Select Committee on Science and Technology Written Evidence


APPENDIX 128

Supplementary memorandum from the Association of Learned and Professional Society Publishers

1a   "According to the data you have collected, how many of your members have already digitalised their back catalogues, and how many have plans to do so?"

  There are two types of digital back-files: those which were created as a "by-product" of publishing in electronic form, at little or no additional cost, and those which have had to be specially retrodigitised from earlier years. A growing number of publishers, particularly not-for-profit, make the "by-product" archive freely available to all after a short period (see, for example, the recent "ALPSP Principles" at http://www.alpsp.org/2004pdfs/SFpub210104.pdf http://www.alpsp.org/2004pdfs/SFpub210104.pdf and the "DC principles" at http://www.dcprinciples.org).

  However, the very substantial cost of retrodigitising a complete back-archive normally makes it necessary to charge for access, in order to recoup the investment. Some journals, such as those of the Royal Society, have been retrodigitised at no cost to the publisher by J-Stor, a non-profit foundation originally funded by Mellon, but now self-supporting; J-Stor recoups its costs through library charges. Completely free access is only possible in cases where the retrodigitisation has been externally funded (eg by PubMed Central or the Wellcome Foundation).

  In our survey in early 2003 of 149 publishers (including all the major publishers, and covering members and non-members, commercial and not-for-profit) we found that 85.2% make back volumes available online, in the majority of cases from 1997 or 1998 (which is when they started e-publishing). However, over 20% had an average start date earlier than 1995, and at least 10 had digitised from Vol 1 No 1 (for cost reasons, retrodigitisation is often a gradual process and some have not yet gone all the way back to the first issues). In the past year, we have seen many further announcements of publishers carrying out total or partial retrodigitisation of their journals (for example, the Royal Society of Chemistry has gone right back to the first issues in 1841; Taylor & Francis are working their way through their journals, and have covered five so far).

1b  "Can you supply an estimate of how much this activity increases the publishing costs of a society publisher?"

  The one-off cost of creating a digital archive is not a publishing cost except in a very general sense. "Publishing costs", as examined by the Committee and covered in much of the evidence, relate to the initial cost of publication (peer-review, production, distribution, DOI registration, marketing etc). Retrodigitisation, on the other hand, is a "re-purposing" of already published content; most of those publishers which have articulated comprehensive retrodigitisation plans have treated these direct costs as quite separate from ongoing publishing, and indeed have priced the resulting back volume collections quite separately from ongoing subscriptions.

  This cost of retrodigitisation of the journal archive depends entirely on the size of the back-file and the kind of functionality to be incorporated into the electronic archive. The print copies must not only be scanned (usually resulting in destruction of the original) but also comprehensively indexed so that the resultant files are retrievable; the archive then needs to be hosted within a suitable system. At the very least, most publishers also create XML header records for each article, and most are planning to assign DOIs to them and deposit them with CrossRef, at a cost of $1.00 per article. Full XML tagging is much more expensive than creating a "static" PDF file, and most publishers to date have tended towards the latter option for this reason.

  A typical scanning cost is 16-25 pence per page, depending on complexity, illustrations etc; ALPSP members have recently been quoted a preferential discount from Apex e-publishing of at least 10% off this figure. Thus the scanning alone of a journal consisting of 6 x 80-page issues per year would cost £70-120 per annual volume; however, the frequency (number of issues) and extent (number of pages per issue), as well as the age of the journal, varies greatly and thus it is impossible to give an average figure per journal.

  There will also be system costs in creating a database in which to host the archive, and an access mechanism for authorised users to search and retrieve content from it. The staff and equipment costs are also very considerable. For example, the Royal Society of Chemistry's recent complete digitisation of back issues from 1841-1996, comprising 1.2 million pages, cost approximately £500,000 for the actual digitisation, and a further £300,000 for staff, equipment and other costs (67 pence a page in total); that of the Institute of Physics, going back to 1874, comprising some 750,000 pages, cost approximately £500,000 including internal costs (75 pence a page in total).

