PLOS Submission to RCUK Review Panel Submitted by Cameron Neylon and Catriona MacCallum, 3 September 214 A. Declaration of Interest 1. PLOS (Public Library of Science) is a not for profit publisher and advocacy organization founded to accelerate progress in science and medicine by leading a transformation in research communication. PLOS generates revenue primarily from publication fees (APCs) 1. B. Introduction 2. All PLOS articles are compliant with the RCUK Policy. On publication, they are made immediately available for reuse under a Creative Commons Attribution licence (CC BY) 2 and deposited in PubMed Central. We are aware that the panel would like submissions to be evidenced-based and have therefore focused on specific questions in their call for submissions from the perspective of a pure Open Access publisher. Specifically we focus on managing workflows related to APC payments to publishers and obtaining the data required to monitor compliance. 3. In addition, we provide an appendix with data (primarily in relation to the current reporting period Apr 213 Jul 214) on papers published at PLOS where the corresponding author has listed a UK Higher Education Institution (HEI) as their affiliation. 4. We point out that for most publishers workflows for APC management and compliance monitoring are new and the data required for both to work efficiently overlap. Implementing appropriate systems will take time. This is not inherently difficult but will require a shared responsibility among all stakeholders about the timescale of the transition and adjustments to expectations on which stakeholders are the authoritative source for specific elements of information. 5. The traditional relationship of publisher as suppliers of published materials and libraries as purchasers of subscriptions has been separate from the supply side relationship between authors and publishers. Publishers have traditionally been able to provide comprehensive metadata on published outputs. APC business models change institutional expectations of what metadata will be provided. 6. The role of administering RCUK funding has been placed with institutions. Institutions and authors are the authoritative source for key pieces of information such as the affiliation of authors and their eligibility for using RCUK funds under institutional rules. Collecting all relevant metadata requires a reassessment of the roles and responsibilities of all players in providing and aggregating authoritative information on the connection between authors, grants, institutions, specific research outputs and publisher metadata. 7. Many of the solutions for monitoring compliance and managing APC workflows, such as providing standardised structured metadata, are already being developed by different stakeholders (see section F below). RCUK has an opportunity to ensure that the developing infrastructure and data supporting compliance and the use of RCUK funding are scalable, openly available and subject to scrutiny. Ensuring that all stakeholders work together with common and open standards will also promote innovation while facilitating a competitive market for scholarly information. C. Monitoring the effectiveness and impact of the RCUK Open Access policy 8. At PLOS, there has been an increasing number of publications since 211 from corresponding authors whose affiliations list a UK HEI or where RCUK is listed in the funding statements (Appendix 1). This is part of a larger trend in the number of CC BY papers published by pure 1 Publication Fees PLOS <http://www.plos.org/publications/publication-fees/> [accessed 26 September 214]. 2 With rare exceptions. PLOS publishers a small number of articles under the cczero waiver, the UK Open Government License and the Creative Commons Attribution IGO license, generally where authors are legally required to do so. These conditions rarely if ever apply to RCUK funded work. 1
