The Journalology newsletter helps editors and publishing professionals keep up to date with scholarly publishing, and guides them on how to build influential scholarly journals.
Journalology #60: Resilience
Hello fellow journalologists, I spent most of last week in Berlin, attending the Academic Publishing in Europe (APE) meeting. As is so often the case with conferences, much of the value was in the conversations with colleagues during the breaks, although many of the sessions were excellent too. The core themes centred around AI and research integrity (and how those two topics interface with each other). I thoroughly enjoyed the conference and would like to thank Marta Dossi and her team for all of their hard work and for sponsoring this newsletter at the end of last year. NewsThe next chapter for Frontiers: to the open access tipping pointIn the current market conditions, we are increasing efficiency further by proposing to significantly resize our workforce of about 2,000 employees across 23 countries to about 1,400 employees. Together with the management and operational changes made in 2023, this makes Frontiers a leaner and more agile organization, with enhanced financial resilience amidst market shifts. Frontiers (announcement) JB: Last year many technology companies downsized, with thousands of workers being laid off. This week’s announcement from Frontiers that 600 jobs will go follows that trend and is indicative of the risks of the B2C (business-to-consumer) open access APC model. I was expecting an announcement along these lines at some point in the first quarter of this year, but it’s always a shock when the news breaks. As I explained in issue 58: In the month of October 2022 the Frontiers portfolio published 8516 research articles, according to Digital Science’s Dimensions database. In October 2023 that number fell to 4170 research articles, a decrease of more than 50%.
The drop in revenues will be significant, especially since the company’s overhead increased from 762 staff at the end of 2020 to over 2500 staff now.
Put another way, in the first 10 months of 2020 Frontiers published around 29,000 research articles (according to Dimensions) and had 760 staff. Over the same period this year, the organisation published 58,000 research articles and ‘over 2500 staff’. So, all things being equal, revenues doubled, but the overhead cost base tripled or perhaps even quadrupled (yes, there are a lot of assumptions in that statement, but it’s surely true to say that costs have increased a lot faster than revenues, because of the recent downturn in article output).
My thoughts are with everyone who’s been affected. Businesses become successful because of their people, and when commercial fortunes change it’s often the rank and file workers who are affected the most. The core narrative in the announcement is that article volumes dropped because Frontiers chose to increase its rejection rate to maintain quality levels. The pandemic also brought along an industry-wide trend for more fraudulent manuscripts. Frontiers responded with sweeping audits of its articles, journals, and Research Topics, introduced additional research integrity controls, and deployed AI-powered quality checks. These additional controls led to an increase in rejection rates, underscoring our commitment to quality. This can’t be the whole story, however, since the downturn in article output was too dramatic for it to be solely due to increased rejection rates. As a reminder, here’s Frontiers’ output (articles per calendar month) between January 2020 and December 2023, based on Dimensions data. Growth was fairly steady between Jan 2020 and Dec 2021. There was much more in-month variation in 2022. For reasons that aren’t entirely clear, there was a massive jump in January 2023, perhaps due to the Christmas backlog, although the spike looks too big for that to be the sole reason; note that there was a similar spike in January 2022, January 2021, and January 2020 (I have theories for why this might be, but I won’t air them here). However, from February 2023 onwards article output dropped. The day before the job cuts were announced, Frontiers published another article entitled Safeguarding peer review to ensure quality at scale which set the scene for what was to come the following day. This, admittedly long, article contains the word ‘quality’ 53 times. The message is clear. Frontiers’ primary focus is firmly on quality. The article includes some data on rejection rates and article volumes, from which we can gain some insight into what happened to submissions between 2021 and 2023. According to these data, in 2021 Frontiers published 85,273 articles and had a rejection rate of 40%; in 2022 it published 125,423 articles and rejected 47% of submissions; and in 2023 it published 90,446 articles and rejected 56% of submissions. Rejection rates likely changed gradually over the course of the year, but we can use these numbers to estimate submission volumes for calendar years.
