The release in February 2022 of the EUA (European University Association) Open Science Agenda 2025 provides a timely snapshot of the current state of play and challenges for Open Research across European academic institutions. The report has several important messages that are equally relevant to Australian academics and academic institutions as to European and international. The Agenda and its key messages are the subject of this week’s post.
Pausing for a moment on terminology, even though the UNESCO definition of “Open Science”, used the Agenda, explicitly includes “basic and applied sciences, natural and social sciences and the humanities”, the term has acquired slightly misleading connotations of STEM-only science. Acknowledging this, terms such as “Open Research” or “Open Scholarship” are perhaps preferable today [1]. Nevertheless, though, the report is not noticeably STEM focused, and may be read to apply equally to researchers from across the spectrum of disciplines.
The report itself is structured around three key priorities: the role of open research in scholarly research publishing; in scholarly research data; and the subtle connections to assessment of scholarly research.
Open Access publication
Open Access publication is at the root of many discussions of Open Research, especially in the wake of the COVID-19 pandemic. The objective seems clear, and something most people would intuitively support: 100% Open Access scholarly research publications.
The pathway for getting there, however, is much less clear. The report cites a depressing finding from the OA2020 initiative that 85% of new scholarly publications in 2017 were still appear behind paywalls. A more recent EU study paints a slightly more positive picture, with 45% of EU27 publications Open Access in 2019, and some EU countries passing 50% Open Access. The Curtin Open Knowledge Initiative Open Access Dashboard provides perhaps the richest and most up-to-date data source on Open Access publication, with Australia reaching 40% Open Access in 2021. However, only about 25% of those publications are Open Access with the publishers, the other 15% made up of harder-to-find Open Access via a preprint server, a university repository, or some other source, such as an academic’s own web page.
Perhaps one initiative that is helping to change that balance is read-and-publish agreements, which aim to shift university library expenditure from paying for publication subscriptions to paying for institutional Open Access publication fees. Led by CAUL (the Council of Australian University Librarians) with strong support from Australian Chief Scientist Cathy Foley, Australia has seen a recent uptick in such read-and-publish agreements. However, the list of participating publishers in CAUL read-and-publish agreements remains depressingly short, with many caveats and limitations even for those few agreements.
While a level playing field between academic institutions and commercial publishers still seems some way off, the EUA Agenda does highlight encouraging trends towards a growing culture of open access amongst academics. European academics’ engagement with Open Access is known to be high and increasing, and anecdotally at least Australian academics seem not far behind.
FAIR data
The second EUA priority area is findable, accessible, interoperable and reusable (FAIR) data. FAIR principles have become a key priority internationally for research data over recent years. Making research outputs easier to access helps safeguard the visibility and value of research into the future. FAIR, Open Access and Responsible Data Management (RDM) principles together form a knot of distinct but interrelated guidelines that are fundamental to research transparency, research integrity, and research quality.
Advances in the quantity of FAIR research data, though, remains stubbornly slow in practice. Once again, the terminology is not always helpful to the cause. In particular, the focus on data, at the expense of reproducibility more broadly—in research methods, procedures, data analysis, algorithms, software, and workflows—is arguably a factor that narrows the focus as well as stifling uptake and culture change. The EUA Agenda highlights four further key challenges facing FAIR-ness in research data, namely:
Gaps in institutional policies for FAIR data and RDM;
Lack of skills, training, and career paths for professional data custodianship;
Lack of infrastructure for supporting FAIR data; and
Gaps in the incentives and rewards for researchers pursing FAIR principles.
Gaps in all four of these areas are evident in most if not all Australian research institutions, and together are no doubt major contributing factors to FAIR today being an aspiration rather than a reality in Australia. In 2017, a joint policy statement in support of FAIR by some of Australia’s key research leadership organisations (including CAUL, ARC, NHMRC, CSIRO, and Universities Australia) was a statement of intent, but ultimately fell well short of its goal of seeing all Australian publicly funded research outputs become FAIR by 2020. Indeed, the ARC’s (Australian Research Council) decision also in 2020 to remove the former requirement to document the “Management of Data” in funding applications—replacing it instead with boilerplate funding agreement legalese—was arguably a step backwards in the profile and visibility of FAIR data for Australian researchers.