Open Access Publishers 3 and does not take into account the rising number of subscription journals providing a hybrid option, the conversion of subscription journals to a full Open Access footing or the launch of new Open Access journals. There can be no doubt that the RCUK policy is having a positive impact on the scale of Open Access publishing, which is likely to be strengthened by the recent HEFCE policy that ties research evaluation to the accessibility of the output 4,5. D. Difficulties in the processes and workflows relating to APC payments to publishers, including ensuring that RCUK funding for open access is used for its intended purpose 9. The challenge of this transitional information environment necessarily means that the management of APC payments is made complex, manual and time intensive. PLOS is one of the few publishers with experience of managing APC payments at scale, dealing with invoices for 1s of articles every month. The challenges include: Matching APC payments to either a specific funder or institution (See also Appendix 1). Different types of invoicing (e.g. directly to institutions, funders or authors). For institutions, PLOS encourages institutional participation in PLOS s program for prepayments where a lump sum is paid in advance and APCs are drawn down against this fund 6 ; this option has been taken up by several institutions 7. Split billing, which is complex and hard to track. It is important to note that authorship, such as who is corresponding author, may have no bearing on the APC payment arrangement as payment will depend on those authors affiliated to institutions or funding agencies with access to funding to pay for APCs. PLOS minimises split billing by limiting the number of split invoices for authors to four (and authors are expected to work out logistics). There is no split billing for institutional accounts; the institution pays the full APC on behalf of the corresponding author either by direct billing (pay as you go, or month to month) or via our prepayment program. That, anecdotally, many authors are not always aware of existence of RCUK APC funds available within their institutions. This is even true in those institutions with exemplary implementation plans. That for some funders (albeit not RCUK), there is no funding mechanism for papers published beyond the end of an RCUK grant. 1. As the metadata standards and infrastructure to mitigate these issues are being put in place (see section F below), it is critical that all stakeholders work together towards creating a model that scales, is sustainable, and benefits all parties. All partners should also work actively with institutions and funders to develop best practices (including reporting and API) and recommendations for processing APCs that can be replicated internationally. 11. PLOS is participating in one such initiative termed ESAC 8 ( Efficiency and Standards for Article Charges ), which is addressing the challenges associated with the management of Open Access article charges (including APCs, article processing charges, article page charges, article fee etc.). 3 Cameron Neylon, The Rise and Rise of Creative Commons: Over 1.2M CC Licensed Scholarly Articles PLOS Opens <http://blogs.plos.org/opens/214/8/15/rise-rise-creative-commons-1-2m-cc-licensed-scholarly-articles/> [accessed 28 September 214]. 4 HEFCE: Policy for Open Access in the Post-214 Research Excellence Framework, 214 <http://www.hefce.ac.uk/pubs/year/214/2147/> [accessed 28 September 214]. 5 Catriona MacCallum, The Stick, the Carrots and the Bear Trap PLOS Opens <http://blogs.plos.org/opens/214/4/1/the-stick-the-carrots-and-the-bear-trap/> [accessed 28 September 214]. 6 Institutional Account Program PLOS <http://www.plos.org/get-involved/institutional-account-program/> [accessed 23 September 214]. 7 Program Participants PLOS <http://www.plos.org/get-involved/institutional-account-program/programparticipants/> [accessed 23 September 214]. 8 ESAC <http://esac-initiative.org/> [accessed 23 September 214]. 2
ESAC aims to start the discussion on efficient workflows involving all parties such as funders, libraries, authors, standardization initiatives, and publishers and to propose good practices and proven workflows. E. Difficulties in obtaining the data required to demonstrate compliance (see also 9 ) 12. As for APC workflows, workflows for monitoring OA compliance are also new. Challenges in obtaining the data required to demonstrate RCUK compliance include: Identifying an output that is subject to the policy: a substantial proportion of the scholarly literature does not have a unique ID that enables discovery via the web. This is particularly true for non-stm disciplines and for research outputs that are not journal articles. Disambiguating between authors of the same name. Identifying the specific funding or grant associated with an output. RCUK provides a publicly available list of their outputs via Gateway to Research, but this relies on author reporting. FundRef will help but our work at PLOS with specific funders has demonstrated that there is a substantial mismatch between the funders and grants acknowledged within an article and those articles that are reported to funders as relating to specific grants. In work with the