This suggests that the downturn in submission volumes between 2022 and 2023 was relatively moderate. However, it’s important to remember that the 2023 submission numbers quoted above are over-inflated by the January 2023 peak. The picture looks far more stark if you compare January 2023 with December 2023: ~15,000 articles were published in January compared with ~5000 articles in December, which is a precipitious drop that can’t just be caused by higher rejection rates. Let’s do the maths. In January 2023 Frontiers published 15,163 articles, according to Dimensions, which equates to 28,500 submissions at a 47% rejection rate. By contrast, in the month of December 2023 Frontiers published 5043 articles, according to Dimensions, which would equate to 11,500 submissions based on a 56% rejection rate. In other words, submissions look to have fallen by around 60% between January and December 2023, once you take into account the change in rejection rates. N.B. This is a back-of-the-envelope calculation with many assumptions; the figures are unlikely to be entirely accurate, but they strongly suggest that submissions to Frontiers journals dropped significantly throughout 2023 and that the increased rejection rates are only part of the reason for the reduced article output. What can we learn from Frontiers’ experience over the past few years and what might this mean for other publishers in 2024 and beyond? First, companies that are largely dependent on one revenue stream are less financially resilient than businesses that have many product lines. The likes of RELX and Springer Nature have diversified businesses within their portfolios, some of which are outside of scholarly publishing; this makes large-scale job losses far less likely. I’ve speculated in the past about whether Frontiers is being readied for sale. The price would presumably be much lower now than a year ago. It’s worth remembering that (1) BC Partners will, at some point, perhaps even this year, need to exit from their investment in Springer Nature and that (2) Holtzbrinck Publishing Group holds a majority share in Springer Nature and a minority share in Frontiers (Stefan von Holtzbrinck sits on the Board of Frontiers). Could Frontiers and Springer Nature be merged in the future? This is pure speculation on my part; I have no inside information here and I may well be completely off the mark. It’s not beyond the realms of possibility, though. Market consolidation is a trend, after all. Second, fully open access publishers that have fast turnaround times are likely to see a drop in revenues very soon after experiencing a negative event. A publisher that primarily has institutional subscription revenue will likely be able to weather a storm better over a period of years than a fully OA publisher that can see its revenue halved within the space of a few months. That’s one reason why Transformative Agreements are so popular: they lock customers in and improve the resilience of a business. Third, reputation once lost is hard to win back. Frontiers has hired some talented publishing professionals in recent years. These job losses will linger long in the memory. It seems likely that some staff who have not been made redundant will be starting to consider options elsewhere; furthermore, Frontiers will find it harder to attract new talent in the future. Fourth, publishers of all types would do well to learn from the Hindawi and Frontiers 2023 case studies. There was a period when scholarly publishing felt like the Wild West with open access being akin to a gold rush. Fortunes can be made and lost very quickly under an APC-based open-access business model. Building financial resilience is even more important now than it was previously. Fifth, rapid growth isn’t necessarily a good thing and it’s hard to come back from a setback. The sudden drop in article output that Frontiers experienced is far from unique. PLOS went through something similar and it has never recovered (graph below is article output for PLOS the publisher; source = Dimensions). Scientific Reports published ~23,000 articles in 2023, which is less than its peak of ~25,000 in 2017 (see graph below; source = Dimensions). I took on management responsibility for Scientific Reports in January 2019 and turning the journal around was one of the most difficult things I’ve been involved in. The Frontiers’ management team and remaining staff have their work cut out for them over the next few years. Will they be able to return to the dizzy highs of January 2023? History suggests that’s unlikely to happen any time soon. A new era at Science Advances
Science Advances is approaching its 10-year anniversary with lots to celebrate. In 2023, the journal received more than 21,000 submissions and published almost 2100 papers. The editorial board, which started with just six distinguished academics, has grown to 45 deputy editors and almost 350 associate editors. After 10 years of outstanding leadership, Ali Shilatifard has stepped down as founding editor. Given the growth of the journal, we are announcing a new model for managing the journal to enable its growth and continued success in the future.
Because of the growth of the journal, Managing Editor Laura Remis and I have decided to appoint section editors to lead the five topical areas of the journal, rather than using a single editor model.
Science Advances (H Holden Thorp) JB: Since we’re at the start of a new year, it’s worth taking stock and looking at the article output of some of the selective open access journals that have been launched over the past decade or so and compare them with the (mostly subscription) journal that was dominant for so many years: PNAS. The graph below plots data from Digital Science’s Dimensions (all article types are included in the search). Here’s my, somewhat biased, interpretation of why Nature Communications has done so well compared with the other journals listed here, despite the fact that its APC is $6790 compared with $4950 for Science, $5800 for PNAS, and $2000 for eLife (which now uses the PRC model, remember).
