End of a Bargain

End of a Bargain

1

Springer Nature has sent an email to authors announcing that they will no longer be depositing manuscripts to PubMed Central (PMC) or Europe PMC (EPMC) on behalf of authors who have opted to publish via the subscription route. Authors who pay for OA, either in hybrid or fully OA journals, will continue to receive this service.

Some historical context is in order. Back in 2008, the US National Institutes of Health (NIH) was rolling out its public access policy (which became a blueprint for the OSTP (US Office of Science and Technology Policy)’s Holdren Memo in 2013). The NIH policy required NIH-funded authors to deposit their accepted manuscripts in PMC within 12 months of publication in a journal. Author compliance with this policy was abysmal. Seeing the policy failing, publishers agreed to help by making deposits on behalf of authors. This dramatically increased compliance, practically overnight, and was a key factor (maybe the key factor) in the success of the NIH’s—and later the OSTP’s—public access policies. 

One of the reasons that publishers agreed to make these deposits (essentially supporting a policy that had the potential to undermine their subscription business model) is that NIH (and later OSTP) agreed to keep the post-publication embargo window to no earlier than 12 months. This bargain held for 14 years, until 2022, when the release of the Nelson Memo moved the embargo window from 12 months to zero. 

Given that OSTP (and by extension NIH and other federal agencies) will no longer be honoring their side of the bargain, it is not surprising to see a major publisher cease to support their side. We expect other publishers to follow Springer Nature’s lead. 

In the case of Springer Nature, it is possible that the end of this bargain presents a timely opportunity to further stake out their vision for the future of academic publishing: one that is focused on the author-pays route to Gold OA. Eliminating deposit on behalf of non-paying authors further emphasizes Springer Nature’s distaste for Green OA routes and makes clear they won’t be taking any actions to support them. This may not immediately drive a lot of authors toward Gold OA (nor away from publishing in Springer Nature journals), as most researchers don’t think a lot about funder compliance at the time of article submission, and the need for a PMCID may only become evident when a progress report or grant renewal is due. But this policy shift should serve as a harbinger for researchers that publishers may no longer actively assist them in meeting the policies of their funders (see also Items 2, 3, and 4 below).

NIH Draft Public Access Plan

2

Last week, NIH has just released for public comment their draft plan for meeting the requirements of the Nelson Memo. The plan is largely as expected, essentially mirroring the NIH’s recently implemented data policy (but makes data release more of a requirement than a suggestion) and continuing most aspects of its long-running publications policy (albeit with a zero embargo). For publications, funded authors are required to deposit to PubMed Central  (or have deposited on their behalf) the author’s accepted manuscript (as opposed to the version of record—though that is acceptable) of any peer-reviewed paper that lists NIH funding and that will be included by the author in progress reports or grant renewal applications. As directed by OSTP, the NIH policy requires “public access” not “open access,” meaning that articles must be free to read by not to reuse (without permission of the copyright holder). No licensing terms such as the Creative Commons CC BY are specified and authors remain free to use grant funding to pay publication charges.

On these two points (deposit of the author’s accepted manuscript and no reuse license requirement) NIH’s public access policy is closely aligned with the policy already released by NASA, suggesting that other federal agencies will follow a similar pattern. This leaves publishers with a decision: whether to allow this zero embargo green OA route (and hope the existence of a free manuscript version of the paper does not diminish the value of their subscriptions) or to require federally-funded authors to pay for gold OA publication. This is a decision that closely resembles the decision publishers must make regarding the Rights Retention Strategy (see Item 3).

Rights Retention Strategy

3

Meanwhile, as Europe (and select private funders) continues its more prescribed shift toward  OA and the use of the CC BY license, it seems counterproductive, or at least a little mystifying, to see the continuing emphasis on retention of copyright by authors. The University of Oxford, for example, has started a new Rights Retention Strategy (RRS) pilot, while the UK’s N8 Research Group of institutions recently issued a statement on the importance of researchers “being able to retain their intellectual property rights when their work is published in a journal.”

