The EA, IES, and UE respond to the REF2029 Consultation on Open Access

The English Association’s Higher Education Committee, the Institute of English Studies, and University English have responded to the REF2029 Consultation on Open Access.

With huge thanks to Professor Greg Walker (University of Edinburgh), Chair of Unit of Assessment sub-panel 27 in REF2021, for leading on this important collaborative work, and to all our colleagues in our subject communities with whom this response was developed.

Times Higher Education cited our response in their article, ‘UK learned societies blast REF open access proposals for books‘, which was published on 18 June 2024:

…Three bodies representing English literature – the English Association, the Institute of English Studies and University English – have also criticised the inclusion of scholarly editions in the open access plans, claiming the “cost of making these ‘gold standard’, multi-volume, collaborative outputs OA would be crushing for publishers and universities, making it unlikely they could be included in a REF return”.

“It is surely not the intention of this policy that lengthy, high-quality, transformative research should be outside the scope of REF,” they explain in their consultation submission.

Warning about the lack of sufficient funding to accompany the proposed plans, the bodies recommend that “longform open access publishing should thus only be encouraged in this REF cycle, and the essential funding and infrastructure put in place ready for the next”.

Key facts

• The Research Excellence Framework (REF) is the UK’s system for assessing the quality of research in higher education institutions (HEIs). It was first conducted in 2014, and replaced the previous Research Assessment Exercise. The most recent REF took place in 2021, and was was conducted jointly by the four higher education (HE) funding bodies: Research England (RE), the Scottish Funding Council (SFC), the Higher Education Funding Council for Wales (HEFCW), and the Department for the Economy, Northern Ireland (DfE). The REF was managed by the REF team, based at RE, on behalf of the four funding bodies and was overseen by the REF steering group, consisting of representatives of the four funding bodies. The primary purpose of REF 2021 was to produce assessment outcomes for each submission made by HEIs. These outcomes deliver the wider threefold purpose of the exercise: (i) To inform the selective allocation of the four HE funding bodies’ grants for research to the institutions which they fund, with effect from 2022-23; (ii) to provide accountability for public investment in research and produce evidence of the benefits of this investment; and (iii) to provide benchmarking information and establish reputational yardsticks, for use within the HE sector and for public information. See the REF website for more information.
• Open access refers the practice of making scholarly research available online without the reader having to pay to read, download, or re-use it.
• The REF Open Access consultation is the process by which members of the subject community can respond to the proposed REF2029 Open Access policy. The policy will outline open access requirements for research outputs to be submitted to the forthcoming REF exercise.

 

The EA, IES, and UE response

SECTION A

4. What are the most important changes in the open access landscape since the development of the REF 2021 open access policy? 1) How do these differ across disciplinary areas? 2) What are the implications of these changes for the REF 2029 open access policy?

The principle that journal articles should be published in institutional repositories is widely accepted in English Studies, with an evolving infrastructure to support it. ‘Read & Publish deals’ have also significantly increased the volume of articles available OA. However, the long-term viability of this compromise position is not yet secure: the uneven spread of OA globally means major territories remain outside TA deals & OA income is insufficient fully to ‘flip’ many hybrid journals with international authorships. Publishers & learned societies predict decreasing income & hence significant risks to the sustainability of AHSS research ecologies.

There has been little movement on the challenges to longform OA publication since the 2015 Crossick Report. Until a large-scale, universally accessible, & equitable funding model is established for longform OA, the burden will fall unevenly on HEIs at a time when resources are increasingly limited.

Variation in funding between HEIs will cause serious harm in disciplines such as English where excellence is distributed widely across institution types. There will be impacts on emerging & non-mainstream fields & sub-disciplines. We are very concerned for ECRs & researchers working outside HEIs who will find it harder than those in well-resourced institutions to comply with OA requirements. Longform OA publishing should thus only be encouraged in this REF cycle, and the essential funding and infrastructure put in place ready for the next.

 

SECTION B

5. Should deposit requirements post acceptance be maintained where publication isn’t immediately open access?

No.

If the requirement to deposit accepted versions (before editing) is maintained, the requirement for deposition within 1 month of publication (Annex A) is too short for most A&H disciplines. Journals are often published by small organisations (many located overseas) and by small teams of largely volunteer editors. Given the limited resources available, authors often do not hear that a volume has appeared until weeks, even months, after publication. And dates of publication themselves are often unclear, with the date on the journal cover often differing from the date of actual appearance in the public domain by weeks, or again months. All of which leads to the creation of significant grey areas around even so seemingly clear a concept as ‘date of publication’. Given this, it is realistic to allow at least 3 months between notional date of publication and the date by which an article needs to be deposited in an institutional repository.

 

6. Do you agree with alignment to the UKRI open access policy in respect of licensing for journal publications by requiring licensing terms equivalent to CC-BY or CC-BY-ND licensing for journal publications?

No.