Setting Demographic Goals for our Editors at AIP PublishingOne immediate area of focus is on editor recruitment. When we recently presented our benchmarking study to our editors-in-chief, there was broad agreement that there should be demographic objectives when journals evaluate new editors and board members. While the pace of change can vary — for example, there is no set term for associate editor contracts — contract renewals provide an ideal opportunity to diversify our talent pool. We have, however, learned that this can be a challenge: Although journals hiring for new editors-in-chief are encouraged to maintain a balanced gender ratio during interviews, there is still a huge gender gap at the requisite level in academia — meaning there is a lack of women available for these roles, which will take time to address. Sometimes, this can lead to practical questions of whether a journal is willing to delay an appointment to ensure equity in the process. AIP Publishing (announcement) JB: Creating a diverse editorial board can take time, as this announcement explains. However, it’s a vital part of journal development. Making public statements, such as this one, keeps the issue front and centre of everyone’s mind. Self-citations in around a dozen countries are unusually highFor the study, they evaluated the citation patterns of the 50 countries with at least 100,000 publications from 1996 to 2019 indexed in the Scopus research database.
They found that in 38 of the countries, the number of ‘country self-citations’ — in which researchers in a particular country cite their own papers — decreased over the 24-year period.
But in Colombia, Egypt, Indonesia, Iran, Italy, Malaysia, Pakistan, Romania, Russia, Saudi Arabia, Thailand and Ukraine, researchers increasingly cited their own papers over the same time period.
Nature Index (Dalmeet Singh Chawla) JB: The PLOS One paper that forms the basis of this story can be found here. In the authors’ words: We argue that these anomalies should be attributed to the aggressive science policies adopted by these countries in recent years, which are all characterized by direct or indirect incentives for citations. Our analysis confirms that when bibliometric indicators are integrated into systems of incentives, they are capable of affecting rapidly and visibly the citation behavior of entire countries. What counts as plagiarism? Harvard president’s resignation sparks debateWhat happened to Gay has prompted some scientists to question the value of requiring scholars to freshly summarize known facts in the introduction and methods sections of each new paper. In one approach, dubbed ‘modular writing’, researchers could sample more liberally from the work of their peers to describe the broader scientific literature, provided that they cite the source. This could particularly benefit those whose first language is not English, theoretical physicist and author Sabine Hossenfelder wrote on the social-media platform X after Gay resigned. “It is entirely unnecessary that we ask more or less everyone to summarize the state of the art of their research area in their own words, over and over again, if minor updates on someone else’s text would do,” Hossenfelder wrote. Nature (Jeff Tollefson) Editing companies are stealing unpublished research to train their AI (paywall)Many academic publishers collaborate with large, private editing firms to provide “author services”, which include English language editing. The arrival of AI has triggered a frantic race to the bottom among such firms, which immediately spotted a way to monetise two resources they had in abundance: research papers uploaded in digital formats and well-trained editors. Client papers could be used to train specialised AI large language models (LLMs) to recognise and correct the characteristic mistakes made by non-anglophone authors from all parts of the world. Editors could help the system learn by proofreading the automatically generated text and providing feedback for optimisation. Times Higher Education (Alan Blackwell and Zoe Swenson-Wright) Understanding key concepts in the UKRI open access policyJisc has created a series of downloadable resources to support publishers in understanding the steps they need to take to offer compliant publication options to UKRI funded authors. We wanted to synthesise the basic information needed to understand the different routes to compliance and outline high-level overviews of the process, with the aim of making it all seem less intimidating for the smaller/less well-resourced publishers. The main message we wanted to share was that there are multiple routes to policy compliance which publishers can adopt based on their circumstances. JISC (announcement) JB: You can read the resources here. I'm not sure if its “less intimidating” to see so many links to policies gathered in one place. Other news storiesCalifornia Universities Partner with Wiley on Landmark Open Access Agreement Journals retract six Didier Raoult papers for ethics violations REPORT: 2024 Publishing Tech Trends Peer review innovator wins APE Award for Innovation in Scholarly Communication Manuscript Manager joins STM Integrity Hub OpinionNot-for-profit scholarly publishing might not be cheaper - And that's OKEven in the most rose-tinted scenario, where several thousand articles per annum are cranked out by a skeleton staff, with minimal spend on marketing, platform development or article production, it is difficult to see ORE’s cost per publication falling below €1,000. Plug in a more realistic set of assumptions, like those used to produce Figure 1, and the cost per publication starts out at almost €3,500 in 2026 before falling to just over €2,000 in 2030 – only marginally lower than the average APC paid in 2022. Impact of Social Sciences (Rob Johnson) JB: This opinion piece summarises the key findings from a report Rob published at the end of last year entitled Scenario modelling for Open Research Europe, which I covered in issue 57 of this newsletter. The work is incredibly important because it gives an independent view of the costs of publishing that are a reasonable approximation to reality, unlike many previous attempts. Over the years there’s been a narrative that the costs of publishing scholarly articles online have decreased over time, which exacerbates claims about the profit margins of commercial publishers (which are a separate discussion). Towards a federated global community of Diamond Open Access – the diamond papersThe proposal outlined here is the result of an initial proposal first presented at the Global Diamond Summit in Toluca on 26 October 2023, and subsequently amended based on consultations to the GDSG, the organizers of the Diamond Action Plan, and the DIAMAS and Craft-OA communities. Further consultations, e.g. the Diamond Open Access Plan community, are on the way.