On the surface, this seems contradictory: authors are required to retain all rights to their written works but also must give up nearly all of those rights through the required implementation of a CC BY license. Ostensibly, the “strategy” part of the RRS dovetails with the Plan S prohibition on the use of coalition member funds to pay article processing charges (APCs) for publication in hybrid journals. The RRS compliance route requires an author to deposit the accepted version of their paper in a repository, making it immediately available under a CC BY license, so that they can then publish the Version of Record (VOR) behind a paywall in a hybrid journal at no cost. To do this, the author must 1) be the copyright holder in order to add the CC BY license, and 2) do so in a manner that overrides, or at least avoids violating the terms of, the license they sign with the publisher. This is a tricky needle to thread,involving legal arguments that have yet to be tested in court. 

Publishers are, understandably, less than enthusiastic about the RRS as it undermines the value of journal subscriptions (by providing OA versions of articles that the publisher has placed behind a paywall), and provides no OA revenues (such as APCs) to cover the costs of the work done. 

Few publishers and journals explicitly support the RRS (aside from some such as AAAS, the publishers of Science, which has a magazine-like section it can continue to sell to subscribers even in a post-OA world). Most publishers require authors making RRS demands to pay an APC and to have their article published on an OA basis in hybrid journals, or they steer the papers to the fully OA titles in their portfolios, making the point moot. Some publishers are more or less ignoring the RRS and allowing authors to post their manuscripts with CC BY licenses but still requiring copyright transfer for the VOR. It is worth noting that the NIH, in their draft policy (see Item 2 above)  suggests the agency is considering plans to develop language for authors “to retain rights to make the peer-reviewed manuscript available post-publication in PMC as soon as processing is complete, without an embargo period,” which may suggest a similar hope to avoid paying author fees. 

To the extent that the RRS succeeds, it will do so by (counterintuitively) accelerating the shift to Gold (not Green) OA. Were that to occur, publishers may welcome the ability to shed the responsibilities of holding copyright on research papers to authors. Under the author-pays APC model for OA, the journal gets paid when the article is published, and, other than the small number of journals that can monetize traffic into advertising revenue, that’s largely the end of any earnings from the article. Secondary rights sales will no longer be viable given the CC BY license, so it makes sense to cut post-publication costs to as close to zero as possible. While a CC BY license allows much, it still places an attribution requirement upon reuse. Someone taking an author’s name off of a paper or claiming the work as their own would violate the terms of the license. Publishers typically employ services to monitor copyright violations and lawyers to defend those copyrights. These costs and the accompanying time burden will now shift to the author. 

While the idea of authors retaining ownership of their intellectual property seems appealing at first blush, CC BY makes that ownership largely symbolic; like most efforts to regulate researcher behavior, it creates unintended consequences and an additional burden on researchers.

Plan S: Ceasing Support?

4

The RRS (see Item 3 above) is meant to assuage researcher concerns about being able to publish in the journal of their choice under hybrid journal prohibitions. These concerns are likely to increase at the end of 2024 when cOAlition S ceases support for Transformative Journals and journals published under Transformative Agreements, which the organization confirmed last month. From 2025 onwards, researchers will not be able to use cOAlition S funds to pay for APCs in hybrid journals published by the likes of the American Chemical Society, Cambridge University Press, Elsevier, IEEE, Oxford University Press and Springer Nature (a full list of publishers can be found here). 

Authors will still be able to publish in hybrid journals under a CC BY license and be compliant with the cOAlition S mandate; however, they will need to find another source of funding to do so. The RRS strategy is therefore, in large part, a work-around for this Plan S prohibition. 

While this prohibition against the use of cOAlition S funds in Transformative Journals may cause some consternation for both publishers and researchers, it is unclear if the same can be said regarding journals with Transformative Agreements. These “read and publish” or “publish and read” agreements, are agreements between institutions (or a consortium of institutions) that allow affiliated researchers both to access the journals of a given publisher and to publish their research in that publisher’s journals. Such agreements are paid for via the institution and not via cOAlition S funds. In 2024, as now, a researcher may publish in a hybrid journal via a Transformative Agreement, and their paper will be published OA with a CC BY license, thereby meeting Plan S funder requirements. Under this scenario, it is unclear how cOAlition S “ceasing support” for Transformative Agreements is meaningful in the same way it is for Transformative Journals.

Gender Diversity on Editorial Boards

5

Journal editors make decisions that significantly affect academics’ careers; being accepted in an impactful journal can lead to grants being approved and tenure being awarded. It is therefore vital that the editorial boards of journals reflect the communities they serve. A recent research paper in Nature Human Behavioursuggests that the vast majority of journals have half as many women serving on editorial boards as are active within their community. Furthermore, this proportion remained constant between 1970 and 2017. In short, little progress has been made to improve gender diversity on editorial boards over nearly five decades.