The sector remains concerned that there are insufficient safeguards in place to prevent the widespread machine searching, harvesting, and unattributed re-use of research published in OA format by agents, individual and institutional, whose aims may be politically, ideologically, legally, or morally unacceptable to the researchers whose work is thus used. The implications of unrestricted OA publication for intellectual copyright, research integrity, and the safeguarding of researchers have yet to be seriously addressed with the communities concerned.

The making available of all research outputs to unrestricted search and re-use on by machine learning software used to train large language models such as ChatGPT also means that such work is necessarily exploited without attribution for commercial gain by global AI creators. This necessarily contravenes the terms of OA licenses requiring the attribution of all work so exploited. There are technical workarounds via which it is possible to publish content on the internet in such a way that this illegal appropriation is blocked. We ask that REF2029 mandates that such safeguards be a required feature of any OA stipulations, and the question of rights retention is addressed squarely throughout. Until safeguards are in place we propose authors should be able to choose their OA license, whether CC BY, CC BY NC, or CC BY NC ND. We strongly believe this is a decision for authors and publishers, not the funder.

 

7. Do you agree with recognition of alternative platforms as meeting open access requirements as primary platform for publication?

Yes

Pre-print distribution is popular in one part of the English Studies subject area (linguistics) and we are open to the inclusion of alternative platforms. However, we are concerned about the growing burden for research support teams and authors of the multiplication of permissible models. For clarity and consistency in the policy we agree that alignment with existing options (author-accepted or version of record) would be preferable.

 

8. Do you agree with the proposed changes to embargo periods for journal publications for main panels A and B (12 months reduced to six months) and main panels C and D (24 months reduced to 12 months), in light of changing standards and practice?

No.

There is, as yet, no sufficiently robust infrastructure in place to protect the financial viability – and thus the survival – of A&H journals published by learned societies, small charitable bodies, or other smaller scholarly-focused organisations. Such journals, and the societies and organisations that they help to fund through their subscriptions, are a vital part of the research ecosystems of many disciplines. They thus need to have exclusive right to publish original material for long enough to make purchasing the journal an attractive option. To ensure the survival of these journals and societies they support, we strongly recommend the retention of the 24-month embargo period for journal articles within the remit of Main Panel D. The distinctive and essential role that A&H societies individually and collectively play in advocating for and developing the fields they represent – and how this work is funded – especially in the current, challenging climate, are still not adequately recognised by the UK HE funding bodies.

 

9. Do you agree that changes to the open access policy for journal-based publications should be implemented from 1 January 2025?

No

There is an inherent awkwardness in a system that applies two different criteria to the same output type over the course of a single assessment exercise, one set for articles published before 1/1/2025, another for those published after that date. The scope for confusion and inadvertent contravention of the rules is clear. We recommend that the various deadline dates for publication of outputs in these proposals be standardised (to 1/1/2026). We strongly recommend eliminating the anomalies suggested above for this REF by implementing the new requirements consistently from the start of the next REF cycle (i.e. from 1.1.2029).

If this is not done, a significant number of high-quality research outputs may fall foul of these requirements for technical, operational, or personal reasons, thus preventing the REF exercise from fully capturing the range of research activity and publications produced in the assessment period. The result will be a reduction in REF2029’s capacity to capture, and report on the quality of UK HE research, and thus represent that quality accurately nationally and internationally.

 

10. Do you consider that tolerance limit for articles and conference proceedings should be retained at 5% of any submission?

No.

5% is too restrictive. We strongly advice that the tolerance threshold for articles and other shorter form pieces should not fall below 5%, and indeed ought to be higher.

 

11. Do you agree with the proposed exceptions for journal publications? Should any of the above be removed?

Yes, these exemptions are a reasonable starting point.

We do not suggest any removals.

The following kinds of article also need to be exempt from the OA requirement.

Articles that are likely to attract hostile responses, abuse, harassment (or worse) from organisations, regimes, and/or individuals, and where the increased global machine searchability made possible by OA publication is likely to expose individual researchers, teams, and their HEIs to risk and danger. Examples would include work on states, groups, and (sub)cultures that are intolerant of scrutiny or criticism, or which operate outside of legal norms, or, conversely, which are the object of hostility from hostile and/or malicious institutions or individuals – e.g. research on aspects of Chinese/Taiwanese history and culture, the Arab/Israeli conflict, terrorist organisations, etc., and, increasingly, work on gender and transgender issues, histories of colonialism, religious (in)tolerance, hate groups and hate speech, and on issues perceived by such groups, or by attention-seeking individuals, to be examples of ‘woke’ culture. Considerable care is needed in this area to ensure these proposals do not endanger researchers. The addition of a ‘sensitivity opt-out’ would be reassuring here.

Articles in (sub) disciplines where the most appropriate journals and platforms are published overseas, or in areas where OA is not yet well established.