UNESCO is supporting this development and will provide a neutral platform that can follow up the principles and some of the framework actions in the UNESCO Recommendation on Open Science. Discussions are on the way and more information will be provided in due time.
Hypotheses (Pierre Mounier & Johan Rooryck) JB: This sounds wonderful in theory, but would be incredibly difficult to implement in practice. Rob Johnson argued on LinkedIn that we should keep an open mind (which I’m struggling to do): If (like most of my network) you read this post from somewhere in the UK or US, do not be too quick to dismiss ideas coming from outside the Anglosphere. As a vision for a more equitable and inclusive publishing infrastructure I see much to welcome in this paper's proposals. I hope it will garner wider interest and support among funders and policymakers. Rooting out scientific misconductSome publishers have become more willing to correct the scientific record. This led to more than 10,000 retractions in 2023—reflecting about 0.2% of the literature across all fields, as indicated in a recent analysis. According to the study, this is a 10-fold increase compared with two decades ago. Not all of these were because of misconduct, but studies have consistently found that two-thirds of retractions are. However, it is not clear whether the incidence of misconduct has increased over time. There is no question that the work of today’s sleuths, who often use software not available 20 years ago, has pushed these numbers higher. And the fraudulent activity of research paper mills that produce fake manuscripts is likely also a factor. On a larger scale, universities are starting to take a harder look at the suitability of perverse “publish or perish” incentives for faculty promotion and tenure. Science (Ivan Oransky and Barbara Redman) JB: Unsuprisingly, research integrity and papermills were one of the hot topics at APE 2024 this week. Signals and Clear-Skies are two tools that could help, which were well received by the audience at the STM’s New Dotcoms to Watch session. These tools, and others like it, can give users a sense of how a journal’s or a publisher’s article quality has changed over time. A choice, not an obligation: Releasing scientific software as open source should be the responsibility of the authorsRegardless, the timing for transitioning code to open source should remain under the control of scientists and their collaborators and should not be an absolute prerequisite for publication. The success and effectiveness of software development projects frequently depend on making the appropriate publishing decisions at the right moment rather than rigidly adhering to prescribed processes that might carry a flair of political correctness within scientific journals and publishers. After all, good citizenship does not automatically guarantee quality. The sustainability of open source is at the hands of the developers when they decide to communicate their work in a scientific paper upon acceptance, and beyond. EMBO Reports (I Kappas, VJ Promponas and CA Ouzounis) I4OA Hall of Fame - 2023 editionThe Initiative for Open Abstracts (I4OA) was launched in September 2020 to advocate and promote the unrestricted availability of the abstracts of the world’s scholarly publications, particularly journal articles and book chapters, in trusted repositories where they are open and machine-accessible. I4OA calls on all scholarly publishers to open the abstracts of their published works and, where possible, to submit them to Crossref. Among society publishers, the American Geophysical Union (AGU) went from 7% to 99% open abstracts for current journal articles last year, which is a great achievement. The publishing arm of the American Institute of Physics (AIP Publishing) joins them in reaching close to 100% open abstracts, going from 41% to 95% in 2023. CrossRef (Bianca Kramer) JB: Surely we have got to the point in the OA transition where abstracts, at the very least, should be free to read for both humans and machines. Many publishers keep abstracts behind the paywall, though, which is a missed opportunity in so many ways. Retraction Watch and Crossref: Collaborating to Improve the Assessment of Scholarly OutputsPublishers remain a key, authoritative source of information on retractions and other important updates to content. Both Crossref and the Center for Scientific Integrity see the Retraction Watch data as a complement to publisher-provided data on retractions. Better information on retractions from publishers supports all of the use cases listed in the previous section.
