Focusing on the 'citation advantage' of open access misses its value in getting science information in the hands of those who need it.
Ever since the open access (OA) movement was launched in the 1990s, support in the scientific community for the principle OA represents that all scientific publications should be made freely available, at least in electronic form has outstripped individual scientists' willingness to put that principle into practice.
The most important reason, according to a recent study commissioned by the European Union, is money. Publishers of OA journals, deprived of income from subscribers, require fees from authors to cover publication costs. Although many funding agencies now provide for these fees in their research grants, not all do and authors' fees are a particular obstacle for scientists in developing countries.
The study found that a second important disincentive to publishing in OA journals is that most scientists still prefer, where possible, to publish in established journals with high citation rates a proxy for quality of scientific publications when it comes to promotion, for example, or measuring how a scientist has contributed to their institution's reputation.
Many such journals, including some of the most prestigious, are owned by commercial publishers with little incentive to give away access to their content for free.
Citation rates are still the 'gold standard' for evaluating the quality of research, and many scientists say there are too few highly cited, or high quality, OA journals in their field. That is despite some of the better known OA journals, such as those published by the Public Library of Science (PLoS), taking steps to ensure high scientific standards.
The overall result according to calculations by Stevan Harnad, one of the closest observers of the OA movement is that even by 2020, only about one quarter of scientific articles will be freely accessible.
'Citation advantage' questioned
A study published last month by Philip Davis, of Cornell University in the United States, has been widely interpreted as throwing OA into further doubt, by questioning what is generally perceived as a major benefit of OA publishing the 'OA citation advantage'.
Some publishers claim that publishing in an OA journal is itself sufficient to increase the number of times an article is cited, purely because it will become more widely available within the scientific community.
Davis' new research, an analysis of articles published in 36 journals over three years, concluded that free access articles were no more likely, or less likely, to be cited by other researchers than those appearing in conventional journals.
Is this a fatal flaw in the arguments for supporting OA, as the press release announcing the conclusions appears to suggest?
Far from it. The debate these results has triggered sidesteps consideration of the full value of OA journals.
This lies not merely in how they benefit science specialists, but also in making scientific research widely available to those who can neither afford high subscription rates for specialist journals, nor get access to scientific libraries but whose work or personal interest depend on having access to the global pool of scientific knowledge.
Those who benefit from OA include many scientists in the developing world, where most university and research institution libraries remain heavily underfunded.
Then there are students, who are equally keen to follow new scientific developments. And finally there are all those who put scientific research to practical use including members of the public, as well as professional groups such as healthcare workers.
As Davis has said, there are many benefits to the free access of scientific information a point long argued by OA advocates, even if a citation advantage may not prove to be one of them.
Leslie Chan, and his colleagues at the Electronic Publishing Trust, expressed similar sentiments in a recent article in PLOS Medicine, where they argue that the sharing of knowledge discovery across borders and the building of a global knowledge commons is increasingly important for solving problems that we all face.
Measuring real value
Putting this social value of science into measurable terms is much more difficult that the relatively simple calculations of citation rates. It is a big challenge and one reason why citations continue to play such a dominant role in assessing individual scientists' or their institutions' achievements.
And as Chan et al. point out, standards for assessing journal quality and relevance are generally based on Northern values that often ignore development needs and marginalise local scholarship.
Until we can find robust ways to measure the success of integrating development goals into the research programmes of developing countries, and of how local scientists contribute to achieving these, the potential of OA is unlikely to be fully recognised.
Even if the evidence for a citation advantage is as weak as Davis' research suggests, this does not undermine the case for OA publishing or even for the less ambitious route to free access represented by creating Open Repositories (so-called 'green OA') controlled by scientists themselves and their institutions.
What the evidence does do is reinforce the folly of using citation rates alone as the main measure of the social value of science and the work of scientists.