November 3, 2003

An Oxford NARrative

Researchers enjoy the process of designing a carefully thought-out experiment and interpreting the data it generates. Oxford University Press (OUP) is currently conducting an experiment in publishing that is likely to provide insights into the transition to Open Access. Martin Richardson, director of the OUP Journals Division, talked to Open Access Now about the design and rationale of the experiment.

"We are not going into the experiment trying to judge what the outcome will be," says Richardson. "We are trying to learn about Open Access from the experience of one particular journal." Richardson is Director of the Journals Division at Oxford University Press (OUP). Structured as a department of Oxford University, OUP primarily publishes books and has a tradition stretching back to the fifteenth century. But the Journals Division publishes over 180 journals, spanning a broad range of subject areas including biomedical research, other scientific disciplines and the humanities; OUP has been publishing journals since the middle of the nineteenth century, and close to 100% of its journals are now published online.

The pressure to experiment
Richardson and colleagues are very aware of demands from parts of the academic community for research to be made freely available online immediately upon publication. They wanted to respond to these calls and to test models for a transition to Open Access publishing with a well-established journal. Richardson notes that the life sciences community was particularly concerned about the need to find viable financial models to fund Open Access, so OUP has chosen one of its flagship scientific journals, Nucleic Acids Research (NAR), for the Open Access experiment.

"The goal of the project is really to learn from practical experience about the possible transition for a well-established journal from being funded primarily by subscriptions to being funded primarily by author charges," says Richardson. "Most experiments with Open Access publishing have been done with newly launched journals. We decided that if Open Access is really going to take off as an alternative model, then it has to apply also to well-established journals that have existing revenue derived from other sources."

"We have had many discussions over the past year about the experiment that we are now embarking on, and NAR seemed like an obvious choice. It has a strong track record of experimentation - NAR was one of the first journals to be published online, it led the way in encouraging authors to deposit data into public databases, and so this was sort of a natural extension." NAR was also one of the first journals to index its full-text content in PubMed Central - initially with a twelve-month delay and, after a period of experimentation, now with a six-month delay.

"So, we decided to start with one section of the journal - the annual Database Issue. This issue contains about 130 papers; we felt that this was a big enough slice to be a good sample," explains Richardson. "The Database Issue also has a wide geographical spread of authors - mostly from Europe and the US, but also from elsewhere - and quite a number of authors who publish in that section on a regular basis. This would give us a sample of previously published authors with whom we could test out our ideas."

OUP will use the experiment to evaluate how authors respond, and to identify any concerns that arise. Richardson says that the information will then be used to decide about a possible transition for the rest of the journal, depending on how the experiment with this one section goes.

The NAR Open Access experiment is still rather tentative. They have not yet fully embraced Open Access publication as defined by the Bethesda principles (see Open Access Now, July 14, 2003).

Experimental design
As with all good experimental strategies, OUP had to do some background research to guide the experimental design. "We put together the experiment after full consultation with authors who have published in NAR over the past three years," explains Richardson. "We surveyed around 300 authors, a reasonable portion of whom are likely to go on publishing in the Database Issue in the future. We wanted to ask people who had experience of publishing in this section what they felt about Open Access, and in particular whether the transition that we had come up with was one that they felt comfortable with."

OUP sent a letter to previous contributors to the NAR Database Issue explaining the funding options available for publication in the journal, and asking the contributors whether they would support the introduction of author charges for the 2004 Database Issue. In return for payment of author charges, the entire Database Issue would be freely available online to all, immediately upon publication.

The response was very encouraging, with a two-thirds majority voting for a move to an Open Access publication model for the 2004 Database Issue. "There was a fair amount of discussion amongst the Editors and ourselves about precisely what the experiment should be," says Richardson. "We did look at a number of other models, and in particular a model where we would offer optional Open Access, giving authors a choice for each particular paper as to whether they wished to pay for Open Access or not." This model had been developed by a number of online journals, and The Company of Biologists has recently adopted this approach for their journals, which include Development and Journal of Cell Science (see Open Access Now, October 6, 2003).

Richardson says that OUP rejected this option - despite it being a perfectly viable way of looking at a transition - because it didn't want to create a two class system based on the ability to pay. "There was quite a strong feeling, from both ourselves and the Editors, that there might be resistance to a twin track system distinguishing between authors who could afford to pay to have their material Open Access and those who could not afford it," he explains. "We did not want to be associated with a system that might be perceived as unfair, and we are very keen not to put any barriers in the way to prevent high quality material being published."

"We are doing the experiment to learn"

Martin Richardson

All the articles in the January 2004 Database Issue will be available with immediate Open Access, and authors of each accepted manuscript will be asked to pay a mandatory charge. "If authors are unable to pay - either if they are in a developing country, or if they are in an institution or funded by an agency that does not recognize this kind of charge - then we will consider waiving the charges," says Richardson.