Publishers can check the Crossmark section of their Crossref Participation Report to see if they are providing information on retractions and other updates to Crossref. If not, there is support documentation that explains the additional metadata and information they can provide in order to do so, which they can share with their production teams or service providers.
Science Editor (Rachael Lammey and Ivan Oransky) Other opinion articlesPutting Data at the Heart of your Organizational Strategy Proper citation of research by journalists is necessary for more trustworthy news Trust in Scholarly Publishing Apply publication-charge waivers across hybrid journals, too (paywall) ALPSP Annual Conference 2023: A Review by George Litchfield Beyond Predatory Publishing: Additional Questionable Offers in Scholarly Publishing Hanging in the Balance: Generative AI Versus Scholarly Publishing Five ways generative AI will transform scholarly publishing Beyond Generative AI: The Indispensable Role of BERT in Scholarly Publishing Journal ClubIdentifying Fabricated Networks within Authorship-for-Sale EnterprisesFabricated papers do not just need text, images, and data, they also require a fabricated or partially fabricated network of authors. Most ‘authors’ on a fabricated paper have not been associated with the research, but rather are added through a transaction. This lack of deeper connection means that there is a low likelihood that co-authors on fabricated papers will ever appear together on the same paper more than once. This paper constructs a model that encodes some of the key characteristics of this activity in an `authorship-for-sale' network with the aim to create a robust method to detect this type of activity. A characteristic network fingerprint arises from this model that provides a robust statistical approach to the detection of paper-mill networks. arXiv (Simon J. Porter and Leslie D. McIntosh) JB: This preprint deserves to be read widely as the implications are important. As generative AI improves, fake text is likely to improve too. Assessing the author network is more likely to reveal the products of paper mills. Publishers will want to take a look at figures 5 and 6 to see how they fared compared to their competitors. Elsevier is the third highest in terms of proportion of content (and highest in absolute terms). Note the upwards inflection between 2021 and 2022 for Frontiers (light blue dashed line). Hindawi, unsurprisingly, fares worst of all. The paper’s discussion section begins with this paragraph: There is some debate within the research integrity community on whether it is prudent to openly share approaches to paper-mill detection on the basis that, in doing so, paper-mill enterprises are able to adjust their strategies accordingly. While this position is understandable, we note that we have benefited significantly from the open Problematic Paper Screener dataset on tortured phrases, and the open investigations on PubPeer that it has seeded. Without access to this data, validation of our proposed technique would have been difficult if not impossible to carry out. We believe that further transparency is important in the development of technologies that are strong enough to combat paper mills. And finally...EASE (European Association of Science Editors) recently announced a survey on Equity, Diversity and Inclusion, which you can read about here and take here. In their words: The study will be open from 8 January to 31 January 2024. Everyone across the community is encouraged to participate by completing a short survey (less than 10 minutes) in this timely and important initiative.
The survey is anonymous and contains questions about your workplace characteristics and perceptions of EDI. Your responses will help us establish action plans and measure progress towards greater diversity, equity, and inclusion. This information will be used to inform further work and the development of tools and guidelines to improve diversity across editorial, publishing, and peer review processes.
Until next time, James |
Journalology
James Butcher
The Journalology newsletter helps editors and publishing professionals keep up to date with scholarly publishing, and guides them on how to build influential scholarly journals.
Subscribe to newsletter Hello fellow journalologists, There’s a strong DEI theme to this week’s issue, with reports from Springer Nature (on editorial board diversity) and C4DISC (on workplace equity) released this week. The newsletter also includes a fascinating map of the biomedical publishing landscape, a primer on COUNTER, and a discussion of F1000’s recently revised editorial model. Thank you to our sponsor, Digital Science In late 2023, Digital Science fully acquired Writefull, which...
Subscribe to newsletter Hello fellow journalologists, This week’s issue of Journalology covers the STM Trends 2028 report, a new journals programme from Harvard University, and a preprint that suggests that preprints get a citation advantage. Thank you to our sponsor, Digital Science In late 2023, Digital Science fully acquired Writefull, which provides AI-powered language and metadata solutions to streamline every stage of the publishing workflow - including triaging, copy editing, quality...
Subscribe to newsletter Hello fellow journalologists, Two years ago I quit my corporate job and started work on Journalology. Many friends and colleagues thought I’d gone slightly mad. If you want to know more about the person who sends you these emails every week, Science Editor has just published the transcript of a conversation I had with Jonathan Schultz earlier this year. We discussed why I decided to leave Springer Nature, how editors and publishers differ in their approach to scholarly...