KEY TAKEAWAYS

• Breaking peer review into stages could decrease the burden on expert reviewers and improve the quality of published research.

Peer review is a key part of scholarly publishing; however, there have been increasing calls to shift away from the traditional peer review model to make the process more efficient and sustainable. In a Nature World View article, Professor Olavo B. Amaral describes an alternative approach to peer review that could improve data quality and transparency, and lessen the burden on peer reviewers.

Conventional peer review relies on expert referees to evaluate an article’s claims and its suitability for publication in the target journal. Due to time constraints, the underlying data are rarely scrutinised, potentially allowing errors and fraudulent results to go undetected.

Prof. Amaral believes that every manuscript should undergo basic checks to ensure that the data are complete and consistent, calculations are correct, and analyses are reproducible, but that only select articles, such as those of special interest, should be sent out for expert review. Such an approach would allow peer reviewers to use their time more effectively, on papers for which the data have been validated.

“Not all research needs to be reviewed by an expert. Much of the low hanging fruit of quality control doesn’t need a specialist — or even a human.”

Although certain aspects of manuscript quality control could be automated, algorithms work best on structured text, and most scientific fields do not have standardised formats for presenting results. A more fundamental problem is that data checks cannot verify that the data were collected as reported and have not been ‘cherry-picked’. To address this issue systematically, Prof. Amaral suggests that the focus should switch from scrutinising manuscripts to quality control of research practices, as proposed by frameworks such as Enhancing Quality in Preclinical Data (EQIPD). Implementing this change could not only make peer review more viable but could also improve data reproducibility and increase trust in published research.

Prof. Amaral calls on field experts to develop guidelines for data standardisation and urges funding agencies to facilitate the efforts to improve data collection and reporting by, for example, rewarding researchers for having specific aspects of their results certified.

—————————————————–