Cancelled Check

Cancelled Check

1

What is a publisher to do in the light of increasingly contradictory messages being sent by funders around open access (OA)? As noted in last month’s The Brief, Jisc recently released a report on “transitional agreements” (which everyone else calls “transformative agreements” but at least the acronym, “TA,” is the same). As discussed by Alison Mudditt inThe Scholarly Kitchen, the report made it clear how unsatisfied the organization is with the slow pace, according to Jisc, of publishers flipping their Hybrid journals to become fully OA. Jisc sees TAs as failing to deliver on their promise of being a “temporary and transitional” activity meant to shift the way an entire industry does business. They are not alone in this sentiment, with cOAlition S declaring a “growing frustration in the community regarding the slow transition to OA of journals included in so-called transformative agreements.” cOAlition S is ending support for such agreements (along with transformative journals, TJs) at the end of 2024, stating that it was “clear that the transformation would have to be completed at a definite point in time, by the end of 2024 at the latest.”
 
It is perhaps beside the point (but we will make it nonetheless) that TAs were not designed to transition or transform journals. If a journal has 3,000 institutional subscribers around the world, and 100 of them, mostly clustered in Europe, wish to license journals under a transformative model (e.g., publish and read), that means the other 2,900 institutions prefer a standard subscription model. A journal publisher is not going to flip the journal to OA for a small fraction of its customer base when most of its customers prefer another model. TAs were designed to transform the research output of a given institution to OA. If enough institutions, over time, opted for TAs, then more and more content would be OA and eventually the journal would flip entirely. Everyone agrees on this point, but there are apparently wildly different expectations over what “eventually” might mean. 1 year? 10 years? 100 years?
 
Several funding agencies and some consortia that have signed TAs lean toward the shorter end of the expectations scale (although a rationale for such expectations has not yet materialized) and seem to expect that publishers would have shifted their business models already! As we puzzle through this expectation for extreme alacrity, we wonder if some research funders have never actually met a publisher? 
 
The thing about flipping journals to fully OA is that it can be hard to do without relying on author-pays models to maintain economic sustainability. At the same time, those agencies (and others) have declared the end of the article processing charge (APC) as an acceptable business model for journal publishing. Jisc, along with cOAlition S, has convened a working group on moving away from the APC model, declaring that it “creates a dysfunctional system.” The Bill & Melinda Gates Foundation, a core member of cOAlition S, has announced a new policy in which the Foundation will no longer pay for APCs, declaring them to be inequitable and that they have led to “some unsavory publishing practices.” 
 
On the one hand, the funders are not wrong. APCs do create inequities and incentivize unsavory publishing practices, as critics of author-pays OA models have pointed out for many years. On the other hand, these same funders have been actively promulgating policies that inexorably lead to author-pays models, including APCs and TAs (which are effectively packaged APCs). 
 
If you’re a publisher that has made an honest effort to work with the demands of these important funders, then perhaps you’re feeling a bit like the rug has been pulled out from under you. Imagine the publisher who signed up for cOAlition S’s transformative journal program and pledged to flip their journals to OA only to find, at the end of the transition process, they receive a cancelled check. 
 
Case in point, Cambridge University Press (CUP), which has been one of the most aggressive adopters of TAs since the 2018 launch of Plan S, signed deals with over 2,000 institutions. CUP jumped into these waters fully knowing that these deals were not sustainable, noting in 2019 that in order to keep the lights on “further publishing services and brand-new models” would need to emerge. However, as a representative of CUP presented at this month’s UKSG meeting, such new models and services for books and journals do not yet exist. As a result, CUP’s business in those areas “is not currently profitable due to the level of investment required.” And these are presumably retrospective numbers, not accounting for the recent Gates announcement nor the 2024 cOAlition S phaseout of support for TAs and TJs (in other words, the numbers were from the period before the checks were cancelled).
 
One lesson here is that while one can’t ignore research funders, they may not be the best source of business strategy. As librarian Arthur J. Boston puts it, the OA movement consists of a global community passing “wait, no, not like that” back and forth forever. Building the plane while flying it may be acceptable to funders whose grant recipients have no choice but to comply with the latest set of demands, but it is a risky business strategy – particularly when funder policies point in different directions and frequently change course. 

Gates Plan U-Turn

2

Policy is a blunt instrument that functions best when requirements are clearly stated and adequate support is provided. The new OA policy “refresh” from the Bill & Melinda Gates Foundation falls short on both these counts.
 
The new policy is essentially the old policy, with two important distinctions. Gates-funded researchers must now post a preprint of any “Funded Manuscript” (effectively adopting Plan U) in addition to following the previous rules that require immediate availability under a CC BY license of any published paper’s accepted manuscript. The other key difference is that (as we discuss in Item 1, above) Gates will no longer pay any publication fees associated with OA, or at least any APCs will not be covered by grants from the Foundation.
 
One could immediately see that there might be some turbulence associated with the policy from its rollout, which consisted of a webinar, and a blog post, followed by an announcement of a new Gates preprint server by F1000. All of this took place in a vacuum, as the actual policy wasn’t made available for nearly another two weeks. Once the policy did become available, more confusion ensued as it is full of ambiguity, creating uncertainty for researchers hoping to remain in compliance. Gates Program Officer Ashley Farley even went so far as to acknowledge that “the policy language is confusing.” 
 
