(No) Impact Statement

Issue 66 · June 2024

Join the conversation on Twitter @Briefer_Yet and LinkedIn.

Did a colleague forward the The Brief to you? Sign up via the button below to get your very own copy.

Journal Market Trends

We are putting the finishing touches on C&E’s 2024 Scholarly Journals Market Trends report. Yes, we know – most trends reports are not worth the time it takes to read them. That is why we wrote this one. There was no good resource available to provide to a journal’s editorial board, an organization’s board, or an executive team that provides an overview of the issues in the market that they need to be aware of. The report is the perfect pre-read for your next editorial board meeting, board retreat, executive team meeting, or other discussion about your journals or journals-related business. 

This report is not a sentiment analysis or underpowered survey of what other people think. This is C&E’s analysis, based on decades of work in this market, of what trends and market drivers your organization needs to understand today to inform decision-making.

We have worked to balance depth with “digestibility” in the report – in three ways:

  • We packed the report with colorful charts and tables to help translate complex data into insights you can use. 
  • The report is engaging. We bring the same narrative style and attention to language to this report that we bring to The Brief.
  • We include an Executive Summary in PowerPoint format with the report. While we think the full report makes a great pre-read, we also wanted to provide a presentation format that can be used in meetings to help frame discussions.

The report will be released shortly (in July). For those organizations that preorder the report before its release, we are offering an early-bird discount of 20% (discount code EARLYBIRD20). 

Please see our website for more information and to purchase the report.

Introducing the Capsule Messaging Collection™

The way marketers approach messaging is evolving, and C&E’s new Capsule Messaging Collection™ is designed to help organizations adapt to these changes. By providing a structured framework for messaging, the Capsule Messaging Collection™ supports both human-driven and AI-powered marketing workflows. 

To learn more about how the Capsule Messaging Collection™ works, please see our website.

