Pathways to Open Access
  • A A A

    REF exceptions

    By Alan Bracey, on 21 December 2015

    As part of our work on exceptions to the REF policy, we’ve done an analysis of UCL’s 2014 REF submission in relation to the following exceptions:

    • 39 (b): embargo periods that are longer than the permitted REF maximums
    • 39 (c): journals that do not allow deposit

    We were keen to get an idea of how many of these exceptions we would have had to report to HEFCE had the policy applied in 2014. We also wanted to see how far our findings agreed with HEFCE’s estimate that 96% of 2014 REF submissions covered by the policy had embargo periods that could have complied (see para 54 of the consultation outcomes document here). As well as answering these questions, we’ve produced some statistics on non-compliant journals: We’ve also put together an editable list of all the non-compliant journals we found, by panel, which we hope will be a useful community resource:

    What we did

    UCL submitted a total of 9,413 publications to REF 2014. Of these, 8,048 (85%) would have needed to comply with the new OA policy, i.e. they were papers published in journals or proceedings with an ISSN.

    We boiled the 8,048 eligible publications down to a list of 2,510 ISSNs, and checked each title’s Green OA policy. The work consisted of manual checks of publisher webpages, Sherpa Romeo record checks, and using some embargo period data supplied by the Sherpa REF project team.

    We recorded journal embargo lengths, where available, and if deposit wasn’t allowed. We also recorded where a Green OA policy was unclear, or missing. Then we checked our findings against the maximum embargo lengths permitted by HEFCE, and created a list of non-compliant publications. From this, we extracted lists of non-compliant journal/proceedings titles, one for each panel. We’re making the lists available here: They’re open to anyone to edit, correct and supplement.

    A selection of our findings

    531 of the 8,048 publications would have fallen under either the ’embargo length’ (39b) or ‘deposit not allowed’ (39c) exception. This includes publications in titles without a clear policy, which constituted the majority of the exceptions (281).

    93.4% of publications submitted to REF2014 that would have fallen under the policy would have complied (leaving aside other exceptions) if they’d been deposited in time. This is close to HEFCE’s estimate of 96%.

    The majority of our exceptions (54%) were in our panel A submission.

    Elsevier was UCL’s top non-compliant publisher. 15% of the exceptions were in their journals. They lead on this score by a wide margin, with Wiley, the 2nd-highest publisher of non-compliant papers, publishing 9% of the exceptions. We suspect that society journals will feature in the lists of these publishers’ journals – but we haven’t checked, yet.

    Of all UCL’s non-compliant papers, more were published in The Journal of Immunology than any other title. Its publisher, The American Association of Immunologists (AAI), was the 4th-highest non-compliant publisher in our list.

    For more stats, see:

    What next?

    We’re running a workshop to look at all the REF exceptions in January. We’re asking delegates to contribute their experiences of working with the exceptions so far. However, we’d welcome feedback from everyone, and will include comments from the whole community in the exceptions toolkit that we’ll be producing. Please add your comments to this document: