Executive Summary: The proposed HEFCE/REF Open Access [OA] mandate -- that in order to be eligible for REF, the peer-reviewed final draft of all journal articles must be deposited in the author’s institutional repository immediately upon publication, with embargoes applicable only to the date at which the article must be made OA is excellent, and provides exactly the sort of complement required by the RCUK OA mandate. It ensures that authors deposit immediately and institutionally and it recruits their institutions to monitor and ensure compliance.
For journal articles, no individual or disciplinary exceptions or exemptions to the immediate-deposit are needed, but embargo length can be adapted to the discipline or even to exceptional individual cases.
Embargo length is even more important for open data, and should be carefully and flexibly adapted to the needs not only of disciplines and individuals, but of each individual research project.
Requiring monograph OA if the author does not wish to provide it is not reasonable, but perhaps many or most monograph authors would not mind depositing their texts as Closed Access.
we propose to accept material published via either gold or green routes as eligible, recognising that it is not appropriate to express any preference in the context of research assessment
Excellent. This is the optimal OA policy and is completely compatible with the OA policies being adopted worldwide. Note, though, that gold and green are not both “publishing routes.” They are both routes to providing OA, but only gold is a publishing route. The green route is to publish in any journal at all, and to provide OA to that publication by depositing it in an OA repository. So for clearer wording I would suggest: we propose to accept as eligible published material that has been made Open Access via either gold or green routes, recognising that it is not appropriate to express any preference in the context of research assessment. It is already implicit in the proposed HEFCE/REF OA policy, but suggest that it is make explicit that (the final peer-reviewed draft of) articles must be deposited in the author’s institutional repository immediately upon publication, irrespective of whether the author chooses the gold or the green route.
We propose to treat as ‘open access’ publications those which meet all of the following criteria:
deposited in the author's own institutional immediately upon publication,
although the repository may provide access in a way that respects agreed embargos
made available as the final peer reviewed text, though not necessarily identical to the publisher’s edited and formatted version;
and presented in a form allowing the reader to search for and re-use content (including by download and for text-mining) both manually and using automated tools, provided such re-use is subject to proper attribution under appropriate licensing
Excellent. I suggest that it be made clear that the re-use might only be after any allowable publisher embargo has elapsed. (I also suggest specifying that the re-use rights may exclude re-publication rights by rival free-riding publishers, otherwise this condition may induce publishers that have no embargo to adopt an embargo.)
We intend that work which has been originally published in an ineligible form then retrospectively made available in time for the post-2014 REF submission date should not be eligible, as the primary objective of this proposal is to stimulate immediate open-access publication.
Excellent. This is precisely the condition that is needed to ensure that deposit is immediate and not delayed, and to ensure that the authors’ institutions are recruited to monitor and ensure that deposit is immediate and not delayed. It is especially useful because its effects go far beyond the 4 papers that authors will ultimately submit to REF: The choice of the 4 papers to submit is not usually made until the end of the REF interval, just before the next REF. So, in the meanwhile, this policy makes it necessary that all potentially eligible papers are deposited immediately upon publication, whether or not they are ultimately submitted. This will go a long way toward ensuring that all UK research output is deposited immediately. (Special congratulations to HEFCE/REF for this especially effective clause!)
The role of institutional repositoriesAs part of our commitment to increasing public access, we intend to require that outputs meeting the REF open access requirement (whether published by the gold or green route) shall be accessible through an institutional repository.
Excellent. But one small suggestion: …(whether made open access by the gold or green route)… As earlier, this is to distinguish publishing from access-provision.
‘All submitted outputs covered by our requirement for open access above, and other submitted outputs that are available electronically, shall be available through a repository of the submitting institution.’
This would mean in practice that each submitting institution would maintain a web facility through which all relevant outputs might be identified and accessed (including items available through a link to another website).
Excellent. Exactly the policy needed. And of course just about all UK institutions already have these repositories. What they lacked was a mandate that would fill them, and empower them to ensure that they are filled. This is precisely what the HEFCE/REF OA policy provides – for the UK, and as a model for the rest of the world.
We welcome further advice on repository use and on techniques for institutional repositories to cross-refer to subject and other repositories.
Institutional repositories need to be OAI-compliant and interoperable, so that subject and other repositories can harvest their metadata for central cross-repository search.
Institutional repositories should also implement the SWORD protocol for importing/exporting contents to/from institutional and central repositories, such as Arxiv, UKPMC or EuroPMC. Repositories should also configure their data to make them maximally visible, discoverable, harvestable and hence searchable through Google Scholar and other major search engines. (I don’t think the function you mean here is to “cross-refer”: it’s interoperability, harvestability, importability, exportability, and rich metadata. As the repositories begin filling in a serious way, these functionalities will be developed and made even more powerful than is currently envisioned.)