(No) Impact Statement


If one were to stop 1,000 random individuals walking by a busy street corner in Washington, DC and ask them, If the federal government were to require a private good to be available for free with no delay, do you think it would impact the market for that good?, we would hazard a guess that 999 of them would respond affirmatively. 
The one person who responded negatively would likely be on their way to the Office of Science and Technology Policy (OSTP) of the White House, where apparently they think about economics very differently.
Readers of The Brief may recall that when OSTP issued the Nelson Memo in August of 2022, there was (as required by law) an accompanying economic impact statement. This statement, however, was essentially a landscape review and not an impact assessment. We noted at the time of its release that, “Curiously absent from the 2022 Impact Statement is any estimate or discussion of the costs associated with the implementation” of the policies contained in the Memo. 
Congress noted the same thing, calling this 2022 impact statement “not serious.”
In the 2024 Consolidated Appropriations Act, Congress specifically directed OSTP to “produce an in-depth financial analysis” of the Memo’s anticipated impact within 100 days of the law’s passing, or else “pause implementation of the memorandum.” 
That 100 days has recently passed and OSTP has produced a new document … that is also not an impact statement.
One of the most basic and essential questions that must be answered for any piece of legislation is, How much is this going to cost? In the case of the Nelson Memo those costs are both direct and indirect. 
As we have previously discussed in The Brief, requiring immediate public access to published research will result in an erosion of subscriptions for publishers. This will, in turn, result both in more journals flipping to Gold open access (OA) and in higher article processing charges (APCs) because APCs in hybrid journals are currently subsidized by subscriptions. More and more federally funded research will therefore be published via Gold OA at higher prices. 
This will have implications for researchers, who will have to pay for such APCs out of their grants – which means money will be diverted away from research to pay publication fees. While researchers can budget for publication fees in their grant proposals, without additional funds at the agency level, this will result in either less money for research or fewer grant recipients. In either case, there will be implications for US R&D policy and ultimately an economic impact on numerous critical industries (pharmaceuticals, materials, energy, etc.). 
The Nelson Memo will also have implications for universities, which will likely be responsible for compliance monitoring (see Item 3 below). Universities are increasingly signing transformative agreements (TAs), which help to mitigate some of the inequities introduced by the Nelson Memo and other OA funder mandates. However, TAs effectively shift the economic burden of these mandates from the funder to the university – who may then try to recoup some of the cost from grant overheads or higher tuition payments (which has implications for many students and their families) or other means. 
The policy will further affect research societies, more of whom are seeking relationships with the larger commercial publishers (few of which are US companies) due to the economies of scale needed to participate in TAs. The Nelson Memo will likely result in more TAs and thereby accelerate this consolidation, which can be clearly seen in the increasing market share (as measured by article output) of the top five publishers. This has implications for jobs at societies and at smaller publishers in the US, many of which are (via consolidation) being shifted outside the US. 
There will also be impacts on federal agencies, which will see increased monitoring and digital archiving costs (see Item 3, below). Additionally, there are the more indirect (but very real) costs associated with research fraud, which has been enabled by author-pays OA models. Sadly, the Nelson Memo will likely encourage more fraud – despite the updated report taking great pains to point out that US researchers are not the cause of most research malfeasance, at least as measured by article retractions – simply by shifting more journals to Gold OA.
These are only some of the implications of the Nelson Memo – there are no doubt many others that an impact assessment would surface.
None of these costs are discussed in the new OSTP report. While we at The Brief had low expectations for this analysis(100 days is a very short amount of time to produce the sort of “in-depth” analysis Congress required), we were genuinely surprised by what OSTP produced. It appears that OSTP has produced not an impact statement but rather a historical trends report for the scholarly publishing market, offering no financial projections of the Nelson Memo’s potential costs nor any acknowledgment of its impact on the market. 
This report reads like the start of a research project on the history of publishing business models and recent trends in research integrity and peer review – many of which, while notable, are irrelevant to assessing the costs associated with the Nelson Memo. The only financial data in the report show spending rates by federal researchers for periods when the policy was not in effect, with no attempt made to extrapolate these rates forward under the new policy. There is also no mention of the potentially enormous costs of the Memo’s open data requirements.
One can only surmise why OSTP has, for a second time, not produced the required impact statement. One possibility is that OSTP does not want to produce an impact assessment because they know the price tag associated with the policy is going to be high. Another possibility is that OSTP no longer sees the Nelson Memo as a priority and (under)resourced the development of the impact statement accordingly. This second possibility fits with our understanding that the statement was largely overseen by one part-time staffer on loan from NIH.
So what happens next? The ball is in Congress’s court. We are eager to hear whether this historical review is accepted by Congress as the “in-depth financial analysis” that was requested and whether implementation work on Nelson Memo policies is permitted to continue. A new draft appropriations bill was issued by the House of Representatives a few days after the release of the OSTP analysis, which contains identical language to that in last year’s appropriations bill issued in July of 2023:

SEC. 552. None of the funds made available by this or any other Act may be used to implement, administer, apply, enforce, or carry out the Office of Science and Technology Policy’s August 25, 2022, Memorandum to Executive Departments and Agencies entitled ‘‘Ensuring Free, Immediate, and Equitable Access to Federally Funded Research.’’ 
It is unclear whether we have come full circle and this represents a negative response to the OSTP’s unsatisfactory efforts, or if the language banning implementation was meant to serve as a placeholder as Congress awaited the analysis, but we will certainly be hearing more about the future of the Nelson Memo soon.