2a  "How many of the societies you represent have journals that are accessible via the Science Direct platform?"

  Science Direct is the proprietary platform for Elsevier's journals. It (or its sister platform MDConsult) includes 189 of the 219 society journals which Elsevier publishes under contract (the remaining 30 remain offline at the societies' request); it is also, in principle, open to other publishers and hosts another 67 journals (50 from one learned society publisher) on this basis. In total Elsevier estimate that they work with over 150 societies.

2b  "Through which other web-based platforms do your members offer access to their journals?"

  Many larger publishers have their own platform (eg Wiley's InterScience, Springer's SpringerLink); the same is true of some of the larger society publishers (eg Institute of Physics, Royal Society of Chemistry, Institution of Electrical Engineers).

  Small and medium-sized publishers typically work with third-party hosts, among which the largest are probably Ingenta (which hosts 4,500 journals and provides access to a further 1,500, both commercial and not-for-profit) and MetaPress (1,800 journals, 40% society-owned; their largest client is Taylor & Francis). Other hosts include Stanford University's HighWire (350 journals, all but five not-for-profit, and many offering free access to back-files after a few months; clients include key players such as Oxford University Press and British Medical Association), BioOne (all not-for-profit), Johns Hopkins University's Project Muse (all not-for-profit), Extenza (approx 300 journals, mixed) and the services offered by larger societies to smaller ones such as the American Institute of Physics and the Biochemical Society. Access to the titles in BioOne and Project Muse can be licensed as a complete package; however, journals on all the other platforms mentioned above are licensed on a publisher-by-publisher basis, both as individual titles and, in many cases, as a "bundle" of that publisher's titles.

  In the USA, some printers (eg Allen Press, Sheridan Press) also offer online hosting services to their (mainly not-for-profit) client publishers, although this is less common among UK printers (Charlesworth is one which does so).

  Access is also provided through the websites of all the major subscription agents, although the content is hosted elsewhere: EBSCOHost, SwetsWise, Harrasowitz, Minerva are examples of such services.

  In addition, aggregators sell collections of content which has been sub-licensed by multiple publishers; these provide good value for money, but the journal content is sometimes provided after a time delay to protect subscriptions. The main aggregators are EBSCO Publishing, Gale Group, ProQuest and Wilson.

  Not-for-profit organisations such as ALPSP itself, HighWire, BioOne, ProjectMuse and American Institute of Physics all offer multi-publisher collections to simplify acquisition and access by libraries. The ALPSP Learned Journals Collection is the only instance, however, which does not include journals hosting; the journals are accessed through a single portal (provided by subscription agent Swets) but continue to be hosted on a variety of platforms at the preference of the individual publisher.

  A publisher's content is likely to be available through a number of these different platforms; in addition, their content is increasingly linked (using the Digital Object Identifier standard and the CrossRef technology) both to and from citations within the literature, and to and from bibliographic details and abstracts in the numerous "abstracting and indexing" databases which provide the preferred route for about half of all journal usage.

3   "According to the data and views you have collected, what proportion of your members would be adversely affected by a move to the author-pays publishing model? What proportion would benefit from such a change?

  This question is impossible to answer because there are so many variables, in particular the amount which it is possible to charge authors (per accepted publication, submission, or both) and how many of them are able and willing to pay; the costs are also affected by the percentage of articles processed but rejected by the peer-review system. Science publishers seem to be more confident of adequate income from payments on behalf of the author (rather than on behalf of the reader) than are those in the arts, humanities and social sciences; in these subject areas, research grants are small or non-existent and thus the funding would simply not be available.