Articles co-authored with researchers working outside the UK or not employed by UK HEIs, and so not contributing to the REF volume measure for the relevant UoA.

 

12. Do you agree that there should be no deposit requirement for longform publications, but that they should be made immediately available as open access upon publication (or no later than 24 months following publication if subject to an embargo)?

Yes, with the additional exemptions discussed in 19 below.

We welcome the waiving of the deposit requirement for longform publication, which respects the considerable difference between author-accepted and final publication versions of longform outputs, given the value added by publishers through editorial and production processes, and mitigates the risk of confusion between variants were both to be available.

 

13. Do you agree with the proposal of a maximum embargo period of 24 months for longform publications?

No

We are concerned that 24 months may be too short for scholarly editions, if they remain in scope for the policy. We believe there may be considerable differences in the underpinning business models for scholarly editions as opposed to monographs. Scholarly editions often have much greater publisher input (e.g. production costs of specialist typesetting) and ‘sell through’ over a considerably longer period of time than monographs. Currently there is no clarity on who would provide the additional funding for these ambitious projects. Their unexpected inclusion in the REF2029 OA rules conflicts with UKRI OA policy, which states that scholarly editions are explicitly out-of-scope, indicating they will be ineligible for UKRI OA funding. The cost of making these ‘gold standard’, multi-volume, collaborative outputs OA would be crushing for publishers and universities, making it unlikely they could be included in a REF return. It is surely not the intention of this policy that lengthy, high-quality, transformative research should be outside the scope of REF. We believe that scholarly editions should be exempt from the policy, though it is possible a compromise (e.g. 60 months+) may mitigate the impact of their inclusion.

 

SECTION C

14. Is licensing for third party materials not being granted a reasonable ground for exemption from open access requirements?

Yes

Third party materials are almost always fundamental and critical to the intelligibility, effectiveness, and reproducibility of a research output (see 15, below).

 

15. Is sharing of a version of an output without third-party materials if licensing can’t be obtained, mirroring the UKRI open access policy for longform outputs, appropriate to meet the open access requirements for REF 2029 policy?

No.

The inclusion of all the relevant third-party material is almost always fundamental to the intelligibility and effectiveness/reproducibility of a research output. The removal of such material is almost certain to render the output, at best, less intelligible and effective, at worst meaningless. Publishing outputs online in such a state is thus likely to impact adversely on the credibility of the research, the reputation of the researcher, and on the institution and culture that produced it.

 

16. Do you agree with the principle of a tolerance level for non-compliant longform outputs?

Yes

 

17. Do you agree with the proposed tolerance level of 10% for longform outputs?

No

Given the current state of the OA landscape regarding longer-form OA publication, it would be more appropriate to encourage the principle of OA publication without mandating it for this REF exercise, with a view to revisiting the issue in the next exercise, once the infrastructure is in place to make such a mandate feasible. Consequently, even a 10% tolerance is too restrictive if significant disruption and damage to the research and publication infrastructure are to be avoided.

 

18. Do you agree with the proposed date for implementation of an open access policy for longform outputs in REF 2029 being for all longform publications for which contracts are agreed from 1 January 2026?

No. See 1 and 16, above.

There is, again, an anomaly created by having one set of requirements apply to outputs published or contracted before 1/1/2026 and another to those published or contracted after that date. This introduces unnecessary complexity and potential for error into the system. Given that the number of outputs contracted and published after 1/1/2026 but available (after embargo periods have elapsed) before the 12/2028 (and which do not qualify for one or more of the exemptions) is likely to be quite small, we recommend that it would be preferable not to mandate OA publication for longer-form publications in REF2029, but rather to encourage it now, and apply the requirement more consistently in the next exercise.

 

19. Do you agree with the proposed exceptions for longform publications?

Yes.

We agree with the proposed exemptions and would not recommend removing any of them.

Are there other exceptions you think are necessary for longform outputs? Please provide evidence in support.

Exemption should be extended to longer-form publications in which the field of research, the topics discussed, or the conclusions reached, are likely to attract hostile responses, abuse, harassment (or worse) from hostile organisations, regimes, groups, and/or individuals, and where the vastly increased global searchability made available by OA publication is likely to expose individual researchers, teams, and their HEIs to risk and danger (see 11b).

It should also be extended to book chapters and other contributions to longer-form outputs where the host volume is published overseas, is edited by or co-edited with overseas researchers, with researchers not employed by UK HEIs and thus not contributing to the REF volume measure for the relevant UoA, or is otherwise exempt from REF OA requirements.

See comments on scholarly editions under 13, which we propose should be exempt.

A clear definition of what constitutes a creative output will be needed to ensure work in the burgeoning fields of critical-creative writing, hybrid and speculative writing, is included within this exemption. Similarly, while the term ‘trade book’ is widely used, there is currently no universally accepted definition that would capture all the various works, from best-sellers to critical-creative cross-over works and textbooks, that might be covered by this important category of output.