The National Institutes of Health (NIH), in responses to comments on their own draft Nelson Memo policy, echoes OSTP in not acknowledging the costs to researchers that will be incurred. Replying to comments suggesting that such costs exist, NIH states that “compliance with the Policy is free,” as “NIH maintains a free pathway for compliance … by submission of the Manuscript to PubMed Central.” While technically this is true (the act of compliance is free), NIH elides the costs (both financial and in time/effort) needed to get to the point where one can do this one “free” action. 
In a previous analysis, OSTP estimated that it will cost authors around $3,000–$4,000 per article to follow this “free” pathway. Hence the NIH policy is free like a puppy, rather than free like a beer. Accepting the puppy is indeed free, but the researcher is left to pay all the food and veterinary bills and to spend time and effort cleaning up the metaphorical droppings of this “free” gift from NIH.
NIH’s revisions are also notable in that they make clear there is no end date for the obligation to make articles freely available. After the grant has ended, the previously funded researcher remains on the hook for paying publication costs for any resulting work even though funds are no longer available. A failure to do so will reflect poorly on any future applications for funding. 
NIH appears to be attempting to head off paying charges for depositing articles in PubMed Central on an author’s behalf. “The NIH Draft Public Access Policy notes that if authors are asked to pay a fee by a third party for submission of Manuscripts to PubMed Central, the NIH Draft Public Access Policy would not permit this fee to be paid from NIH funds because it is not a legitimate publication expense.”
This is surprising given that the success of the current NIH policy is largely dependent upon third parties depositing papers on behalf of authors. Rather than allowing for a less expensive route that would improve compliance and reduce author burden, NIH has chosen instead to drive authors to the more expensive APC Gold OA route, where deposit by the publisher comes with an apparently acceptable OA payment. This part of the policy would not seem to exclude the American Chemical Society’s article development charge (ADC) as that does not include deposit on behalf of the author, rather just permission to deposit. Presumably traditional page charges would also be acceptable, so long as deposit was not included. The prohibition on paying for deposit services leads librarian Lisa Janicke Hinchliffe to wonder whether this also bans the use of federal funds for services to ensure deposit from the researchers’ own libraries.
The proposed effective date for NIH’s policy has been moved up as well, and compliance will be required for any manuscript accepted for publication on or after 1 October 2025. Perhaps most importantly, the language used makes clear that the new Nelson Memo requirements will apply to all existing grant recipients rather than just those funded after the policy is implemented (“This approach has the benefit of capturing all Manuscripts accepted for publication regardless of whether the award or contract is new or ongoing”). 
Aside from the confusion created by changing the terms of a contract midstream (NIH guidance does state that “At any point during your grant, NIH can institute a new policy that affects you”), this retroactive implementation directly disadvantages grant recipients who were not given the opportunity to budget for the necessary publication charges. Those researchers will instead need to pay the newly required expenses of APCs out of their existing funds, requiring the sacrifice of things such as purchasing reagents and equipment or paying salaries. It also rapidly advances the timetable for journals to get ready for these massive changes, as all papers will be subject to the new policy, rather than phasing it in over time through new grants. The comment period for these changes is open through August 19, 2024.

Administrative Burdens


One of the more confounding aspects of the Nelson Memo and resulting agency policies is the lack of awareness (or at least acknowledgment) by US funders and policymakers of the results of comparable EU policies already in place. Just as with OSTP’s “free” route to compliance, so too does Plan S offer a supposedly free option: the Rights Retention Strategy (RRS). Perhaps US funders should note that this method has largely been obviated by most publishers (by requiring APC payments and by steering papers with RRS statements to fully OA journals) and has seen little uptake from cOAlition S–funded researchers (down to 7% as of last year).
Also notable is the lack of recognition of institutional costs in monitoring policy compliance. These were raised by leadership at the University of Oxford in response to the UK’s proposed OA requirements for its 2029 Research Excellence Framework (REF). The Oxford administrators suggest that OA requirements would cost some £20 million in compliance and administrative costs for Oxford alone, which would translate into hundreds of millions of pounds across the sector. Oxford was joined in opposing the new mandates by the British Academy, the Royal Historical Society, and a collaborative response from the English Association, the Institute of English Studies, and University English.
The administrative costs (whether compliance is monitored by universities or funding agencies) generally seem to be off the radar for policymakers. Back in 2013, one of the “surprise” outcomes from the Research Councils UK (RCUK) OA policy was the enormous administrative costs involved. More than 10 years later, that lesson has not yet sunk in. Because the Nelson Memo calls for zero increase in funding, universities will need to pay any compliance monitoring costs out of existing overheads, and funding agencies will have to divert funds from other areas to track authors and presumably remediate noncompliance. 
Any policy that inadequately plans for likely outcomes, that fails to offer a realistic financial analysis, or that does not provide the required funding to achieve the desired outcomes (or all three!) is unlikely to be effective.
Perhaps then it should come as no surprise that cOAlition S has announced their support for OA in the 2029 REF, with suggestions to better align the activity with Plan S, a policy that the cOAlition suggests has not had the desired outcome.



