Magazines and books on a bookshelf. Image credit: chuttersnap (CC0)
Peer review is an integral part of scientific publishing in which active researchers – who may also serve as editors at scientific journals – review and assess the standard of submitted research. Peer review is therefore central to science, as it determines which articles are published in high-profile journals, and in turn, influences the careers of scientists.
Reviewers are expected to be relevant experts in the field of research they are reviewing. But this can sometimes create a conflict of interest as the reviewers’ own articles may be cited in the manuscript under review, which could influence their peer review. Conversely, reviewers whose work is not cited may feel that relevant work has been overlooked and may suggest their own papers be added – a practice that has been exploited in the past.
To find out whether citations influence peer review recommendations, Barnett analysed more than 37,000 peer reviews from four journals that operate fully open peer review, making all versions of submitted manuscripts and reviews publicly available. Using a matched design, he compared two or more reviewers who evaluated the same manuscript.
The analysis showed that being cited in the first round of review did not increase the likelihood of a favourable recommendation. However, in the second round of review, reviewers who were cited were more likely to approve the article. Contrary, reviewers who requested a citation to their own work were much less likely to approve the article.
However, a similar pattern was observed for reviewers who suggested citing work other than their own. This indicates that some citation requests may reflect legitimate concerns about missing context rather than purely self-serving behaviour.
These findings provide valuable insights into how the peer review process can be improved. Some journals – including those analysed by Barnett have already taken steps to reduce inappropriate requests for citations from reviewers. Other journals could consider implementing automated systems to flag self-citation requests, which could help improve the fairness and integrity of peer review systems, ultimately benefiting both scientists and science.
