Not so long ago, most papers were published in scientific society owned journals, and only for their most important discoveries did scientists send a short version to one of the weekly magazines. Even then, the full paper with all the data and details was generally published soon afterwards in a traditional journal. The entire enterprise was financed in large part by library subscriptions, with only minimal page charges for authors. Access to papers meant heading to a library to find journal issues on the shelves. This was not ideal, but academic libraries allowed public access and provided copies through interlibrary requests from academics and the general public as well. This system served the needs of the scientific community for many years, but the advent of the Internet offered a brave new world. With online access and the digitization of journal archives, we no longer had to make trips to the library to read articles or make copies, and we could get papers of interest anytime and anywhere we wanted. Libraries wondered why they were paying for subscriptions if no one was reading the print journals, and many institutions found a way to save money by cutting print journals, sometimes not replacing print with institutional online access.
In the many discussions I have had with colleagues about scientific publishing, there seems to be an underlying assumption that the costs of running an online journal are minimal. It is true that authors are not compensated for their manuscripts, and editorial boards and peer reviewers of scientific journals are largely volunteers as well (at Biophysical Journal, there are modest stipends for the Editor-in-Chief and Associate Editors). However, there are still significant costs required in scientific publishing, from managing a secure submission and review site to copy editing and typesetting the final manuscript to providing archived access to papers and supplementary material. The actual cost varies with investment in editorial and presentation efforts, but appears to be at least $2,500 per paper, and likely much more for papers with dynamic content such as videos and animated figures. Compare this to the average author page charges at Biophysical Journal, which are less than $1,000, a lower direct cost to authors that is supported by a large base of library subscriptions.
As the number of library subscribers decreases, journals can balance their budgets by raising fees to authors, typically in the form of an article processing charges (APCs). At the same time, many journals offered open access, either included in the APC or for an additional fee, as a carrot to attract papers. Now, some open access journals charge APCs of $5,000 or more, and with these payments, the journals no longer need to sell library subscriptions. This has contributed to growing stables of cascading journals (such as the Nature subtitles) that lead to expensive open access journals at the end of the line. In the past, outstanding papers that were not accepted by one of the top high-impact journals would usually be submitted to the appropriate society journal. Today, these papers often move down the chain of subsidiary journals with little effort required by the authors, and this contributes to decreased submissions to other journals.
For many of us, achieving 100 percent open access is an ideal goal for scientific publishing, and the arguments in support of an open access model are many. Some scientists have been unapologetically pushing the transition to open access, but the rest of us have largely been complicit by chasing after the status of “high impact” journal articles. Regardless of how we feel about open access, however, the path towards 100 percent open access publishing has, directly or indirectly, at least three harmful side-effects for science and the scientific community.
First, it has moved significant financial support for scientific publishing from an indirect cost (library subscriptions) to a direct cost (author-paid APCs). For those of us who are funded by government agencies or disease-based philanthropies, this means we must spend more of our research grant dollars for publications. Some deep-pocketed private foundations, such as the Wellcome Trust and Howard Hughes Medical Institute might provide additional funds for publication, but in general, direct costs are a finite pool. The cost differences between the average open access journal and the traditional model means that for about every 100 open access papers, the additional publishing costs would fully fund a National Institutes of Health R01 grant. That might not sound like much, but in 2017, there were over 50,000 papers published in PLoS One, Scientific Reports, Nature Communications, and Cell Reports altogether, which is the equivalent cost of approximately 500 research grants.
Second, as detailed recently by James Zimring in Scientific American (https://blogs.scientificamerican.com/observations/were-incentivizing-bad-science/), the economics of open access publishing provide a perverse incentive against rigorous manuscripts and peer review. Open access journals make money based on the number of published articles, rather than the number of journal copies sold. In contrast to traditional publishing models where revenue depends on subscription sales, open access journals generate more revenue by publishing more articles. Library committees evaluating new and renewing subscriptions placed a high value on the long-term reputation of a journal that depended on publishing only the most rigorous science. In my experience, academic and professional scientific editors do their best to ensure rigorous peer review, but it is difficult to push continually against business incentives. Failure to maintain the highest standards opens a pathway down a slippery slope that would be difficult to undo.
Third, an indirect effect of these changes has been to move much of the arbitration of publishable research from professional scientists (editors of society based journals) to professional editors (journal house staff). After completing their scientific training, professional editorial staff members have focused on publishing rather than discovery, and thus they do not have the depth of experience found with academic editors. Of course, professional editors become broadly knowledgeable and can bring advantages, especially in terms of bridging between authors and reviewers. However, these editors are under pressure to choose papers that will be highly cited in the nearterm (and improve their journal’s Impact Factor that depends only on citations during the first two years following the year of publication), rather than the most impactful work for the long-term growth and health of the field.
We are well down the road towards maximizing open access science. As we continue, though, we need to keep in mind and advocate for the traditional strengths of our peer reviewed publication system: institutional support for publishing, editorial decisions made by practicing scientists, and placing scientific rigor over financial exigencies. Failure to do so might lead to short-term flash at the expense of long-term damage to scientific research.
Together, our Society of diverse and creative thinkers will always achieve a better solution than any single person. The issues facing us that I have written about this year can only be addressed effectively by all of us thinking and working together. Personally and on behalf of the Society, I welcome your thoughts, criticisms, and questions.
—David W. Piston