Misperceptions about what AI and large language models (LLMs) in particular actually do are driving a lot of confusion in how we think about both their potential value and danger. A controversial (but fun!) article this month in Ethics and Information Technology offers a great reminder about just what it is that LLMs such as ChatGPT do. They are in no way designed to offer information, accuracy, or truth. LLMs are word-prediction machines, meant to produce text that resembles human writing, with no actual concern for the content that writing includes. In “ChatGPT Is Bullshit,” Hicks et al. argue that what LLMs are doing is akin to the concept of bullshit as defined by Harry Frankfurt’s acclaimed book, On Bullshit. At the core of the argument is that because, like a politician or a student who hasn’t read the assigned book, LLMs are not concerned with the truth, but rather “designed to produce text that looks truth-apt without any actual concern for truth,” their outputs are best labeled as “bullshit.”
The authors argue that the terms we use when LLMs give us false information – “lying” and most predominantly “hallucinating” – are misleading and spur further confusion. To lie one must know what is true and what is false, which an LLM cannot do. To hallucinate, one must perceive something that is inaccurate, which an LLM does not do (they don’t “perceive” anything but instead predict which word is most likely to come next in a sentence). These terms lead us to believe that LLMs are doing things that are beyond their function. This argument ties in well with recent suggestions that scholarly authors should not cite LLMs in research papers.
Misunderstandings of the true nature of LLMs are leading us down the wrong path, at least according to Steve Sinofsky. The market seems obsessed with AI-driven discovery tools, both generally (Google using Gemini to create direct answers to search queries – it did not go well) and specific to the world of scholarly communication, as nearly every company in our market has announced some sort of AI-driven enhancements to search results. Sinofsky suggests that the “idea that the current state of the art LLMs would simply replace search might go down as one of most premature or potentially misguided strategy ‘blunders’ in a long time.” As with the “Bullshit” authors, Sinofsky makes clear that using a tool “designed without a basis in facts” that is meant to give answers because “they sound good” is the wrong technology when asking a question you don’t know the answer to.
Ben Evans also writes this month about what needs to be done to develop useful products using LLMs, a tool we should assume gets things wrong. Evans, echoing Sinofsky, argues that the problems with LLMs 1) are inherent in LLMs (we should be skeptical of the AI evangelist claims that the technology will get better and these problems will go away), and 2) might be mitigated by thinking of generalized tools such as ChatGPT as infrastructure not products. 
Evans frames two alternative approaches to the generalized LLM:
The first is to contain the product to a narrow domain, so that you can create a custom UI around the input and output that communicates what it can and can’t do and what you can ask, and perhaps also focus the model itself (hence RAG [retrieval augmented generation]). This gets us the coding assistants and marketing tools that have exploded in the last 12 months, as well as the first attempts at knowledge management tools…
But the other approach is that the user never sees the prompt or the output, or knows that this is generative AI at all, and both the input and the output are abstracted away as functions inside some other thing. The model enables some capability, or it makes it quicker and easier to build that capability even if you could have done it before. This is how most of the last wave of machine learning was absorbed into software: there are new features, or features that work better or can be built faster and cheaper, but the user never knows they’re “AI” … Hence the old joke that AI is whatever doesn’t work yet, because once it works it’s just software.
In a follow-up article, Evans looks at Apple’s recent announcements of how they’re pursuing both of these approaches in integrating AI into the iPhone. Apple has essentially unbundled the LLM into a set of features and capabilities within other products (some of which may be visible to the user and some of which are so baked in they are simply “software”). At the same time, Apple also announced they are working with OpenAI to bring a ChatGPT app to the iPhone – essentially covering their bets (and potentially setting up a blockbuster licensing deal, similar to their arrangement with Google). 
These are useful lenses for organizations working in scientific and scholarly publishing. While licensing content to technology companies is attractive given the high margins (nice work if you can get it!), the more lasting play may be in product development. How does generative AI help enable products and services focused narrowly on solving problems for your customers and members? How does it lower costs by making existing products faster or cheaper to maintain? Organizations that can answer these questions creatively may find themselves better positioned for an AI future than those that only pursue licensing deals in the expectation that generalized tools will take over. 

