Reading Time: 5 minutes
How one university got three different results to the question “how many OA articles did we publish in 2022?” Graphic by Ben Rawlins of University of Kentucky. The “sources” are three different aggregation services, Dimensions, Web of Science, and Taylor and Francis.

Mark Carden deserves congratulations for organising one of the most distinctive conferences on the scholarly calendar. This is one a smallish conference, with around 160 attendees, but the attendees tend to return, because they know what to expect, and everyone I spoke to agreed it has a special atmosphere. What makes this conference different? Partly the eclectic line-up of the talks, covering a wide range of well-informed presentations (a clear library bias this year), and little overt or concealed sales pitch. Part of the success was a kind of equality to the event; I chatted to company CEOs, to professors, to an admittedly small number of early career researchers, but all of them open to talk. I think that openness derives largely from the informality, a light tone to the event – which derives from the chair’s accessibility and tone, for example, his witty comments about the attendee names (“Only one of you managed to spell your name wrong on the registration form”).

Of course, elements of the standard expected conference format continue. The opening keynote said nothing unexceptional, and nothing exceptional either (“these are turbulent times”, with a PowerPoint graphic of turbulence, but I found Antonia Seymour’s description of the scholarly publishing environment (“shared purpose”, “common goals”, and “trust”) contrasted greatly with some of the subsequent presentations, which revealed some shocking conflict and disparities in the scholarly landscape. In particular, a presentation by Ana Heredia on Latin American journal publishing showed that 80% of Latin American journals are not indexed in Web of Science or Scopus, and 90% of them have no impact factor. That doesn’t look to me like a shared purpose and common goals across all scholarly publishing; it looks more like a kind of systematic exclusion of much global scholarly content.  Kaveh Bazargan’s account of the many and varied ways in which articles are produced fraudulently makes it clear the phrase “community of trust” is clearly interpreted differently by many stakeholders: he identified no fewer than 4,700 phrases that are commonly used to disguise plagiarism to detection tools such as iThenticate. To make matters more complex, it’s not just goodies and baddies; Rashna Bhandari from India revealed that SciHub is used by researchers as a matter of course to find scholarly content, since they would not otherwise not be able to access. A shared purpose, perhaps, but not shared access to content.

Another great presentation was a quick overview of SCOAP3 (it’s actually SCOAP3, but I don’t want to encourage scientists, who should know better, to misuse scientific nomenclature). SCOAP3 is the CERN-based publishing system, which has now been running for almost ten years. Instead of separate deals between publishers and each institution for article publication, CERN negotiates centrally for most if not all of the papers authored by their researchers. The result looks like a clear win for researchers: the cost per article for publishing high-energy physics articles was well below the average APC cost in other sectors, and that cost had risen by only 7% in the ten years of SCOAP3’s existence, well below inflation, let alone by comparison with APC prices direct from publishers. What is more, 10% of the SCOAP3 budget (paid by the 45 wealthiest countries) is used to pay for countries that “cannot reasonably be expected to contribute”. Every SCOAP3-funded article, comprising about 90% of all the papers in high-energy physics, is published with full open access. The CERN approach may not be perfect, and it may not be appropriate in all respects for other subject domains, but it represents a fascinating example of what can be achieved by collective actions of researchers.

There was a rather inconclusive debate, on the motion that “academic libraries are no longer necessary”. The debaters ended by agreeing, with votes for and against the motion unchanged before and after the debate, because with or without the library building, the roles played by the library staff are clearly essential (and pretty much agreed by everyone in the room). What I learned, incidentally, was that today half the library staff at one major US university have no library training – which suggests to me how the library has imperceptibly changed its role during the last 25 years or so. Perhaps the institution should remain but should no longer be called a library.  

The workshops, a regular theme of R2R, were a mixed bag. I attended the AI theme, which and I expected more to be achieved from three one-hour sessions, but AI is a hot topic in the media, and everyone has an opinion, and much of the time was spent at quite a superficial level: there were many more questions than answers. I suspect that some of the other workshops, with a more restricted subject, were able to achieve more. The eventual summary of the AI workshop back to the full group was more coherent, I think, than the sessions deserved, including reasonable recommendations as the need for transparency and explainability, the need for AI literacy, and the value of empirical trials – but none of these really needed three hours of discussion.

For me, one of the most illuminating (if depressing) presentations were from Michael Levine-Clark from Denver and Ben Rawlins of the University of Kentucky, describing the valiant but labour-intensive attempt to gather some statistics about the benefit of transformative agreements. Drawing on Peter Drucker’s famous saying, ”you can’t manage what you can’t measure”, the presentation provided a clear demonstration of why the library is important. Yet at the same time, obtaining stats for one institution turned out to be way more difficult than expected. The library is confronted by a hodgepodge of incomplete metadata; even today, in 2024, we don’t have consistent metadata to identify the article type in many cases with certainty (is it a research article? A review article? An editorial?) nor even the corresponding author of an article, let alone knowing how much a university is paying publishers for APCs to publish articles. Would any other industry tolerate such vagueness? Ben Rawlins, of the University of Kentucky, described one publisher’s journal licensing tactic that was shocking for its clumsiness and lack of dialogue – again, hardly a community of trust.

Were there tangible outcomes from the conference?  My optimistic hope is that better systems, better checks, and better metadata pull scholarly publishing into a more effective position in coming years than it is today. But the conference suggested there are fundamental issues to be resolved.  

As for the event itself, I needed no convincing of the benefit of the face-to-face format, but even within face-to-face events, some conferences manage things better than others. Despite the dreaded call from the organiser to “go and network” in the breaks, the best conversations were indeed during those breaks, whether by intention or (in my case) largely through serendipity. Via some undefinable process, the organisation and feeling of the event enabled several fascinating discussions to take place. It’s a conference worth attending. I look forward to next year’s event.