Farley’s attempts to clarify the policy unfortunately resulted in creating further ambiguity by suggesting in the same tweet that deposit of “a preprint will fulfill the policy” (as opposed to deposit of an accepted article in PubMed Central or another open repository, as stated in the policy). It is early days and perhaps they will update the policy to match what’s suggested in the tweet. Or not. As librarian Rick Anderson observes, “Sometimes ‘the language is confusing’ means ‘you have understood the language correctly but we’d prefer not to publicly acknowledge its meaning so please stop talking about it.’” 
 
Gates states that the new policy is driven by a goal of increased equity, which is problematic as the new requirements are clearly less equitable than the ones they’re replacing. By requiring OA to the article’s accepted manuscript but not paying to support it, Gates now creates an uneven playing field for its own researchers – those with funding outside of Gates or at an institution with TAs in place will be able to publish where they want, using those funds or agreements to cover the necessary APCs. Those who aren’t quite so well-off (or lucky to be in an institution that has signed numerous TAs) will be severely restricted in their publication outlets to those journals that allow immediate Green OA deposit. 
 
It is worth remembering that most actual scientific research is done by graduate students and postdocs, not the principal investigators (PIs) who run laboratories and are awarded Gates funding. A preprint may indeed suffice for those senior, tenured PIs looking to stay in the good graces of their benefactor, but it’s certainly not enough for the students/postdocs hoping to build a career and who don’t get to keep Gates funding once they move on to their next station. They will need as strong a publication record as possible if they hope to land one of the increasingly rare jobs in academia, and so will have to find separate OA funds to pay for those publications in order to keep their bosses compliant with Gates.
 
Despite the recent proliferation of preprint services, Gates has created another one with F1000 called VeriXiv – a “verified preprint platform.” VeriXiv will be open only to Gates-funded research to start with, although F1000 does appear to have plans to make similar services available to those able to pay a fee. VeriXiv purports to be different from other preprint servers as “Twenty different ethics and integrity checks will assess a range of issues, including plagiarism, image manipulation, author verification and competing interests” in order to “ensure the scientific validity of research.” Supporting research integrity checks is laudable but there are many problematic issues with scientific validity that have nothing to do with the kinds of issues found by these checks. It is possible (and all too common) to write a completely original paper with unmodified data that either deliberately or mistakenly misinterprets that data, in whole or in part, and draws unsupported conclusions. That is something peer review and a good editor are designed to catch.
 
It is hard to know what to think about a research policy that on the one hand implies that the research Gates funds does not require peer review (at least by one reading of their policy-by-tweet, which states that a preprint alone is sufficient for compliance), and on the other hand suggests that the research they fund needs 20 different ethics and integrity checks.
 
There is an urgent need to explore new routes to OA and to move away from the incentives and inequities created by the author-pays APC model. Bringing preprints more widely into the workflows of researchers offers intriguing possibilities for policies that better serve the research community. But as with all policies, the details matter, and here we are left wondering if this policy might itself benefit from a round of peer review. Revise; resubmit. 

Shadows and Light

3

Some new data on consolidation in the scholarly journals market emerged this month in a preprint by Rob Johnson and Elle Malcolmson. While we are biased toward looking favorably on any paper that quotes Joni Mitchell in its title, the small sample sizes and low quality of publicly available financial data for the publishing revenue of research societies makes many of the preprint’s conclusions difficult to accept.

Looking at a dataset of 277 UK research societies tracked from 2015 to 2023, the study does indeed show increased levels of outsourcing of journal publishing operations from societies to third parties. The number of wholly self-publishing societies in the dataset dropped over the eight-year period from 68 to 44 – a 35% decline, which to be fair, includes 6 societies that stopped publishing altogether. The study notes an increased “diversity of partners” as more societies are farming out their journals to multiple publishers than was the case in the past.

The study concludes that partnering has largely resulted in lower revenue for societies than seen for those that remained independent. The numbers, however, are not convincing. Only 21 of the 50 largest societies by revenue provided any public data about publication revenues, and the most recent financial data used ranges from 2020 to 2023, so different timespans are being compared for different publishers. Of that varied small sample size analyzed, 7 of the 8 societies that stayed self-published over the period studied grew their revenue roughly in line with inflation. Four of the 6 that outsourced publishing through a partner in both years saw lower recent revenue listed than that for 2015. But raw revenue numbers don’t really tell the whole story.

Total revenue is not what matters in analyzing the performance of a publishing partnership, because partnering with a larger publisher generally results in significantly lowered costs. Independently published societies must pay for everything – a submission system, hosting platform, production, sales, etc., etc. In a partnership model, these expenses are typically covered by the publishing partner, meaning that a society can bring in less overall revenue yet net significantly more because expenses drop. 