Lower JIFs


Happy JIF (Journal Impact Factor) Day wishes go out to all our readers, as Clarivate has released the 2024 Journal Citation Reports, which confusingly (as always) show 2023 citations to papers released in 2022 and 2021. We’re consistently seeing declines in JIF scores as compared with those from last year. Much of this is likely driven by the post-pandemic return to typical publication volumes. Excluding meeting abstracts, the Web of Science lists 3,319,004 research articles published in 2023, which is a 4.4% decline from the 3,473,127 listed in 2022. Fewer articles mean fewer citations to count in the JIF numerator, while the JIF denominator remains counting articles from pandemic level highs. We recommend paying more attention to how your journal performed in relation to others in the field – its rank in category rather than its raw JIF.

Briefly Noted


In other Nelson Memo policy news, the National Endowment for the Humanities (NEH) released details on their plans to make a plan for a compliance policy. The plan to make a plan makes clear that the policy will only apply to journal articles, and that full books and book chapters are excluded. According to Dimensions (an inter-linked research information system provided by Digital Science, https://www.dimensions.ai), over the last three years, NEH funded an average of 99 journal articles per year. One has to wonder how well the costs of writing, maintaining, and enforcing this policy (not to mention the potential shift of hundreds of thousands of dollars from research budgets to publication fees) compare to the value of immediate (rather than delayed) access to less than 100 humanities papers per year.

America isn’t the only nation that seems to be in denial regarding the funding necessary for immediate OA to research papers. While funneling ¥10 billion (around $63 million) into university repositories, a spokesperson for Japan’s Ministry of Education, Culture, Sports, Science and Technology (along with the usual suspects such as cOAlition S’s Johan Rooryck) suggests that immediate Green OA is somehow feasible without payment of APCs. Unless Japanese authors are willing to restrict themselves to a limited number of journals, we expect to see significant use of the policy’s approved use of funds for publication expenses paying for APCs in 2025 when the requirements go into effect.

While the US proceeds headlong into embracing the author-pays OA model, Adam Marcus and Ivan Oransky call for its elimination in order to protect the integrity of the scientific record.

Bloomsbury acquires “the academic imprints and associated titles” of Rowman & Littlefield. The deal does not include Rowman & Littlefield’s other business lines. 

American Chemical Society and the Couperin consortium (which represents 250+ institutions across France) get credit for a new acronym: ZEGOA, which (despite sounding like a new weight loss drug) represents zero-embargo green OA. ACS and Couperin have signed a memorandum of understanding for a new agreement that provides reading access to ACS journals plus ZEGOA: author rights to “immediately self-archive the accepted manuscript in the repository of their choice for articles published within ACS’s hybrid journals.”

A new preprint concludes that MDPI and Frontiers journals rely on an inordinate amount of self-citation to build their reputations as compared with journals from other publishers. MDPI’s and Frontiers’ journals “exhibit much greater self-citation rates, particularly to articles contributing to the JIF.” This points out another fragility in the publication volume–based strategies of these companies. Given the decline in publication volume from both companies beginning in mid-2023 and continuing this year (currently down year-on-year 28% for Frontiers and 31% for MDPI), it will be worth monitoring how the lower number of articles, and hence a lower number of self-citations available, impacts their rankings. 

As noted in The Brief back in MarcheLife’s new publication model, which resulted in decreased submissions and an increase in the percentage of desk-rejected articles, is likely to have had a significant negative financial impact on the journal. In response, it appears that eLife is raising its prices for authors by some 25%

The Economist takes an in-depth look at the growth of science in China, in terms of both quantity and, increasingly, quality. This should perhaps be somewhat tempered by a report discussed in Nature suggesting that, at elite universities, researchers are pressured to engage in unethical behavior to protect their jobs. The growth is also seen as likely to impact Chinese author behaviors as the country looks to expand its own science publishing activities and bring more papers into China-owned journals.