Embargoes and licences
Some publishers introduce embargo periods before work can be made available in an open-access form. Where embargoes apply we propose to determine eligible periods with regard to the practice of other major research funders at the time. Outputs will be eligible if they are still under an acceptable embargo at the REF submission date
Exactly the right strategy: Separate deposit mandate (immediate, no exceptions) from the date at which the deposit must be made OA (allowable embargo length to be decided discipline by discipline).
While we expect that sufficient clarity and reassurance on embargoes and licences will be achieved through the Research Council discussions, we welcome responses which address these issues.
The crucial thing is to separate the immediate-deposit requirement from the question of embargo length or nature of license.
Exceptions
Ideally we would like to see all outputs that are submitted for research assessment published in a form which meet the criteria. However, we recognise that there may be some exceptions during this transitional period.
We have considered three possible ways of handling exceptions:
a. Identifying categories of material which we think may be exempt from the open access requirement described above (paragraph 11). We would give careful consideration to reasoned suggestions for sub-categories of material that should be exempt.
b. Allowing individual outputs to be exempt from the requirement on a case-by-case basis, based on guidelines.
c. Specifying that a given percentage (for example, 80 per cent) of all outputs submitted by an institution meet the requirement.
Books and data can be handled differently from journal articles, because the case for (and timing of) OA for books and for data is very different. But on no account should any exceptions be allowed for the immediate-deposit requirement for journal articles. Special treatment or exceptions should only pertain to the embargo length (i.e., the date at which the deposit is made OA). And on no account should the immediate-deposit requirement be applied only on a percentage basis. Immediate-deposit should be 100%. Embargo-lengths and rights/licenses can be adapted to disciplines or special individual cases.
Some have asked that particular disciplinary groups should be exempt from this requirement, but we consider that research in all subjects has equal importance and therefore equally merits receiving the benefits of open-access publication. As with other aspects of the REF we expect the details relating to exemptions to be sympathetic to particular disciplinary issues; but in this instance we consider it will be most appropriate to identify which types of output should be exempt, looking across all disciplines, and we welcome advice on this.
Again, exemptions and exceptions based on disciplines should only be considered for OA embargo length and for further rights licensing, over and above free online access (and of course for books, data, and other special content other than journal articles). But not for journal article deposit date, which must be immediate.
Taking account of publication timescales and that the start of the next REF period is 1st January 2014, it may be that some notice is needed before these requirements apply. We propose to set a date which provides reasonable notice. Outputs published before that date will be automatically exempt from these requirements. We welcome advice on an appropriate notice period, taking account of the publications cycle.
If the HEFCE/REF mandate is adopted soon, there is no reason at all why 1 January 2014 cannot be the start-date for the immediate-deposit requirement (for journal articles).
(If a researcher’s institution does not yet have a repository, there is OpenDepot, created precisely for that purpose – and still patiently awaiting mandates in order to put its resources to use!)
Monograph publications
No pressing issues with books: No harm would be done if monographs too had to be deposited immediately, but with no requirement to make them open access – neither immediately, nor ever, unless the author wishes.
(As green OA for articles grows, more and more monograph authors will want the benefits of OA too, in terms of increased usage and impact.)
Open data
We invite comment on whether respondents feel this is the appropriate approach or whether they feel that sufficient progress has in fact been made to implement a requirement for open data as well. We will consider any representations that such a requirement may reasonably now be developed but would also need advice on how this might be achieved.
There are special complications with data that do not exist at all for journal articles (or even for books): Researchers are researchers, not mere data-gatherers. They gather data in order to use, data-mine and analyze it. If they are forced to make their data OA for use by one and all immediately, then there is a “Prisoner’s Dilemma”: It’s much better for me if I don’t take the time or trouble (nor spend the time seeking the funding) to gather the data myself: Just let someone else do the work, and then I can help myself to the data immediately, because it is mandated! In other words, embargoes are a much more serious matter in the case of data than in the case of journal articles.
Journal article embargoes are merely ways of allowing publishers to ensure their current revenue streams and modus operandi instead of letting research and researchers derive the full benefits of the web era. (It is not clear that that is good for anyone but the publishers). But with research data not only is the existence and length of an exclusive “1st-expoitation right” for the data-gatherer fair and important, but the length of the fair “embargo period” will vary substantially from project to project, not just from discipline to discipline.