The Copim community would like to offer some preliminary reflections on REF’s recently announced consultation on its updated open access (OA) policy for the REF 2029 exercise, in particular relating to longform outputs. We welcome any comment and feedback ahead of the mid-June deadline for responses to the consultation.

As a community dedicated to enabling the funding, creation, and circulation of OA books, Copim is deeply invested in these issues. We are keen to promote a range of viable, alternative models to the established, dominant mechanisms for commercial OA monograph publishing. Partly this is about ensuring the infrastructures and funding mechanisms are in place to support the expansion of Diamond OA publishing. Book publishers that use Diamond OA models do not require authors to obtain funding for OA publishing fees — Book Processing Charges (BPCs), as they are often termed.

We make the following arguments:

• We welcome the principle of OA by default that the draft REF 2029 policy proposes, arguing that it provides an opportunity for the UK higher education sector as a whole to begin to fully embrace more ethical, community-led forms of publishing in research and teaching.

• We suggest that action needs to be taken to encourage open research cultures within universities and more widely. We are keen that debate around the specific guidance for OA outputs does not distract from the need to make a positive case for expanding OA publishing across the sector.

We also offer initial views on details of the policy for longform outputs. We argue that the policy should:

• provide a clear roadmap for eliminating the use of embargoes before OA versions of a longform outputs are available;

• not rely on Green OA to deliver an expansion of OA for longform publications, given the risks of creating poor quality experiences for readers and harming perceptions of the importance of OA as an approach;

• more explicitly encourage and incentivise a transition to Diamond OA, as a more cost-effective and fairer route to the widespread adoption of high quality OA longform publishing than Gold OA;

• include more explicit guidance about how the OA version of a text is linked and shared;

• allow greater flexibility in which licenses are permitted.

In line with our emphasis on encouraging experimental forms of publishing, this text is enriched with additional information and context we have added as comments.

1. Welcoming the principle of OA by default

First, we are very pleased to see the draft policy strongly endorse the fundamental principle that OA should be a default option for longform publication in the UK sector. For too long, huge volumes of longform scholarly knowledge produced in the UK have remained closed off to the many potential readers that exist beyond the subset of academic institutions with the budgets needed to subscribe to and/or acquire paywalled/closed access content, with dominant publishing models — including dominant open access models — generating both local and global inequalities in scholarly publishing.

The draft policy also addresses an imbalance from the previous REF exercise. In REF 2021, only shorter form outputs — mostly journal articles, but also conference proceedings — were required to be made available on an OA basis. By including an OA requirement for longform outputs, REF 2029 will bring a measure of parity across disciplines. The previous set of requirements had the effect of keeping swathes of knowledge closed off from some disciplines more than others. A reader based outside the UK higher education system wanting to find the latest research from a UK-based scientist or a mathematician would be more likely to be successful than if they were seeking work from leading UK sociologists, historians, and philosophers, for example.

We still have concerns about the wider place of REF exercises in the UK higher education system. However, we realise that there is little prospect of such exercises losing their place in the UK system any time soon. As such, our intention is to focus on the opportunities that REF 2029 does present with respect to longform OA publishing.

2. Encouraging open research cultures

When REF announced the consultation on longform outputs, we were disappointed to observe how quickly some scholars (perhaps a minority, but a vocal one) bundled together their understandable fatigue with REF exercises with wider (often misconceived) objections to OA.

Our colleague Lucy Barnes recently addressed some of these views, which we summarise:

• Open access does not inevitably require the payment of publication-specific fees by either authors or institutions. Copim has written about and provided resources to publishers on the diverse range of revenue models that exist for OA book publishing beyond fee charging and is actively working on collective funding models to provide further alternatives. This includes the Open Book Collective, which is currently supporting the work of 12 OA publishers and infrastructure providers by distributing financial support around supporting universities. It also includes the Opening the Future revenue model, which is being used by Liverpool University Press and Central European University Press to radically increase the number of books they can publish on an OA basis. We have also seen a number of other models launch, seeking to achieve some similar aims, for example JSTOR’s Path to Open and MIT Press’ Direct to Open.

• Research has shown that making a digital book free does not prevent it from selling hard copies (as also explored in a 2013 report by Ferwerda, Snijder, and Adema).

• OA books are used more than closed-access titles, with potential benefits for visibility and citation.

There are further reasons why OA is often the better option for academics looking to either disseminate their research or integrate OA content into their teaching:

• OA books are globally accessible, helping address the entrenched inequalities in the global publishing system.

• OA texts are born-digital and therefore offer unparalleled opportunities to experiment creatively with the form the scholarly book, as members of our Experimental Publishing Group are currently highlighting through ongoing pilot projects.

Researchers and writers will never embrace OA solely because they are required to do so as part of a national exercise. Any mandates need to be combined with efforts to recognise and respond to entrenched structures in publishing and higher education.

For that reason, we are keen to see wider actions to foster a culture of open research and community-governed OA publishing across UK higher education. The separate consultation on the ‘People, Culture and Environment’ part of the REF 2029 exercise may be relevant to this, and we will contribute as we can. In particular, we would like to see institutions being encouraged to demonstrate their investment in their open research environments, for example via their allocations of QR research funding, and to encourage the use of open content in teaching. Universities could also be asked to include their open research strategies as part of their People, Culture and Environment submissions.

In preparing this post we canvassed university library colleagues on their personal views. For example, our colleagues Dominic Broadhurst and Wendy Taylor suggested that institutions “develop coherent strategies and practical means of proving the ‘academic value’ of OA monographs to academics’ own careers and research profiles”. This could mean developing detailed open research strategies that support the legitimacy and academic value of OA monographs to an academic’s own career and research profile. We would also encourage institutions to reorganise budget lines to facilitate Diamond OA as a default option, as a way for them to be able to demonstrate their investment in high quality peer-review, copyediting, and production of longform OA publications rather than (embargoed) access to intellectual content.

