The Dark Side of Scientific Publishing?

Nature News recently highlighted the fact that scientific publishing is getting slower – even though in the digital age everything about publishing and communication should be faster (1). A closer look, however, shows that the process of individual review is fast enough. Email communication, online manuscript management tools, outsourced and digital design, plus the now common procedure of making PDFs of accepted articles available way ahead of print are all modern advances of the Internet age. The problem lies in the trend of “journal shopping”. 

Hiring committees and grant agencies have placed so much weight on the journal impact factor (IF) that CVs are screened for publications in those journals with the highest standing (with reference to this indicator). Much has been written about the fallacies associated with the IF; however, the pressure is still on the individual scientist. Publishing in shiny, high-impact journals is a mighty weapon in the hands of a scholar who tries to snatch that grant or get that all-important permanent contract at a university of their choice. Obviously, high-impact journals receive an overload of submissions by eager scientists and can only publish a fraction. Scientists take their chances and submit anyway: one journal at a time, down the IF ladder from high to low. Many will be rejected, and fairly often this only happens after external review – something that consumes the time of the reviewers as well as the authors. Chances are, a set of new reviewers will hand the submission to the next journal on the authors’ wish list, while time is lost in a serial fashion.

One solution to delays caused by serial events is parallelization. In computing, parallel processors are used to permit faster calculation (or more calculations at one time); parallel DNA sequencing has dramatically increased the supply of genetic data. While the Ingelfinger rule prevents simultaneous submission to more than one journal, third-party review organizations (such as Axios Review, Peerage of Science, and Rubriq) can assess the ‘fit’ of a paper to multiple journals simultaneously and pass the paper to the outlet that is thought to be most suitable. Early evidence suggests that the parallelization approach significantly shortens the review process.

I am an academic editor at Axios Review (https://axiosreview.org) and so know firsthand that this trial project has run quite successfully since its official launch only two years ago. Our open letter in a follow-up issue of Nature covers the essence of this idea (2). For instance, 85 percent of papers reviewed by Axios Review get accepted at the first journal to which they are sent, and more than half of the accepted papers are not peer reviewed again by the journal. These are impressive numbers and support the general idea that independent, external review can be de-coupled from processes at the individual journals.

We scientists know that other solutions to the problems associated with journals and impact factors are currently tested and discussed, too. These include ideas to either get rid of the impact factor altogether, or to get rid of pre-publication peer review in favor of a system where manuscripts may be openly discussed by the community and improved in an iterative forum fashion on preprint servers. These measures appear equally suited to doing away with journal shopping. However, there are many reasons to keep a system in which journals rely on invited expert opinion and the associated review process. Peer review is a tool that acts like a filter for high quality presentation of experiments and analyses, as well for balanced and neutral conclusions. Within this setting I urge colleagues to try out the new system of third-party review. As a reviewer you will be grateful for potentially receiving fewer submissions, and as an author you will appreciate taking much fewer steps when writing and revising your manuscript.