Untransformed

Untransformed

1

If a university sets a pass mark for its students and 68% fail to meet the grade, is that the fault of the university for setting too stringent a target? Or the fault of the students for not being able to procure the resource they needed to prepare for the exam? What would the reaction be if students who missed the target by 0.1% were expelled from the university?

On June 20th, cOAlition S announced that 68% (1,589) of the 2,326 transformative journals (TJs) failed to hit their targets and will be removed from the program. The targets (set without, so far as we are aware, a substantive market study) were to increase the proportion of OA research content in 2022 by at least 5 percentage points in absolute terms and at least 15% in relative terms, compared with 2021. Between them, Springer Nature (1,721 journals), Cambridge University Press (240 journals), and Elsevier (182 journals) publish 92% of the 2,326 participating journals. 77% of Springer Nature’s journals missed the target compared with 63% for Elsevier and 35% for Cambridge.

Authors who choose to publish OA in those 1,589 delisted (“untransformed”?) journals can no longer use cOAlition S funding to pay for associated article processing charges (APCs). Authors can still publish in such journals and adhere to the policies of Plan S, as long as the journal is in a transformative agreement; the author pays for the APC using other funding sources; or the author deposits a copy of the author accepted manuscript under a CC BY license in an appropriate repository(assuming the journal allows that to happen). 

cOAlition S published the full data set, which contains some interesting stories. Perhaps most notable is that half of the 1,589 journals that were expelled from the TJ arrangement published more OA papers in 2022 than in 2021, but not enough to meet the Plan S targets. Between them, the delisted journals published 33,020 research papers OA in 2021 and 36,436 OA papers in 2022, a 10% growth rate in OA content. However, the number of subscription articles only reduced by 948 articles (from 237,740 in 2021). In other words, the delisted journals, taken as a collective, published additional papers OA rather than flipping subscription content to OA. Publishing more papers on an OA basis seems like a success story to celebrate for an organization that is advocating for more OA publishing. It is puzzling to us that rather than celebrating this as a success (10% OA growth!), Robert Kiley, Coalition S’s head of strategy, has called the results “clearly disappointing.” We can understand that it is not an ideal outcome from the coalition’s perspective but it is forward motion.

A perhaps less equivocal success story is that of the Nature Portfolio. Of the 36 Nature journals, 35 met the targets and remain as TJs. Some of the growth in OA uptake is impressive, especially given the $11,690 APC required to publish OA in a Nature journal (excepting authors covered by transformative agreements or by the waiver policy). For example, 23% of Nature Medicine’s content was OA in 2021, but a year later that had increased to 59%; subscription research articles fell by 42% during that period. Across the 36 Nature journals, the number of OA articles tripled from 484 in 2021 to 1,461 in 2022. Meanwhile, the number of subscription articles fell from 4,092 to 3,619 – a 12% decrease across the portfolio. 

Cell Press journals did not fare as well in the cOAlition S announcement. Elsevier submitted only its biomedical Cell Press journals to be TJs: roughly half of those met their targets and the other half were delisted. Other aspects of the TJ report are examined in James Butcher’s Journalology newsletter.

It is notable that hardly any of the leading medical journals applied to be TJs. Partly this is because the OA business model is especially problematic for journals that publish low numbers of articles and that rely on commercial revenue streams, such as reprints, advertising, and sponsorship. Their exposure to cOAlition S–funded authors also tends to be low, as many of the top medical journals are based in the US (see our discussion of the Nature Index in #2 below).

Those hybrid journals that have met targets and kept their TJ status will only be able to receive payments from cOAlition S funders for the next 18 months; at that point authors will need to publish in fully OA journals if they want to use cOAlition S funding to pay for APCs. In other words, even though they have met targets (so far), the journals participating in the TJ program will nonetheless soon be in the same boat as non-participating journals. It is unclear how problematic this is for researchers. Many researchers with cOAlition S funding are also covered by transformative agreements – which are, we emphasize, not funded by cOAlition S. Other authors may have access to institutional funds for APCs. Still others may publish in hybrid journals that allow deposition of author accepted manuscripts with CC BY licenses.