Wellcome Trust we determined 2-3% of papers that acknowledge funding are not reported to funders and a similar proportion of those reported to funders do not contain acknowledgement in the article record. Ensuring the output is deposited in an appropriate repository within the permissible embargo period (if an APC has not been paid for a CC BY publication) It is currently impossible to survey repositories to determine whether a full text copy has been deposited and which licence is associated with it (i.e. the metadata provided by most OAI-PMH feeds does not provide sufficient information). Where information is provided on copies that are deposited but not accessible, this is not provided in a consistent form. OpenAIRE specifies requirements for repository metadata that define whether a full text copy is currently embargoed but only a subset of repositories are currently OpenAIRE compliant. Further, the OpenAIRE metadata does not provide sufficient information to determine whether the repository copy is licensed in a way which is compliant with the RCUK policy. Ensuring the output is associated with a licence that is compatible with the RCUK policy. Different publishers, different journals and different outputs from the same publisher provide licensing information in a range of inconsistent ways. F. Infrastructure Solutions 13. Efficient and rigorous systems to manage APC workflows and monitor compliance require that the metadata across different platforms and across different publishers, institutions and funders are standardised. 14. The tools, standards and systems that need to be in place to ensure this are already being developed: The FundRef 1 initiative is a key means of improving on funder acknowledgement in the bibliographic details of the paper. PLOS is working alongside other stakeholders and key vendors to implement FundRef. Ringgold 11 provides an institutional identifier for the publishing industry and became the first contracted ISNI 12 registration agency for institutions in 212. 9 Cameron Neylon, Policy Design and Implementation Monitoring for Open Access PLOS Opens, PLOS Opens, 214 <http://blogs.plos.org/opens/214/9/3/policy-design-and-implementation-monitoring-for-open-access/> [accessed 28 September 214]. 1 FundRef <http://www.crossref.org/fundref/index.html> [accessed 3 April 213]. 11 Ringgold Inc <http://www.ringgold.com/> [accessed 26 September 214]. 3
ORCID 13 provides unique identifiers for authors. The NISO Accessibility and Licensing Information Working Group 14 has proposed a metadata framework to standardise how licensing information is expressed by publishers. CrossRef 15 enables publishers to upload licensing information for existing DOIs. PLOS has funded a tool (the Open Article Gauge 16 ), built by Cottage Labs which aims to check the license for individual journal articles that have a DOI or PMID. G. Details of issues arising from the requirement for use of the CC - BY licence in gold publications (e.g. not being able to publish, not being able to reproduce required third - party materials, or researchers not being able to publish in a researchers first choice of journal or publish using the specified licence) 15. PLOS asks all authors to ensure that any third party content, such as figures, is compatible with a CC BY licence. It is technically possible to tag third party material that is not CC BY within an article that is otherwise CC BY but PLOS has chosen in principle not to do this in order to facilitate reuse of its entire corpus. Issues have arisen where authors do not understand the implication of the licences and use third party material they mistakenly think is available for distribution under a CC BY licence. Many publishers, for example Nature Publishing Group, will grant our authors permission to republish figures in PLOS articles under a CC BY license. 16. There are however cases where a publisher refuses this permission. We would emphasise that this does not prevent inclusion of figures or other elements within a CC BY licensed article either on the basis of a specific license or under a claim of fair use. PLOS has taken a policy position that such elements should not be included so as to maximise the clarity of downstream reuse rights. H. Cost of Publishing 17. PLOS APCs are $2,9 USD for PLOS Biology and PLOS Medicine, $2,25 for PLOS Genetics, PLOS Pathogens, and PLOS Neglected Tropical Diseases and $1,35 for PLOS ONE. PLOS believes that lack of funds should not be a barrier to Open Access publication and provides support for those who cannot afford the charges via two programs 1) the PLOS Global Participation Initiative 17 and 2) the PLOS Publication Fee Assistance program 18. 