Open access : evidence-based policy or policy-based evidence ? The university press perspective

Based on a paper presented at the UKSG seminar 'Scientific Publications: Free for all?', The Geological Society, London, Tuesday 23 November 2004 Open access has both advocates and sceptics. It could be said that we are still at a 'superstition and witchcraft' state of knowledge: everyone has an opinion, but hard facts are lacking, and few who speak either for, or against, open access can provide detailed evidence for their viewpoints. Oxford Journals is keen to experiment with any new business models that might help us fulfil our mission of maximizing accessibility to research information. The objectives of our experimentation are both to test the long-term financial viability of new models, and also to ensure that they really do achieve wider dissemination than existing models.


Introduction
There are many evolving models for publishing and distributing research information and, whilst this can seem confusing, ensuring that many avenues are explored rather than just one helps us avoid moving too far along one road only to find that we meet a dead-end.It would be wrong at this stage to advocate any one model over another until more evidence is available to support a proposed large-scale move away from the traditional subscription-based model.
At Oxford Journals, a division of Oxford University Press, we are taking the approach of active experimentation to the problems of open access, using a variety of different distribution methods.These include: Through these experiments we will be able to test the real market appeal, the impact on usage of 'open' v. 'closed' access journals, and the financial and administrative issues surrounding such models.

a) Subscriptions and free access
In common with many publishers, we offer print and online subscriptions, which can be purchased together or separately.Single journals can be bought by an individual institution, or our entire online collection, or sub-sets, can be purchased at discount by library consortia.In addition to this, single articles can be obtained through document delivery services, or purchased immediately online through our payper-view facility.Our online journal content is also widely licensed through aggregator services; we offer free online access to developing countries and free access to back-issue archives.

b) Open access
We are experimenting with converting an ) and of market research (both with authors -from whom there was overwhelming support to make this change -and librarians).
Traditional full-priced subscriptions to NAR had been in decline for quite some time, and this decline was accelerating with the introduction of free online access to articles six months after publication.In distribution terms, this loss was more than made up for by the additional circulation achieved through consortia sales and free access to developing countries.However, the long-term financial viability of the journal was in question, so open access provided us with the potential both to stabilize the revenues and maximize dissemination.A complete overnight switch was not going to prove possible until further changes were made to research funding, so we decided on an interim model with author charges subsidized by institutional membership and print subscription revenues.
For 2005, we have introduced a publication charge of $1500 per paper, which is reduced to $500 for authors from institutions with print subscriptions, or if the author's institution joins our membership programme (annual charge of $2459).We simultaneously publish in the PubMed Central archive, and offer unlimited re-use of articles for research and educational purposes.

Case Study Two: Journal of Experimental Botany
In this experiment, authors are given the option of paying £250/$400 for their article to be made freely available online.In addition, JISC funding supports a waiver for this charge for all UK authors.Review articles and special issues are not open access.We are keen to test the uptake of the open access option at this price, and aim to hold the subscription price at current levelswith the longer-term objective of decreasing it should the revenue from authors become sufficient to do so.

Complementary & Alternative Medicine
This new journal (eCAM) was launched in June 2004 and made freely available online through the sponsorship of the Ishikawa Natural Medicinal Products Research Centre in Japan.As a result, there are no charges to authors.The print edition continues to be sold on subscription.
Our initial analysis of the impact of this approach on usage remains inconclusive.Average usage per article remains very similar to other medical journals that we publish; although there is an interesting difference between research and review articles, with research articles in eCAM attracting 41% higher usage, but review articles attracting 26% less usage -in effect, balancing each other out.Perhaps the established brand and readership helps the better-known journals achieve higher usage for review articles, even with paid subscription access, whereas readers looking for research articles are more likely to search across a variety of journals.

c) Free access for authors
We offer authors free e-offprints by automatically sending them a toll-free link on publication of their paper.They can e-mail the link to colleagues, or post it on their web site.

d) Institutional repositories
Oxford Journals are working in partnership with Oxford University Library Services (OULS) in support of the national SHERPA project.We enable free online access for OULS to articles by Oxford University-based authors that have been published in many of our journals from 2002.Articles are searchable via the OULS pilot institutional repository, and are available free of charge to researchers across the globe.

Does free access mean higher usage?
One of the key questions we hope to answer through our experimentation is the question of

Martin Richardson
Open access: the university press perspective whether or not free, or 'open', access leads to higher usage.This effect would be important in helping publishers to judge whether a changed model was required, but information on this remains inconclusive.
In our own experimentation we have found that the database issue of NAR (which was published under open access in 2004) has indeed attracted greater usage than subscription-based issues; but then the nature of the content lends itself to greater usage, so it is difficult to draw firm conclusions.We will know a lot more in a few months' time, once we can compare 'like with like', as regular issues of NAR, as well as the database issue, become freely available online.In addition, we will be able to compare the usage of open access articles published within a single issue of the Journal of Experimental Botany with subscription-based articles.We will continue to monitor the usage trends of eCAM to see the long-term effect of free access for a newly launched journal.

Does free online access mean cancelled subscriptions?
Studies of subscription circulation trends for 28 of our journals with free back issues imply that where the subscription barrier is lifted after six months or less there is a significant impact on paid subscriptions.For those journals offering free back archives after six months, the average attrition rate from 2002-2003 was 6.1%.At 12 months it was 2% and at 24 months 2.4%, so six months does appear to be enough to have a negative impact.

Next steps
At Oxford Journals we want to continue experimenting, and encouraging others to join us -both in additional experiments, and also in the sharing of our findings.It is in all our interests to explore evolving models, and to seek to find new ways of maximizing the accessibility and impact of the research information we publish.Over the next few months, we plan further usage analysis along with a detailed review of the relationship between usage and citations/impact factor.In addition, we will carry out further research into the financial impact of various models and complete a cost/benefit analysis for each.We look forward to sharing the results of our experiments with the community in order that we can play our part in defining new methods of dissemination to ensure the widest possible, cost-effective access to research articles.
a) subscriptions and free online access b) several different forms of open access c) free access for authors d) working in collaboration with SHERPA (Securing a Hybrid Environment for Research Preservation and Access) to test the viability of institutional repositories.

Open access has both advocates and sceptics. It could be said that we are still at a 'superstition and witchcraft' state of knowledge: everyone has an opinion, but hard facts are lacking, and few who speak either for, or against, open access can provide detailed evidence for their viewpoints. Oxford Journals is keen to experiment with any new business models that might help us fulfil our mission of maximizing accessibility to research information. The objectives of our experimentation are both to test the long-term financial viability of new models, and also to ensure that they really do achieve wider dissemination than existing models.
MARTIN RICHARDSON Managing Director Oxford Journals, Oxford University Press Research), partial open access (see case study two, below: Journal of Experimental Botany) and sponsored open access (see case study three, below: Evidence-Based Complementary & Alternative Medicine).