The impact on publishers that have delisted journals depends largely on their journals’ exposure to researchers who receive cOAlition S funds, inclusion in transformative agreements, and current subscriptions. The vast majority of hybrid journals are looking to manage the transition to OA in a steady and sustainable way. Despite cOAlition S’s best efforts, demand from authors for OA is not growing fast enough to justify a “flip” to fully open, at least in hybrid journals. In addition to the math not working, many publishers are concerned about equity and the impact on authors without sufficient funding for APCs. The OA transition is a marathon, not a sprint. And so this likely marks the end of the TJ route to compliance for most participants. As we noted two years ago, few journals were likely to succeed in meeting Plan S’s targets (as they were not based in market reality), but applying for TJ status for a journal was a “sure, why not?” way to get a couple of years of Plan S compliance with no downside or penalty. The extra time that TJ status allowed to bring in cOAlition S funds for APCs contributed, if minimally, to the journals’ bottom lines – and the good press associated with the TJ effort was likely a net positive.

We anticipate that fully OA publishers will view dwindling participation in the TJ program as an opportunity to further boost market share (indeed, Frontiers was quick to weigh in on the TJ assessment). That said, some perspective is perhaps helpful here: all 2,326 TJs combined published 56,805 OA articles in 2022, which equates to only slightly more than a month’s output from MDPI.

China Research Metrics

2

The growth of academic research from China is one of the defining trends of the past decade. Like all countries, some of the research from China is good and some of it is bad. However, there is mounting evidence that an increasingly large proportion is truly excellent, with the latest data point coming from the Nature Index, which released its 2022 results this month.

Nature Index’s key metric is called Share, which measures a nation’s (or academic institution’s) contribution to the articles published in a cadre of 145 influential journals selected by independent editorial boards. This year, China’s Share in the natural sciences increased by 21%, eclipsing the US for the first time. The full data set shows that most countries in the top 10 of the natural sciences category saw their Share decrease year-on-year. China and India were the only two countries to improve between 2021 and 2022.

The Nature Index is based on the articles published in 145 journals. This year, for the first time, the Index includes journals from the health sciences. Nature Communications, the largest journal included in the Index, contributed 7,341 articles (10% of the total). By contrast, Arthritis & Rheumatology, the smallest journal, contributed just 14 articles. This means that journals that publish a large number of papers contribute disproportionately to the Index. Indeed, the top 10 journals, ranked by article output, published 40% of the articles included (the full list can be found in this spreadsheet). Many of the large journals are in the fields of chemistry and materials science and so the natural sciences category is heavily weighted toward those subject areas. China’s research strength is in the physical and applied sciences, so it does particularly well there.

China is much weaker in the health sciences category of the Nature Index; the US has a Share that is more than four times higher than China’s. This is unsurprising given that many of the most influential clinical journals are based in the US, with many of them being included in the Nature Index.

A better way to assess a country’s output, in our opinion, is to assess the most-cited papers, regardless of which journal they are published in. In March 2022, a paper published in Scientometrics found that China had participated in more highly cited (top 1%) papers than the US in 2019. 

But perhaps the best data source for publishers who want to understand how many high-impact articles the leading research nations produce, by discipline, is the annual Digest of Japanese Science and Technology Indicators. This report has a global perspective and is an important resource for publishers. Page 13 of the 2022 annual report lists the fractional share for the top 10 countries over the past two decades, based on the top 10% and top 1% most-cited articles.

The chart at the bottom of page 14 clearly demonstrates how strong China is in materials science and chemistry, especially in the high-impact space. China’s share in clinical medicine is much lower, which tallies well with the Nature Index findings.

Every publisher, regardless of discipline, should have a growth strategy for China. Publishers can use bibliometric tools to identify key individuals and institutions to target with editorial outreach and marketing. Publishers should be working with their editorial team to create an outreach strategy that includes face-to-face meetings, desk outreach, commissioning, and peer reviewing. However, the biggest challenge for western publishers is understanding the cultural and political landscape in China. In that regard, C&E’s recent report was written to help publishers successfully navigate the complex, but vitally important, research landscape in China.