The University of Cambridge has launched an initiative that seems aimed at creating a competitor to its own university press. Launching the Diamond Open Access Journals at Cambridge platform through the university’s Office of Scholarly Communication and Library would seem to deprive the effort of the ample expertise and infrastructure already available on campus at Cambridge University Press. We at The Brief have a hard time seeing a lot of authors flocking to these new journal launches that lack even the most basic of features such as ISSNs, DOIs, or any sort of long-term preservation plans.

The UK staff of the Nature portfolio are on strike against Springer Nature because they say their pay has not kept pace with the cost of living. The strike threatens to see Nature miss an issue for the first time in its 155-year history.
Wiley released their FY24 financial results. Overall, the publisher reported $53 million in operating profit though a $200 million loss after interest and various one-time “impairment” charges are factored in. The publisher’s “Research” group reported a 3% drop in revenue (to $1,042,705) attributable to the cleanup of Hindawi. “Excluding Hindawi, Research revenue for the year was flat.” Wiley breaks out the two business units that make up it Research division: “Research Publishing” (down 4% to $892 million) and “Research Solutions” (down 2% to $150 million). Wiley’s outlook for the current fiscal year for the Research division is “low to mid-single digit growth.”

“There is progress but it is slow. At the current pace of change, equality remains too far away and further action is needed to accelerate change.” So says Elsevier CEO Kumsal Bayazit in introducing the company’s wide-ranging report on Progress Toward Gender Equality in Research & Innovation – 2024 Review. There’s an enormous amount of data in the report worth digging into, but one point that caught our eye was that patents are largely (three-quarters) filed by men or by teams consisting entirely of men, yet publications from women are more likely to be cited in policy documents than those by men.

Recent industry moves include Daniel Kulp starting a new position as Director of Publications at the American Urological Association and Jasmin Lange becoming Managing Director of Bohn Stafleu van Loghum, part of Springer Nature Group. Congratulations to both!

After reports exposed that the research society Optica was distributing money to researchers from Huawei, a company blacklisted from doing business in the US, the organization has cut ties with the company and will return funds received from 2022 to 2024.

With 2,500 citations, a 2006 Nature paper on Alzheimer’s disease was set to become the most-cited paper ever to be retracted, but now will rank second, as Nature has retracted a 2002 paper on stem cell pluripotency that has been cited 4,482 times. The Alzheimer’s paper (Lesné et al.) includes several figures with manipulated images, and all authors, with the exception of Lesné, have agreed to a retraction. The paper is a key (though not the only) piece of evidence in a line of research investigating whether the buildup of amyloid proteins causes Alzheimer’s disease. Phill Jones wrote about the tribalism and factions in the world of Alzheimer’s research and how it has led to a loss of perspective that has damaged progress toward a cure. Though bad for medicine, it does make for good drama, at least as far as Grey’s Anatomy is concerned.

A new lawsuit filed by textbook publishers gets to the issues of Google’s ad integrity and its alleged unfair competitive stance. Cengage Learning, Macmillan Learning, McGraw Hill, and Elsevier have sued Google over its “circus of failures” in refusing to police the “pervasive advertising of unauthorized, infringing copies.” The plaintiffs allege that Google not only ignores their reports of illegal activity, it also promotes pirated copies of textbooks while restricting ads for legitimate ones.

More depressing analysis for writers. Your chance of making a living by writing books full-time is roughly equivalent to the odds of you becoming a self-made billionaire. 

The conduct of civilized life, and the vitality of the institutions that are indispensable to it, depend very fundamentally on respect for the distinction between the true and the false. Insofar as the authority of this distinction is undermined by the prevalence of bullshit and by the mindlessly frivolous attitude that accepts the proliferation of bullshit as innocuous, an indispensable human treasure is squandered. – Harry Frankfurt (2002) “Reply to G.A. Cohen.” Contours of Agency