Looking at two data points also fails to tell the whole story that a trend analysis might illuminate. How many of those societies were already seeing revenue declines before partnering (and perhaps reduced their rate of decline by partnering)? Often a society will accept lower overall returns in a partnership if those returns are guaranteed, lowering financial risk in an increasingly turbulent market. Worth also considering is the year-to-year variance in some publishing agreements, which due to signing bonuses or other royalty structures, can be significantly front-loaded, so year one of a deal can look a lot stronger than later years.

There are also some qualitative differences between many of the societies compared. Perhaps those societies that were doing well financially were less driven to partner than those that were struggling, thus biasing the samples. Does comparing the Institute of Physics Publishing, with over 100 journals, to societies like The Physiological Society or the Royal College of Radiologists, who each publish three journals, tell us more about the publishing model or the scale of operations? The study does make the point that scale matters, as the 10 biggest societies in the 21 examined saw median growth while the 11 smallest saw median declines.

In the end, the study’s “compelling evidence for retaining, or even reclaiming, their [societies’] independence” is perhaps less than absolutely compelling, and certainly not broadly applicable. The decision to partner with a larger publisher is often complex and nuanced, and it cannot be understood merely by looking at topline revenue figures. There are many trade-offs to consider in the decision to partner or remain independent, and in our experience at C&E, careful, in-depth quantitative and qualitative analyses are necessary to understand what’s best for each individual publishing program. To stay with Joni Mitchell, “Every picture has its shadows, and it has some source of light.”

Dealmaking

4

American Journal of Preventive Medicine Acquired by Elsevier. C&E represented the Board of Governors of the American Journal of Preventive Medicine (AJPM) and AJPM Focus in the sale of both titles to Elsevier, AJPM’s long-time publisher. The journals have long been official titles of the American College of Preventive Medicine (ACPM) and the Association for Prevention Teaching and Research (APTR) – and they will continue as such. C&E is honored to have advised the AJPM Board of Governors on this notable deal. 

ASPET Partners with Elsevier. The American Society of Pharmacology and Experimental Therapeutics (ASPET) has signed an agreement with Elsevierto publish the society’s four journals. The decision resulted from a carefully considered strategic analysis, supported by C&E, of the benefits and risks of remaining self-published vs. entering a publishing partnership, following by a highly competitive RFP process. Long-term growth and security for the society guided ASPET in its ultimate decision to partner with Elsevier. C&E is proud to have supported ASPET in this process and congratulates both parties on the new partnership.

Briefly Noted

5

As we await a proper financial analysis from the US Office of Science and Technology Policy, budgets for US federal funding agencies across the board appear to be mostly flat, if not actually declining, providing scant resources to support the policy requirements of the Nelson Memo. 

Angela Cochran and Holden Thorp both want to see more effort from research institutions put toward ensuring that their researcher employees are behaving ethically. The US federal government, through the Office of Research Integrity, agrees. Universities, however, seem adamant in avoiding as much responsibility as possible.

Is Science Reddit the new Science Twitter? This may be problematic given the ongoing spam issues seen for Reddit. 

A new study published in Proceedings of the Royal Society B that looks at the language barrier and biases against non-English speaking scientists offers some interesting correlations. The higher a journal’s Impact Factor, the lower the number of inclusive policies it is likely to have. Society ownership of a journal showed a positive correlation with inclusivity. And neither the proportion of OA articles published, nor the number of editorial board members from non-English-speaking countries, had a major positive association with the adoption of linguistically inclusive policies.

Science is self-correcting, but sometimes the wheels grind very slowly. Just this month, the Elsevier journal New Microbes and New Infections posted a stunning 101 expressions of concern on papers by Didier Raoult, the researcher chiefly responsible for the promotion of hydroxychloroquine as a COVID cure. This brings Elsevier up to speed with PLOS, which flagged 50 of Raoult’s articles back in 2022.

David Clark from Oxford University Press suggests that scholarly publishers need to take a more active role in the development of generative AI to help ensure the accuracy of GenAI results, as well as the sustainability of those that create the materials used for training AI. Wiley would seem to agree, signing a $23 million content rights project “with a large tech company.” 

Tasha Mellins-Cohen offers a tremendously useful framework for keeping track of the ever-expanding variety of OA business models.

Another year, another OA manifesto: The Barcelona Declaration on Open Research Information calls for free and open bibliographic information and databases to better track the scientific enterprise. 

While journal article retractions reached an all-time high in 2023, Christos Petrou points out that these retractions are not evenly distributed, and that the industrial-scale work of papermills and the growth of research from China are at the heart of retraction growth.

No surprise here – a preprint survey of how biology researchers assess credibility of grant applications or on hiring committees found that the majority use journal reputation or the Journal Impact Factor.

Universities continue to gut their humanities programs, much to their own detriment. Even if you buy in to the focus on STEM education above all other subjects, Holden Thorp suggests that the key to building public trust in science is to teach more philosophy.

In a welcome but sadly rare moment of unity, every major publisher has backed a lawsuit meant to block book banning in Iowa.

Wait – PLOS ONE didn’t have an author proof stage in its workflow? Apparently not, so after 18 years of a higher correction rate than other journals, this essential step will finally be added.

***
If an academic community wishes to control the business decisions of a journal, the best way to do that is to own the journal. – Michael Clarke, Nature