Scientific journal publishing has undergone significant changes in the last couple of decades with the digital revolution and the rise of open access journals. However, the process of manuscript peer review remains essentially the same as it was in the age of typewriters, even if we now do it by email rather than by post. Typically two or three copies of the manuscript are sent to two or three referees, and the decision to publish is based on their reports. Is this system quaint and outdated in today’s world of instant communication, social media and crowdsourcing? Or has it prevailed because it actually works?
John Bohannon’s sting published in Science, the rise in retractions and disillusionment with the top journals have sparked much discussion on whether the peer review system is actually broken. Only this month, I took part in two separate events where I had the chance to exchange views on these issues. On Wednesday 2nd April I participated in a panel debating ‘Is peer review broken?’, organised by students on the Science Journalism course at City University.
It was interesting to hear that most of the panellists at the City University debate agreed that peer review is ‘chipped’ or ‘stretched’ but not completely broken. These included Tom Reller, Richard van Noorden, Tiago Villanueva, Peter Ayton, and myself. Nikolaus Kriegeskorte provided a thought-provoking view that scientists should abolish journals, simply publish all manuscripts on the Web and let readers evaluate, rate and peer review. However, based on further discussion with the panellists and the audience, the scientific community does not seem to be quite ready for that. See the summary in the blog written by my colleagues Tim Sands and Anna Perman who were in the audience and you can watch the event here.
Similar issues were also discussed during the SfEP event. Anna Sharman showed how the focus is shifting from journals serving as filters of scientific content towards megajournals that don’t filter but publish all articles that are technically sound. Julia Donnelly mentioned how hard it is to publish negative results of a clinical trial in any journal, even though it is widely recognized that publication bias towards positive results is a serious issue. Although BioMed Central offers solutions to that problem by publishing the Journal of Negative Results in BioMedicine and Trials.
Some people express the sentiment that publishing as such is no longer a job or an industry — it’s a button, and commercial science publishing is doomed. Perhaps it is not surprising that some authors would like to move away from journals towards pre-print servers. We are seeing a proliferation of initiatives that provide evaluation of manuscripts decoupled from a given journal, such as F1000 Prime, Peerage of Science, Axios Review or Rubriq.
Only a year ago I participated in a Research Hive Seminar organized by the Library of the University of Sussex where I presented on “Future of Peer Review”. It is amazing that even within this one year there has been further innovation in peer review. We have witnessed the rise of new developments such as Axios Review, who provide referrals for authors to help them identify which journal would want to publish their work. Even different publishers who have so far competed for manuscripts are actually able to work together to enable more efficient peer review by promoting portability.
Journals are introducing various innovations to take pain out of peer review, from re-review opt-out through open peer review to manuscript transfers. Both eLife and Frontiers journals enable referees to discuss the manuscript among themselves, to provide better guidance for the authors on how to revise the manuscript and limit the rounds of revision. F1000 Research go a step further and publish the manuscript first and seek referees afterwards.
Journal publications are still hugely important for scientific careers. From my viewpoint, journal publishing and peer review are definitely here to stay, but editors and publishers need to make sure that authors and readers understand the benefits that it brings. The process must be as pain-free and efficient as possible for authors, reviewers and editors alike.