Research assessment
Perhaps the most subtle and interesting priority area highlighted by the EUA Agenda is the role of Open Research in our approach to research assessment. As every academic today will be keenly aware, research assessment is dominated by citation metrics and journal rankings and impact factors. These metrics are often used for promotion, for recruitment, by funding assessors as blunt proxies for research performance, research impact, and even research quality. EUA estimate that 75% of European universities rely on journal-level impact factors to evaluate individual academic performance, a particularly corrosive distortion prevalent enough in the academic community to have become a key focus of DORA, the Declaration on Research Assessment.
In fact, most academics would agree such citation metrics provide scant if any reliable information about research performance or quality. Anyone who has served on academic recruitment panels or as funding proposal reviewers over the past decade will hopefully have become adept at spotting citation rings, where semi-organised groups of researchers collude in reciprocal citation cartels, often with surprisingly conspicuous success. While most researchers would agree that more nuanced and qualitative research assessment methods are needed, the EUA Agenda highlights the bind that many universities find themselves in. While being ostensibly autonomous in how they approach research assessment, in fact universities face shared external pressures, such as competition for research funding, research rankings, and research talent, that constrain unilateral action and impose a first-mover disadvantage.
Ownership and bibliodiversity
Finally, one of the underlying themes woven throughout all three priority areas is the importance of growing academic ownership of a diversified portfolio of Open Research pathways and outputs. The danger associated with the current period of transition is of historic publisher-dominated, copyright-driven closed research models being replaced simply with new straight jackets to Open Research. Publishing behemoth Elsevier, for example, posts annual profits exceeding $1Bn AUD, with margins of over 30%—beyond that of many luxury goods companies. Through the CAUL read-and-publish agreements, Elsevier offer Australian academics a derisory 12.5% discount on open access publication charges. This leaves thousands of dollars still to pay from research funding for each and every Elsevier article.
Bibliodiversity, in contrast, means supporting a range of different pathways to Open Research. The past two decades have seen a raft of community Open Access journals, for example, that have expanded the envelope of what is possible with Open Research. Journals in my field, such as the Journal of Artificial Intelligence Research (JAIR), the Journal of Spatial Information Science (JOSIS), and most notably the Journal of Statistical Software—frequently ranked in the top ten journals worldwide in applied mathematics and statistics—are run by academic communities for both community and wider public good. These journals operate on a shoestring budget in stark contrast to commercial publishing houses, estimated to be extracting more that $10Bn USD annually from research funders and institutions for publishing primarily closed-access research. Whether for community or commercial journals, academic researchers write the papers, review the papers, frequently edit the journals, and usually make up the biggest readership. So it is not hard to see why reclaiming ownership of academic Open Access publishing, while attractive, faces stiff opposition from vested commercial interests.
Changing publication models and pathways are of course not without risks. The EUA Agenda rightly highlights the future importance of research blogs, preprints, and other Open Research outputs throughout the research lifecycle. Understanding and managing the quality of these more diverse, less centralised forms of research output will clearly present new challenges. However, I would argue that the nexus of Open Research, FAIR, and RDM at the root of the EUA vision has research quality and research integrity more deeply embedded and incentive-compatible than was ever the case with our past tradition of commercial publishers as the gatekeepers of research outputs.
References
[1] Tennant, J., Beamer, J. E., Bosman, J., Brembs, B., Chung, N., Clement, G., … Turner, A. (2019). Foundations for Open Scholarship Strategy Development. MetaArXiv Preprints doi:10.31222/osf.io/b4v8p
Share this post