The researchers parsed more than 173,000 editorial pages from Elsevier to create a database of over 100,000 editors. They then used an algorithm to infer the gender of 81,000 editors and 4,700 editors-in-chief. These editors worked on more than 1,000 journals across 15 disciplines. The authors used the Microsoft Academic Graph (MAG) to identify the publication records of 20,000 editors and 1,600 editors-in-chief who had a unique match in MAG. This allowed them to compare the publication output of editors with the general population of researchers.

Between 1970 and 2017, on average, 26% of scientists were women; during that same time period only 14% of editors and 8% of editors-in-chief were women. Crucially, the gap remained stable throughout this time period. In 1970, 11.7% of researchers and 5.7% of editors were women; in 2017, 36% of researchers and 18% of editors were women.

It is possible that progress has been made since 2017. Organizations are now taking diversity, equity, and inclusion (DEI) issues far more seriously. However, the gap is unlikely to have closed significantly and publishers still have a long way to go. Indeed, according to data from C&E’s 2021 benchmarking study, which contains data from 47 publishers representing 427 journals, only 22% of editors-in-chief are women. Furthermore, only 11% of journals said that they had a formal DEI journal policy, although 56% of journals plan to introduce one.

Editors are often highly productive researchers. According to the Nature Human Behaviour paper, editors tend to have published seven times as many papers as an average scientist of the same academic age working in the same discipline. They also have eight times as many citations, an h-index that’s four times higher than the average, and more than five times as many collaborators. There is good evidence to show that women are more likely than men to leave academia or to take a career break after they have had children, which affects their publication output. This in turn may reduce the likelihood that women will be appointed to editorial roles. However, the authors showed that for editors-in-chief, disparities in productivity or career length are unlikely to be entirely responsible for the low proportions of women in senior leadership roles. This indicates “a systematic role for non-meritocratic factors in the selection of editors-in-chief,” the authors write.

The changing demographics of the academic workforce will likely have a significant impact over time. More than 46% of doctorates were awarded to women in 2021 according to the US National Science Foundation. As long as women do not leave academia at a higher rate than men, the proportion of female senior editors will likely increase over the next decade. To improve gender diversity on editorial boards more quickly, one suggested step is to increase the number of early-career and mid-career researchers on editorial boards. This has the added benefit of better representing the needs of junior researchers, as well as training the next generation of editors, as well as likely increasing the proportion of women on editorial boards. This also makes sense from a commercial perspective. The move to OA means that most publishers are trying to increase the number of submissions to their portfolios. Authors submit to journals that they trust and that represent their communities. Productive female researchers are perhaps less enthusiastic about submitting their papers to journals that are entirely made up of men who are at advanced stages of their careers. Successful journals will likely have editorial boards that appropriately represent the gender, racial background, and career stage of the researchers they serve.

Authored by ChatGPT?

6

Last month we reported that ChatGPT had already been listed as an author on at least one published academic paper. Since then, the editors of Nature, Science, and JAMA have all weighed in with a definitive “no” on the question of whether the prolific chatbot (or other large language models, LLMs) might achieve authorship status in their respective journals. 

JAMA has, for example, updated its Instructions for Authors to read, “Nonhuman artificial intelligence, language models, machine learning, or similar technologies do not qualify for authorship.” 

Science has likewise updated its editorial policies “to specify that text generated by ChatGPT (or any other AI [artificial intelligence] tools) cannot be used in the work, nor can figures, images, or graphics be the products of such tools. And an AI program cannot be an author.”

Nature has outlined two principles for thinking of AI and authorship:

First, no LLM tool will be accepted as a credited author on a research paper. That is because any attribution of authorship carries with it accountability for the work, and AI tools cannot take such responsibility.

Second, researchers using LLM tools should document this use in the methods or acknowledgements sections. If a paper does not include these sections, the introduction or another appropriate section can be used to document the use of the LLM.

Authorship policies at many other publishers, as well as preprint servers, are currently under review but a consensus view aligning with that of JAMA, Nature, and Science appears likely (sorry ChatGPT!).

Dealmaking

7

De Gruyter has acquired the Mercury Learning and Information publishing house. De Gruyter has also signed an agreement allowing ResearchGate to redistributeits OA content.