Tectonics

3

Speculation that the scholarly communications market will coalesce into a single “supercontinent” dates back at least eight years, and using the “Pangea” metaphor, the tectonic movements of geological shifts take time but grind inexorably onward. Recent news on this front includes a report from Elsevier on their first year of efforts to bring in non-Elsevier content to ScienceDirect. 100,000 Organic Chemistry and Transportation articles from 38 journals (coming from the American Chemical Society, the Royal Society of Chemistry, Taylor & Francis, Wiley, and the International Union of Crystallography) were made available through ScienceDirect (the full text of the article was available for OA papers, and a link to the PDF was shown only to subscribers for content behind a paywall). Overall, the papers saw an unsurprising increase in reach, with notably more visits from countries outside North America and Europe. The experiment is slated to expand in the next year to further research topics, with participation from the American Society for Microbiology and Brill.

Elsevier’s primary competitor for supercontinent dominance is ResearchGate, who, in an act that would have seemed impossible just a few years ago, signed on to the STM industry-led GetFTR service. ResearchGate’s prior refusal to check content for copyright infringement led to scores of takedown notices from publishers and a 2017 lawsuit that went against the company and in favor of Elsevier and ACS. Over that time, ResearchGate has slowly capitulated and been assimilated into the mainstream of the publishing community, appearing on panels at industry meetings and signing more and more partnership agreements with publishers to seek their cooperation in hosting their content (the latest last month with Frontiers). 

What’s interesting about these deals is that most of them only allow hosting of OA content, something that’s already legally within ResearchGate’s rights without further permission from the publisher. Why bother negotiating an agreement for using CC BY content when the license already grants you the ability to do whatever you want with it? We suspect this is more of a long-term strategy. ResearchGate has built a huge user base and, unlike publisher websites, users must sign in individually to use the site’s full set of features. This means that ResearchGate, in many cases, has much better information about who is using publisher content than publishers. Providing usage data to publishers may be a small price to pay to build relationships with them while ResearchGate bides its time and positions itself as an aggregator. The integration of the GetFTR access control system is a further move to that end.

A potential shift in the supercontinent market, however, has emerged this month with CORE. Funded by Jisc and the Open University, CORE is a not-for-profit, offering the world’s largest collection of OA articles. Although seemingly aimed more at powering text and data mining for machine reading rather than discovery for human reading (at least initially), CORE’s open APIs (application programming interfaces) offer a route for the fast accumulation of a corpus of content, without having to rely on cooperative agreements or user uploads, to potential supercontinent competitors to ScienceDirect and ResearchGate. As we move toward more of the literature being published OA, open infrastructure such as CORE may make for a more competitive market.

People

4

Jane Bunker, Director of Cornell University Press, has taken office as the new President of the Association of University Presses.

Mandy Cohen has been chosen to serve as the next leader of the Centers for Disease Control and Prevention.

Jessica Lawrence-Hurt has been named the new Director of Marketing for GeoScienceWorld.

Jennifer Kemp is now the Director of Consulting Services for Stratos.

Digital Science did some reorganization of its executive team, including shifting Steve Leicht from the COO role to that of the first President of Digital Science.

Robert “Jay” Malone has stepped down as Executive Director of ACRL, and is succeeded by Interim ACRL Executive Director Allison Payne.

LIBER announced that Martine Pronk, current Interim Executive Director, will take on the position of Executive Director permanently. 

Stefan Tochev, formerly MDPI’s Head of Marketing and Communications, has been named the publisher’s Chief Executive Officer, with Shu-Kun Lin stepping aside to focus on his role as Chairman of the Board.

Briefly Noted

5

Just as The Brief was going to press, Clarivate released the 2022 edition of its Journal Citation Reports, and along with it 2022 Journal Impact Factors (JIF). The biggest change this year is the expansion of the JIF to additional indices. Journals indexed in the Arts & Humanities Citation Index (AHCI) and the Emerging Sources Citation Index (ESCI) will be issued JIFs for the first time. A second notable change is that Clarviate now displays the JIF to one decimal place, which is likely to create more ties but is an improvement in our view as it eliminates the false-precision of the previous three-decimal reporting. We’ll have more to say about this year’s JIFs in next month’s issue.

In further Clarivate news, after facing criticism for their opaque approach to journal delistings, Clarivate has responded with some efforts toward transparency, now offering a monthly download showing changes in Web of Science coverage.