  Responses so far to our current survey indicate that 67% (20/30) of society publishers do make a surplus from their publishing (we understand that the figure may be lower in the USA); both those who publish on their own account and those which contract out their publishing are highly dependent on the payments from their publisher to subsidise their other activities. By and large, it seems to be their expectation that they would make a reduced surplus from Open Access publishing, even bearing in mind the concomitant savings which could be realised if the shift were combined with the abandonment of a print edition; one estimate is that surpluses in excess of 12% would be likely to be reduced. In the above survey, median surplus was 17%, representing 33% of society income. There would also be additional one-off costs involved in the transition from the existing business model to Open Access.

  The effect of such a reduction in surplus would be to reduce expenditure on various vital aspects of the societies' mission. In our survey, 82% of respondents used part of the surplus to support the general expenses of the society. 42% used some for research grants and bursaries; 32% used some to subsidise conference fees, and 29% used some for public education. Thus any move which led to a substantial reduction in publishing surpluses could damage or even destroy many learned societies; vulnerability will to some extent depend on the range of other revenue streams which the society has been able to develop.

  We have just commissioned a major research project from Kaufman-Wills Group LLC to analyse all the data which can be made available to us from the wide variety of Open Access journal experiments (under various models) currently being carried out. This will enable us to arrive at some conclusions not just about the financial effects of moving to OA, but also the effects on author and reader behaviour. The results will be published in the first half of 2005, and should provide publishers of all types with a better basis on which to make informed decisions about any change to their own business model, and to plan the transition—if they decide to make the change—in a way which minimises any negative impact. We believe that such investigation will be assisted by any recommendation to funding agencies to permit their grants to be used for Open Access publishing charges; however, we also believe that there is not yet enough information about the effects of Open Access journals for it to be appropriate to mandate grant recipients to publish under this model.

  It may also be relevant to note an informal survey being carried out by one of our members, asking fellow societies about the extent of demand from their own members for Open Access. So far the figure stands at 0/387,000 from 17 societies—responses are still coming in; however, the Physics societies do report that many of their authors are also self-archiving (usually preprints, rather than final versions of articles) in the ArXiv subject repository.

  We do not, in fact, believe that a significant reason why Open Access might damage learned societies is through erosion of their main member benefit—the member's free or cheap copy of the journal. Those societies for which the journal is still the main or only benefit have already experienced this erosion through site licences, which provide most of their members with online access to the journal at their desktop, thus reducing the requirement for a personal copy. The reaction has generally been to concentrate on offering other attractive member services.

CLOSING STATEMENT

  On behalf of its 280 members, ALPSP would like to reiterate the following points:

    —  The current system of journal publishing has much that is of proven value, in particular the system of peer review. Any evolution of publishing models needs to preserve this value.

    —  Learned society and other not-for-profit publishers are of key importance. At least half of all journals are published by or on behalf of learned societies, university presses, non-governmental organisations, research foundations and the like. Many studies have shown that their journals are on average significantly less expensive, and of higher quality, than commercially published journals.

    —  Transition to electronic-only journals would help to reduce libraries' funding problems, and would be welcomed by publishers. However, there are two major obstacles to this:

      The absence of an assured system for long-term archiving and preservation; the most coherent and cost-effective approach is via our national library, rather than piecemeal by individual publishers and libraries, and adequate funding for this is essential.

      Full-rate VAT is charged on electronic journals, while print journals are zero-rated and combined subscriptions fall somewhere in between; this makes electronic journals more expensive to libraries, rather than less. We advocate exempting the recipients (non-commercial educational and research establishments) in the short term, and working within Europe to change the VAT rules in the longer term.

    —  Any measures taken in the UK alone will be of relatively minor effect; the majority of UK authors are published in non-UK journals, and the majority of UK journal sales are to non-UK customers.

    —  We believe that the market should be allowed to evolve. Permitting recipients of grant funding to use part of the money to pay publication costs would be helpful in enabling realistic testing of the Open Access model. However, we are concerned that the model may not be viable in all circumstances and could, indeed, be damaging to some learned societies; we therefore recommend a period of experimentation and research, and would strongly advise against any mandatory action at this early stage.

March 2004



 
previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2004
Prepared 20 July 2004