Oxford University Press joins the many publishers partnering with Chinese presses to publish science journals, here taking on three OA titles with Higher Education Press.

Digital Science has acquired metaphacts, a knowledge graph and decision intelligence software company.

Association Management Software company Enforme has merged with the web design and development firm Wood Street.

Springer Nature has acquired TooWrite, a scientific writing tool.

People

8

Alondra Nelson has stepped down as Deputy Director of the OSTP and returned to her faculty position at the Institute for Advanced Study. Implementation of Nelson’s landmark memorandum requiring public access to published research will be left to others at OSTP.

John Martin has resigned as CEO of Brill.

The University of Toronto Press has merged its books and journals divisionsunder Antonia Pop, who has been named Vice President, Publishing.

Madeline McIntosh has resigned as CEO of Penguin Random House US, the country’s largest book publisher.

Usama Dar has joined Morressier as CTO.

Digital Science has named Leslie McIntosh Vice President of Research Integrity.

Adam Sewell joins Institute of Physics Publishing as CTO.

Briefly Noted

9

More than 50% of Cambridge University Press’s articles are now published OA, which is particularly impressive given that 60% of its publications are in the humanities and social sciences.

RELX released its 2022 financial report, showing a 9% year-on-year increase in revenue and a 15% increase in adjusted profit before tax.

Wolters Kluwer’s 2022 financial report shows revenue up 5%, profits up 7%.

eLife’s new publication model (covered by The Brief in October of last year), in which all papers selected to send out for peer review may be “published” at the author’s discretion (regardless of the review outcome) is now officially open for business.

MDPI published three of the top five largest journals by publication volume in 2022 and announced that it was the first OA publisher to publish one million articles. C&E’s James Butcher outlined the implications for scholarly publishing in this slide deck.

In largely unsurprising news, a study in Psychological Science suggests that rewarding research papers with open science badges is not a particularly effective route to ensuring reproducibility.

PLOS has updated their Plan S Price & Service Transparency disclosure to include 2021 figures. While PLOS should be applauded for their continuing efforts to share information with the community, by their own admission in the report these disclosures are at best a blunt instrument. There seems to be no category for many of the community-supporting activities PLOS provides, and no listing of any financial surplus generated, an essential component of sustainability for all organizations, both for-profit and not-for-profit. 

Touted as a “program designed to flip existing subscription-based journals to a diamond open access publishing model,”  MIT Press’s  shift+OPEN offers funding to cover costs for one journal to move to Diamond OA for three years while it tries to figure out a strategy to survive beyond that period.

The American Association of Publishers has announced this year’s PROSE Award winners for “professional and scholarly excellence” (thus PROSE). Congratulations to the winners! We await selection of the R.R. Hawkins Award winner, who will be chosen from among the category winners.

How is ChatGPT like a fuzzy JPEG? Ted Chiang explains in an excellent piece in The New Yorker.

In an article in the Italian Journal of Library, Archives, and Information Science, author Andrea Bonaccorsi offers a scathing rebuke over the lack of economic analyses accompanying OA proposals and regulations, which has resulted in so many unintended consequences. While it should be noted that Bonaccorsi’s explanation of how a shift from intrinsic to extrinsic motivators would shift peer review behavior largely mirrors a 2015 Scholarly Kitchen blog post by C&E’s David Crotty, the author does offer an intriguing solution of having research societies help organize pools of qualified reviewers in return for a share of author payments for editorial services.

Cory Doctorow offers a dismal explanation of the life cycle of online platforms such as Amazon, Google, Facebook, Twitter, and more recently, TikTok: “…first, they are good to their users; then they abuse their users to make things better for their business customers; finally, they abuse those business customers to claw back all the value for themselves. Then, they die.” We observe that all of these platforms remain alive and profitable but Doctorow may be right about the direction of travel for value on (at least some of) the platforms. 

Congratulations are due to PeerJ on their 10th anniversary, marking an unexpected longevity as an independent publisher for an organization that was founded with venture capital funding. PeerJ’s ability to pivot multiple times and implement new business models has likely been key to its ongoing success.

Which is harder, defeating the final boss in the video game Elden Ring or navigating the career path of a tenure track professor? McSweeney’s notes that the two activities are not all that different.

***
All the secrets of the world are contained in books. Read at your own risk. — Lemony Snicket (Daniel Handler)