3. Initial views on the REF 2029 OA Policy for longform outputs

Here follows an indication of our current views on the key elements of the policy. We expect to expand on this in our final response, taking on board any feedback we receive.

1.       “If not published as immediately open-access, in-scope longform publications must be made available to freely read, download and search no longer than 24 months after the date of publication”

We welcome the broad principle of OA by default for longform publications. However, this aspect of the policy does not mandate full OA at the point of publication, which Copim strongly advocates for. In our view, both academics and the wider public would benefit from eliminating the option for an embargo before an OA version of a longform output is available, allowing research to be circulated, engaged with and used/reused more rapidly.  If a transitional period is required, we would urge that there is a clear roadmap for the phasing out of permitted embargoes, ideally commencing during the current assessment period.

2.       “Should be the version of record or the author’s accepted manuscript”

This aspect of the policy is significant, in that it opens the door to so-called ‘Green’ OA (commonly involving authors submitting their final draft, tending to be a Word document or a PDF to their institutional repository) being presented as equally acceptable to full open access (i.e. where the version of record is published openly) for the purposes of REF compliance.

We see issues with this approach. As a group which includes a number of publishers, we know how much work is involved in cleaning and formatting final manuscripts prior to publication and how this is a much greater issue for longform manuscripts than for journal articles, for example. While Green OA performs an important role in archiving and preservation, allowing Green OA as a widely acceptable version of OA for REF 2029 runs the risk of creating poor quality experiences for readers and may harm perceptions of the importance of OA as an approach. It would also in many cases relegate the OA version of a manuscript to becoming a second-tier variant, as well as posing significant practical challenges for university librarians working to support research staff. In addition, this approach will in many cases cause delays in manuscripts being publicly available, unless the publisher allows the Green OA version to be available immediately without embargo.

Instead of relying on Green OA to deliver a widespread uptake of OA for longform research and scholarship in REF 2029, we would advocate that the policy more clearly incentivise, and articulate the preference for, the distribution of an OA version of record at the point of publication wherever possible.

To support this, the policy could more strongly encourage support and funding for Diamond OA models that aim to achieve this. Diamond OA offers a cost-effective and equitable way of expanding the circulation of OA longform versions of record.

This is in marked contrast to the alternative potentially preferred by large commercial publishers: a significant expansion of Gold OA publishing, in which publishers charge BPCs for a book to be published OA, commonly in excess of £10,000. This would become huge financial strain on the sector, while also creating new inequalities within the UK scholarly system.

3.       “[Should be] available on a publisher website, repository or other appropriate platform”

In our view, the policy should encourage OA digital versions of record to be not only freely available, but also made as visible, and therefore usable, as possible. In many cases, it would make most sense for this version of the text to be on a publisher’s own website or on a third-party platform as well as a repository. While we agree that the policy should permit any of the given options, we would advocate for more explicit guidance that the OA version of the text should be linked, shared, and marketed on the publisher website (and wherever else relevant) to have parity with the closed (e.g. print) version of the text.

4.       “(Preferably) licensed CC-BY, but will accept CC-BY-ND, CC-BY-NC and CC-BY-NC-ND”

We would advocate for greater flexibility in permitted licenses. Many of our publishers use ShareAlike licenses (e.g. CC-BY-NC-SA), partly to support the principle of widespread knowledge sharing. In our view, such licenses should be included as permissible, alongside alternative licenses, such as CopyLeft or CopyFarLeft, or collective conditions for reuse, for example CC4r. These can readily translate into Creative Commons frameworks. Enabling a plurality of licensing options would better support bibliodiversity, experimentation and boundary pushing within the sector.

5.       “There will be a tolerance band of 10% at unit submission level”

We agree that submission units should be allowed a tolerance band. We are keen to collaborate with institutions in making a positive case for OA. Allowing some tolerance will help offset some of the logistical difficulties that any such policy will inevitably create, at a time when many academic colleagues are already working at or above capacity. We do not have a strong view on whether the proposed tolerance level is optimal or not. However, we would suggest that the policy consider mechanisms to ensure that scholars earlier in their careers and/or employed via precarious employment contracts (e.g. on a fixed term) are provided with greater scope for tolerance within unit submissions than their more securely employed colleagues. Similarly, consideration should be given to how to ensure equity across the sector, given a flat tolerance percentage may disadvantage smaller institutions submitting only a small number of longform publications.

6.       “Can exclude third party materials, if licensing can’t be obtained”

As a group including publishers with experience of attempting to negotiate licenses with third parties, we understand the challenges of doing so in some cases. However, we would encourage the policy to consider mechanisms to ensure that such provisions are used only in circumstances where it can be clearly demonstrated that alternative options that do not require third party licensing have been appropriately considered.

7.       Exceptions

The policy includes a number of potential exceptions. We will not dwell on these here. We would simply point out that the policy in its current form is forgiving in its approach. In our view, while exceptions may indeed be required in some instances, the policy could more strongly support OA by default and mechanisms to ensure a shift in cultures of scholarly publishing. Exceptions should be prioritised for certain categories of researcher, including those who are precariously employed or early career, given they often lack the flexibility of publication options that their more securely employed colleagues have.

4. Concluding reflections

The work required to change scholarly publishing in the UK and more widely should not be underestimated. We hope that whatever form REF 2029’s final OA policy takes, it becomes a prompt for institutions to creatively explore local approaches to delivering a fairer, more sustainable future for scholarly publishing alongside their pursuit of research and teaching excellence. And we hope they do this not just because they are compelled to but because they recognise that this is an urgent priority for the sector.

Header image by Eléonore Bommarton Unsplash