18. PLOS has not raised its charges for APCs since 29. We provide information about our revenue and operating costs available in financial statements published each year 19. In the Appendix we also provide information on the average APC (Table 1) and gross APCs (Table 2) charged for PLOS publications where corresponding authors have listed a UK HEI as their affiliation. Please note that this does not provide an indication of who paid (author, funder or institutions), if the bill was split across different parties, if there was a waiver request or if the invoice was unpaid. 19. We note a substantial contrast between the average price charged by PLOS and that charged by traditional competitors. The average cost to UK institutions for PLOS articles in our dataset is 12 ISNI <http://www.isni.org/> [accessed 26 September 214]. 13 About ORCID <http://orcid.org/> [accessed 3 April 213]. 14 NISO Access and License Indicators - National Information Standards Organization <http://www.niso.org/workrooms/ali/> [accessed 26 September 214]. 15 Crossref.org <http://www.crossref.org/> [accessed 12 June 213]. 16 How Open Is It? Open Article Gauge: A Service to Determine the License for Journal Articles <http://oag.cottagelabs.com/> [accessed 11 June 213]. 17 PLOS Global Participation Initiative PLOS <http://www.plos.org/publications/publication-fees/plos-globalparticipation-initiative/> [accessed 26 September 214]. 18 PLOS Publication Fee Assistance PLOS <http://www.plos.org/publications/publication-fees/plos-publication-feeassistance/> [accessed 26 September 214]. 19 Financials PLOS <http://www.plos.org/about/financials/> [accessed 26 September 214]. 4
$1542 (~ 95; Table 1). This amount differs substantially with those reported for other publishers in the Jisc submission to this panel 2 and in individual institutional datasets. Analysis by Björk and Solomon 21 and Aspesi 22 points to the failure of an effective market in APCs to emerge and constrain prices. A deeper analysis that compares quality of service with pricing is required but existing work in this space 23,24 suggests that prices relate more closely to perceived prestige than quality of service or quality of the journals in question. I. Recommendations 2. APC workflows and the Cost of Publishing i. Promote best practice for APC reporting and workflows through international and independent initiatives, such as ESAC. ii. We welcome that RCUK has encouraged institutions to report on the APC payments they are making. Transparency of pricing and payments is key to developing a competitive market. We therefore recommend that RCUK mandate that institutions make public information on payments they have made for APCs to publishers 21. Compliance Monitoring v. Require consistent metadata to ensure that the accessibility status of any output can be machine read. RIOXX metadata standards will address many aspects of this and illustrates the need for consistent metadata profiles to enable automated auditing. vi. Develop technical requirements documents that support the automation of policy compliance tracking. vii. Work with national and international peer funders to standardise policy language to avoid ambiguity and ensure that policy requirements are measurable. viii. Expand the CrossRef initiative on licensing across all publishers. 22. Future Monitoring xv. RCUK should work with national and international peer funders to aggregate, record and publish public data on scholarly communications ecosystem. Such an observatory would provide a platform that would improve compliance and enable evidence based policy design. The fragmentation of data collection and review between funders globally is a major barrier to high quality implementation of Open Access policy. 2 Neil Jacobs, Jisc s Evidence to the Review of the RCUK OA Policy Jisc Scholarly Communications <http://scholarlycommunications.jiscinvolve.org/wp/214/9/18/jiscs-evidence-to-the-review-of-the-rcuk-oa-policy/> [accessed 29 September 214]. 21 Bo-Christer Björk and David Solomon, Developing an Effective Market for Open Access Article Processing Charges, March 214 <http://www.wellcome.ac.uk/stellent/groups/corporatesite/@policy_communications/documents/web_document/wt p5591.pdf> [accessed 12 March 214]. 22 Claudio Aspesi, Reed Elsevier: Goodbye to Berlin - The Fading Threat of Open Access (upgrade to Market-Perform), 214 <http://www.richardpoynder.co.uk/aspesi.pdf> [accessed 29 September 214]. 23 Jevin D. West, Theodore Bergstrom and Carl T. Bergstrom, COST EFFECTIVENESS OF OPEN ACCESS PUBLICATIONS, Economic Inquiry, 52 (214), 1315 21 <http://dx.doi.org/1.1111/ecin.12117>. 24 Zoë Corbyn, Price Doesn t Always Buy Prestige in Open Access, Nature, 213 <http://dx.doi.org/1.138/nature.213.12259>. 5