With the post-COVID decrease in article publication and the ongoing fallout from problems with special issues at Hindawi, Wiley’s financial difficulties continue, as detailed in their latest financial report.

In AI news, Nature has banned the use of generative AI images and video in published papers. Japan has announced that using content for AI training does not infringe upon copyrights. SciWriter has introduced a new seven-session course on AI tools related to authoring, discovery, and research integrity (hosted by founder Avi Staiman). And a new preprint in arXiv suggests that, while human-created content on the internet has been an essential training tool for AI, as the internet increasingly fills with AI-created content, it will result in “model collapse.” The idea is that as more and more AI-created content is used to train AI, the “polluted” training data will eventually result in gibberish within a few generations.

The National Science Foundation has released its public access plan. As expected, it closely follows the draft National Institutes of Health plan in requiring public access (not open access) to the author accepted manuscript version of research papers funded by the agency. As per the Nelson Memo, of course, it requires immediate access without a publication embargo. The only surprise is that the effective date for the policy is January 31, 2025, which is 11 months earlier than required by the Nelson Memo. Juried conference proceedings are also subject to public access requirements, and while the National Science Foundation is considering non-peer-reviewed proceedings as well as book chapters and other types of documents, they are not yet included.

NASA has released a request for input on its draft Public Access Plan. Responses are due by August 17, 2023.

Project COUNTER has opened up their member survey to everyone, whether you are a member or not. They are seeking input to help shape COUNTER’s tools and services for the future.

Oxford University Press and Silverchair unveiled Sensus Impact at the Society for Scholarly Publishers. The aim of this initiative is to aggregate article usage statistics, alternative metrics, and citation data – from across both publishers’ websites and aggregator platforms – in a dashboard. The dashboard will provide a view of these data for research funded by particular agencies (e.g. National Institutes of Health, National Science Foundation) for use by funders, authors, and policymakers (among others).

A study on published papers using the registered reports format concludes that the approach is “still marginal and few journals make use of it.”

An article in El País looks at “hyper-prolific” researchers – those who publish papers every few days throughout the year. Researchers profiled seem to be putting their name on work, in subjects in which they are not experts, from coauthors they’ve never met. We particularly note the use of the term “MDPI Professors” to describe those who have risen in their careers through CVs filled with “shoddy work.”

We agree wholeheartedly with the argument that journals not already accepting free-format submissions are creating unnecessary friction for authors, but some of the methodological approaches to assigning dollar figures to issues in scholarly communication remain vexing. In a recent, widely publicized study, a group of epidemiologists concluded that $230 million was “wasted” by authors in 2021 reformatting their papers for submission to new journals after rejection. Many of the assumptions made in the process of coming to that dollar figure are dubious. For example, the authors use an estimate of 1,750 work hours per year for researchers based on an OECD study of labor in general – personally, we at The Brief know of few biomedical graduate students working 35-hour weeks (or at least few who end up publishing in the top journals that were studied). The analysis makes a seemingly unfounded assumption that all papers are rejected and submitted to a different journal an average of once per paper, based on a similar unfounded assumption from a previous paper rather than actual data. There’s also no control performed for the journals in their experimental set that offer free-format submissions (nearly 500 Wiley journals offer free-format submission and more than 650 Elsevier journals offer “Your Paper, Your Way,” for example). These quibbles to one side, their ultimate point is a good one: it’s dumb to make researchers waste time reformatting figure legends, references, and other components of the paper until it is close to acceptance.

EDP Sciences announced that their journal Radioprotection failed to meet its Subscribe to Open threshold and will remain a subscription journal for the next year. 

From the department of You Can’t Make This Stuff Up: Scholar who studies dishonesty accused of fabricating findings.

Finally, in a bizarre intellectual property overreach, Apple seems to be trying to block the use of images of apples by everyone, including those selling the actual fruit. We hope this attempt to restrict produce-related imagery is not a trend that will expand (looking at you Plum Analytics).

***
There was an exchange on Twitter a while back where someone said, ‘What is artificial intelligence?’ And someone else said, ‘A poor choice of words in 1954.’ And, you know, they’re right. I think that if we had chosen a different phrase for it, back in the ’50s, we might have avoided a lot of the confusion that we’re having now.  –Ted Chiang (in conversation with Madhumita Murgia, FT)