The green and the gold roads to Open Access

The crisis in university journal budgets first brought to light the problem of access to published research. But the problems of affordability and access, although often confused, are distinct. We describe here a practical solution to the access problem.

There are some 24,000 peer-reviewed journals¹,², publishing around 2.5 million articles per year. Spiralling price rises mean research libraries can afford to subscribe to fewer journals and, in turn, users have access to fewer articles-whereas in the online age, we might have expected the opposite. Even if all journals were sold at cost price, almost no university could afford all or even most of them³. This is the journal affordability problem.

Moreover, as users' institutions cannot afford access, the potential impact of many articles is lost. The impact of an article is the extent to which it is read, built upon and cited. Researchers 'vote' on the relevance and utility of an article by using and citing it. But before one can cite an article, one must first be able to access it. That is the research access/impact problem.

Impact is a metric of scientific progress and productivity. Researchers' careers depend on their citation track record, and universities and research funding agencies reward impact, as they are accountable for the returns on taxpayers' money.

Is there a way to make research articles accessible even to those users whose universities cannot afford journal access to them? Before answering this, we need to ask: How big is this access/impact problem?

One way to estimate it is to compare citation counts for Open Access articles with pay-to-access articles. Lawrence⁴ found that in computer science citations were three times higher for Open Access articles than for papers only available for payment in print or online. Kurtz et al.⁵,⁶ have since reported similar estimates in astrophysics, and Odlyzko⁷ in mathematics.

We are carrying out a much larger study across all disciplines, using a 10-year sample of 14 million articles from the Institute for Scientific Information (ISI)'s database; initial results, for the field of physics, show Open Access articles being cited 2.5 to 5 times more than articles that users' institutions must pay to access online, with this advantage peaking within about 3 years of an article's publication⁸.

All these articles were published in subscription-based journals, but some were made accessible because authors had 'self-archived' copies on the Web-see http://www.eprints.org/self-faq/. Physicists have been self-archiving in growing numbers since 1991 in a central archive called ArXiv. Computer scientists have been self-archiving on their own websites, which are then harvested by Citeseer.

Another self-archiving method, and the one we believe has the highest potential, is for authors to self-archive articles in their own institutional archives (http://software.eprints.org/handbook/). Such archives are compliant with the Open Archive Initiative (OAI) metadata-tagging protocol. This makes distributed institutional archives 'interoperable', meaning that they can all be searched as if they were one, using search engines such as OAIster.

The functionality of open archives can be enhanced by citation indexing (see Citebase), which ranks articles using either their number of citations by other articles in the archive, or the number of downloads⁹. Both measures correlate well enough so an article's current download impact can predict its future citation impact. Such performance indicators can be customized and collected by research evaluators to chart the progress and direction of research, and to assist decisions on promotion and funding¹⁰.

In a recent survey¹¹, the majority of authors reported that they would self-archive if their employer, or funder, required it - 20-40% of them were already self-archiving. We believe the most promising way to achieve the goal of Open Access is for institutions to introduce policies requiring that published articles be self-archived. It is they and their researchers who will benefit from maximizing research impact and eliminating the costs of lost impact. This should motivate authors and their institutions to create and fill more archives --100 universities worldwide already have).

Over 80% of journals are already 'green', that is, they give authors a green light to self-archive in some form . About 5% (almost 1,000 journals) are 'gold', that is, they are Open Access journals. However, to recover publication costs no longer covered by subscriptions, many of these Open Access journals charge authors. The riskiness of this untested model makes publishers more willing to go green rather than gold.

An unfortunate tendency has arisen to equate Open Access itself with Open Access journal publishing. But there is more than this one, golden, road to Open Access: there is the faster, surer and more heavily travelled green road. We think that authors may be confusing the journal-affordability problem with the access/impact problem. Let us hope that the prospect of increased impact will persuade the majority of authors and their institutions to take to the green road so we can all enjoy its rewards.

A fuller version of this article, with figures, is available at the first author's website.

Stevan Harnad Canada Research Chair in Cognitive Science, Université du Québec à Montréal, Canada, harnad@uaqm.ca

Tim Brody Doctoral Candidate in Computer Science, Unioversity of Southampton, tdb01r@ecs.soton.ac.uk

François Vallières Computer Analyst, Observatoire des sciences et technologies (OST), Centre interuniversitaire de recherche sur la science et la technologie (CIRST), Université du Québec à Montréal, Canada, vallieres.francois@uqam.ca

Les Carr Senior Lecturer, University of Southampton, lac@ecs.soton.ac.uk

Steve Hitchcock Postdoctoral Fellow, University of Southampton, S.Hitchcock@ecs.soton.ac.uk

Yves Gingras Professor, Université du Québec à Montréal, Canada, gingras.yves@uqam.ca

Charles Oppenheim Professor, University of Loughborough, C.Oppenheim@lboro.ac.uk

Heinrich Stamerjohanns Postdoctoral Fellow, University of Oldenburg, stamer@uni-oldenburg.de

Eberhard R. Hilf Professor, University of Oldenburg, hilf@physnet.physik.uni-oldenburg.de

1. Tenopir, C. 'Online Scholarly Journals: How Many?' Library Journal (1 February 2004).
   
   
2. Ulrich's Serials Analysis System
   
   
3. Association of Research Libraries Ranked Lists
   
   
4. Lawrence, S. Free online availability substantially increases a paper's impact. Nature (Web Debates), (2001). (edited version appears in Nature 411, 521 (2001)).
   
   
5. Kurtz, M.J. et al. The NASA Astrophysics Data System: Sociology, Bibliometrics, and Impact. (2003).
   
   
6. Kurtz, M.J. Restrictive access policies cut readership of electronic research journals articles by a factor of two. (2004).
   
   
7. Odlyzko, A. The rapid evolution of scholarly communication. Learned Publishing 15(1), 7-19 (2002).
   
   
8. Brody, T. et al. The effect of open access on citation impact. (2004)
   
   
9. Hitchcock, S. et al. Evaluating Citebase, an open access Web-based citation-ranked search and impact discovery service. (2003).
   
   
10. Harnad, S. et al. Mandated online RAE CVs linked to university eprint archives: Enhancing UK research impact and assessment. (2003).
    
    
11. Key Perspectives Ltd, JISC/OSI Journal Authors Survey Report. (2004).