Appendix 1. Data on papers where the corresponding author has listed a UK University as their affiliation. Data collation 1. The following tables and figures provide summary statistics of papers published by PLOS where the corresponding author has indicated an affiliation to a UK Higher Education Institution (HEI). We identified 17 individual UK HEIs. The data are restricted to the current reporting period (Apr 213- Jul 214) and the equivalent period of months before (Dec 211 Mar 213). Fig 1. B provides additional information on published papers from the beginning of 211. 2. We have comparable information where corresponding authors have listed their affiliation to other UK organisations (e.g. some ROs, the Wellcome Trust etc) and can provide that on request. 3. The data set was cleaned by hand to disambiguate between the > 1 different names of universities and different types of institutions provided by corresponding authors. 4. The DOIs of these papers were then used to obtain publicly available data via PLOS s ALM reports, which includes a single free text field listing all the funders acknowledged on the paper (as well as ALM data on each paper and a range of other fields, which are excluded from this analysis). 5. The data were further coded according to a free-text search of whether the funding statement acknowledged any of the 7 UK Research Councils. This included a search on acronyms and full spellings of the Research Council (and variations of either and or & ) plus RCUK (to pick up e.g. RCUK fellows). The aim, for example, was to explore how many papers where the corresponding author had listed a UK affiliation also listed one or more UK Research Councils. 6. Other selected funders that might be mentioned in the funding statement were coded separately to illustrate potential cross funder support for a study. For most of these, the search was for only one variant of their name (Wellcome, NIH, NSF or the Royal Society). Papers with EU funding were based on a search of European EU, FP7 or H22. Caveats to interpreting the data 7. Gross APC, where provided, does not indicate who was billed (author, institution or funder), if the invoice was split, if there were any waiver requests or if the bill remained unpaid. It is based on the total standard APC charge for the papers concerned. APC charges vary by journal (see section H of main text). 8. The HEI assigned by the corresponding author was clear in some cases but many also gave conflicting information (e.g. naming a UK University but providing a US city). In addition, some authors indicated that a RCUK Research Organisation was their primary affiliation but their email indicated a UK University. A best guess was made, which depended on a combination of different address/contact fields and e.g. how they listed their affiliation in the published paper. 9. Information about whether a particular funder was listed in the funding statement of the paper does not indicate that the corresponding author was a recipient of a grant from that funder (it would have taken too long to match the initials of the authors to particular funders within the funding statement). 1. The free text search of the funding statement will have missed any misspelt funder names. 11. Only tentative conclusions can be drawn from the data, especially about the impact of the RCUK Open Access policy on the growth of publications given that this is a snap shot from one publisher and the publishing landscape has changed rapidly over the period (e.g. the growth of hybrid offerings in subscription journals and the launch of new OA journals). 6
Table 1. Average APC for published papers at PLOS (Jan 211- Aug 214) where corresponding author lists a UK HEI. No. of HEI Institutions Total No. Papers Average APC ($ USD) 17 5248 1,542 Table 2: Number of PLOS published papers & Gross APC by affiliation, where the corresponding author has listed a UK HEI as their affiliation in the current RCUK reporting period (Apr 213- Jul 214), and separated according to whether an RCUK funder is listed by any author in the acknowledgements. Note that gross APC doesn t include waiver requests, who was billed (author, institution or funder) or if the invoice was split or unpaid. Only institutions affiliated to 2 or more papers in total are included. Institution (CA) No. Papers