The charges in the first year will be £300/US$500 per article. OUP is keen to stress that this represents a substantial discount on the charge that would be required to fully fund publication. Richardson says the eventual Open Access charges will depend on the uptake and how many people turn out to be willing to pay. "If we have to waive the charges for a substantial number of authors then the costs will have to be subsidized by those who can pay and by subscription revenues. But if most authors pay, we envisage that the charge will have to be about four times this year's figure."

The kinetics of transition
OUP emphasizes that this is an experiment and it is keeping an open mind about the results. "We don't yet know how things are going to turn out," says Richardson. "The deadline for submitting papers for the database issue has now closed, but we are still in the review process. We are very keen to separate the editorial process from any discussion about fees. As far as we can tell it hasn't affected submissions, and submissions to that section have been increasing quite rapidly in recent years. Until the editorial decisions have been made for these papers we are not planning to contact the authors to see if they are willing to pay the charges."

Richardson is cautious about making predictions about the outcome of the experiment. "We will see whether it works or not. We have tried to put the experiment together in such a way that we can assess any problems - and we do anticipate there being problems. A number of authors during the survey expressed concern about the ability of authors and funding agencies to pay for publication in this way."

OUP has expressed a commitment to extend Open Access to the whole of NAR if the experiment proves successful. Richardson says that success will be monitored from two perspectives. "The first, and probably the most important, is whether authors who publish in NAR are happy with the kind of structure that we come up with. The second is whether it is a viable model from our point of view, so that we can see a way that the costs of publishing those papers can be recovered. At the beginning it won't cover the costs, but we have made it clear that provided we get a good response to this in the first year, we would envisage making the transition to move more of the burden of the costs to the authors and reduce the proportion coming from subscriptions."

OUP believes that the adoption of this transitional approach is a sensible strategy for both authors and readers. It is aware that changes in funding conventions may be required so that all authors have access to sufficient financial resources to pay for the full costs of publication of their research. OUP has stated that "the 2004 Database Issue experiment is very important to us; if successful, it could be the key first step in migrating the journal to author-funded Open Access publication."

But it is likely to take some time for a full transition. "We don't have a precise time-frame," says Richardson. "But we think it will be hard to do it in less than four or five years." He expects that the mix of subscription and author charges will be very different in five years time. "My personal feeling is that Open Access is not as black-and-white as the discussions that have taken place so far would suggest." Richardson predicts that different subject areas may require different models, and he is not ruling out the possibility of a hybrid model that would include both author- and subscription based revenues.

Richardson's experiment has had interest from many people in the publishing and library communities who are keen to see how things turn out. He predicts that other publishers will also be conducting experiments. "Experimentation is something we are very comfortable with," emphasizes Richardson. "It's hard to know about these things until you actually try them."

nar.oupjournals.org
www.oupjournals.org

Anthony Maxwell, John Innes Centre in Norwich, UK, is an Executive Editor at Nucleic Acids Research. He talked to Open Access Now about the transition experiment.

"Over the course of the past couple of years at our annual editorial meetings there has been a significant body of opinion amongst Executive Editors to push the journal in the direction of Open Access. There was some initial resistance from Oxford University Press (OUP) - for understandable reasons - but they were open to discussion about new ideas."

"The ultimate goal is to make Nucleic Acids Research an Open Access publication. But OUP cannot do that in one step. So, we came up with a compromise situation that allows a stepwise transition to Open Access. This starts with the Database Issue, which I think should be seen as a litmus test for the viability of this type of approach."

Maxwell acknowledges that much of the driving force came from the senior US editor Richard Roberts, and that there are some differences of opinion amongst individual Executive Editors about how best to proceed. "I personally am slightly frustrated that we cannot proceed more quickly, but I understand the commercial constraints that have to be brought into consideration."

"What we have developed is a mechanism for trying to persuade authors that this ultimately is the way it is going have to be. If authors agree to come up with a contribution to publishing costs then OUP will be able to go down that route," explains Maxwell.

Maxwell emphasizes there are ways to deal with authors concerns about costs. "I think that many institutions will end up paying a blanket authorship fee for the institution, which will then liberate the author from that responsibility. In my view this is by far the best way to go. But it is generally felt that universities and funding institutions are not ready yet to take that step."

Maxwell expresses some doubts about the strategy of starting new Open Access journals that have to build up a name and reputation. "There is no doubt in my mind that if one could get two or three established journals with high impact factors to switch to Open Access then everyone else would follow. This would generate the critical mass needed to kick everyone else in the Open Access direction."