evident if publishers’ claims to add value by coordinating reviews and enhancing text are to withstand scrutiny.

Mr Grappone, associate university librarian for digital initiatives and information technology at UCLA, said that it was up to researchers to ensure they were getting “the most value for our campuses”. “It’s not in the interest of the publishing industry to change; they are very profitable companies,” he told Times Higher Education.

Commenting on the findings, Caroline Taylor, university librarian at the University of Leicester and chair of the UK’s Society of College, National and University Libraries (Sconul), warned that institutions would find it “increasingly difficult” to pay the “ever-increasing sums” expected for journal access.

Ann Rossiter, Sconul’s executive director, said that publishers “need to be much more realistic about their contribution”.

“While no one denies that they do add some value, this is minor compared to funders and institutions, authors, editorial boards and peer reviewers, all of whom provide their expertise for free, often signing over their intellectual property in the process,” she said.

However, the study’s findings have met with scepticism from some quarters.

Tim Vines, a consultant with Origin Editorial who advises journals on peer review, highlights on the Scholarly Kitchen blog that the study was unable to consider papers that had been uploaded as preprints but rejected by periodicals, a quality control process that was “the other key value proposition of peer review”.

Also, he added, a significant proportion of preprints might have been peer-reviewed before they were uploaded.

rachael.pells@timeshighereducation.com