without RCUK acknowledgment No. Papers with RCUK acknowledgment Total No. Papers Gross APC ($USD) University College London 114 81 195 292,25 Oxford University 16 74 18 288,85 Imperial College London 8 75 155 257,8 Cambridge University 61 67 128 23,2 Edinburgh University 62 66 128 21,15 London School of Hygiene & 81 22 13 178,75 Tropical Medicine King's College London 66 31 97 148,75 Manchester University 41 42 83 117,45 Glasgow University 44 19 63 14,35 Liverpool University 35 28 63 93,3 Newcastle University 28 33 61 88,65 Bristol University 29 32 61 88,15 Nottingham University 35 25 6 85,9 Birmingham University 29 26 55 77,85 Warwick University 16 28 44 67,25 Aberdeen University 25 18 43 65,25 Queen Mary University of 26 16 42 62,25 Leeds University 17 25 42 58,5 Leicester University 28 14 42 59,4 Southampton University 25 16 41 58,5 Sheffield University 21 16 37 55,1 York University 21 14 35 55,1 Cardiff University 18 14 32 45,9 Exeter University 16 14 3 44,75 Sussex University 13 16 29 47, Liverpool School of Tropical 24 3 27 5,1 Medicine Queen's University Belfast 2 7 27 36,45 Swansea University 21 5 26 35,1 Dundee University 8 17 25 41,6 St Andrews University 13 8 21 3,15 East Anglia University 7 13 2 28,8 7
Fig 1 : For PLOS published papers where the corresponding author has listed a UK HEI as their affiliation and split by whether they also list an RCUK funder in their funding statement: A. Comparison of the number of papers in the reporting period with that of the equivalent time before (total N = 4,28); B Comparison of the number papers published since 211 (total N = 416; publications in 214 are omitted as the year is incomplete). Note that in each graph the corresponding author may not have been a recipient of RCUK funding. A. No. Papers 16 14 12 1 8 6 4 RCUK not listed RCUK listed 1131 8 1386 963 No Papers 12 1 8 6 4 2 RCUK not listed RCUK listed B. 2 < Reporting Period (Dec 211-Mar 213) Reporting Period (Apr 211-Jul 214) 211 212 213 Year Table 3: No of PLOS published papers where the corresponding author has listed a UK HEI as their affiliation in the current RCUK reporting period (Apr 213- Jul 214) and the equivalent period of months before (Dec 211 Mar 213), and where that University has received more than 2, as a block grant for APCs from RCUK 25. Note that the corresponding author may not have been a recipient of RCUK funding. Institution (CA) RCUK Block Grant (Rank) No. Papers < reporting period (Dec 211-Mar 213) No. Papers in Reporting period (Apr 213-Jul 214) Cambridge University 1 148 128 Imperial College London 2 127 155 University College London 3 151 195 Oxford University 4 167 18 Edinburgh University 5 91 128 Manchester University 6 67 83 Bristol University 7 63 61 Nottingham University 8 52 6 Southampton University 9 32 41 Sheffield University 1 39 36 Leeds University 11 32 42 King's College London 12 67 97 Birmingham University 13 4 55 Glasgow University 14 46 63 Liverpool University 15 49 63 Warwick University 16 21 44 Newcastle University 17 63 61 Cardiff University 18 17 32 Durham University 19 17 17 Queen Mary University of London 2 39 42 Total 1328 1583 25 RCUK APC Block Grant (212) <http://www.rcuk.ac.uk/rcukprod/assets/documents/documents/rcuk_apcfunddistribution.pdf> [accessed 28 September 214]. 8
Fig. 2. For published papers in the current reporting period, where corresponding author is affiliated to a UK HEI: A) The number of papers that list a different number of UK Research Councils in their funding statement (blue= no RCUK funder; red, 1 or more; total N = 2349); B) No papers listing each UK Research Council (total N= 963 papers). Note that funding is that listed by any of the authors and does not indicate funding status of the corresponding author. A. No. Papers 16 14 12 1 8 6 4 2 1386 784 151 28 1 2 >3 No. of different RCUK funders listed No. Papers 6 5 4 3 2 1 494 329 121 12 86 Funding Type 5 2 B. MRC BBSRC EPSRC ESRC NERC AHRC STFC Fig. 3. For published papers in the current reporting period, where corresponding author is affiliated to a UK University: A) the % of papers that list different UK Research Councils (N=963); B) the % of papers listing other selected funders in funding statement of papers and separated by whether a RCUK funder is also listed or not. Note that funding is that listed by any of the authors and does not indicate funding status of the corresponding author. A. MRC BBSRC EPSRC ESRC NERC AHRC STFC % % 8% 9% 11% 29% 43% % Papers 35 3 25 2 15 1 5 Wellcome EU NIH Royal Society NSF 2 17 13 3 RCUK not listed (N = 1386) 3 22 13 6 B. 1 1 RCUK Listed (N = 963) 9