“Open Access publishing is widely believed to increase the visibility, dissemination and, eventually, the citation and impact of research findings.” With this statement, I summarize the acclaimed benefits of OA publishing in a recent letter to the Editor of Nature (Nature 458, 967; DOI).
Philip M. Davis, of the Department of Communication, Cornell University, Ithaca, reacts:
“Last year, we reported in the first randomized controlled trial of Open Access publishing1 that OA may increase article downloads, but doesn’t appear to increase citations. The study, done when articles were 9-12 months old has been replicated in many other science journals, and the results still hold now that all articles are 2 years old.”
Davis and colleagues illustrate that not all acclaimed benefits to OA publishing are true. OA may increase access (downloads and number of unique visitors), but not necessarily citation.
While there are many possible reasons for this, I want to highlight one that may be related to the wide availability of online scientific databases and even online reference managers:
You don’t necessarily need full text access to cite subscription papers.
It is perfectly possible to cite directly from the abstract without accessing the original publication. Abstracts are often available for free, and can be seen as the open access part of subscription articles. If citing directly from the web or databases is something that is occurring widely, this may be a particular reason why OA articles are not cited more than subscription articles.
1 BMJ 2008;337:a568 (DOI)
Raf,
Thank you for the reference to our paper. While we mention the fact that a citation can be made without viewing the full scientific article — as one may cite the abstract or merely copy a reference — this behavior cannot explain why we were unable to detect a citation advantage for freely-available scientific articles.
If your citing-the-abstract explanation were correct, then there would be absolutely no evidence for an Open Access citation advantage.
The strength in our study was that we were the first to conduct a true experiment to test the citation advantage hypothesis. All previous studies employed observational methodologies (many even without statistical controls). As a result, they may have been unable to control for possible confounding variables and may have gotten the directionality of the causal relationship backwards.
Does free access to scientific articles lead to more citations?
More likely, higher quality articles are more likely to be made freely available.
Sure it cannot explain everything – it may be a contributing factor. Maybe OA has a bad reputation. Maybe good papers are more frequently published in traditional journals. Maybe OA advocates tend to waste too much time on the web discussing OA and publish less interesting papers. Maybe authors may want to cite strategically and cite more papers in Nature (et al.) as reviewers and editors may like this. OA is a great movement, but surely not meant to replace traditional publishing. Higher quality articles are more likely to get more citations, irrespective of their availability. What I suggest in my correspondence, is that authors should be encouraged to upload their manuscripts on public archive servers such as PubMed Central or institutional repositories such as our own Lirias since most publishers grant that right to authors. It makes the discussion on OA vs. subscription, in fact, a bit redundant – virtually every article can be OA if only more authors take up their responsibility to make their articles available for free.
I remember the discussion after the BMJ paper was published, e.g. the comment thread at the BMJ. The topic is very political, which makes it difficult to openly and calmly discuss this.
One other explanation why there was no citation advantage: many authors that cite other papers work at institutions with institutional subscriptions to journals.
Or, Martin and others, have access via the WHO programmes AGORA, HINARI and OARE.
If you want to read a lot more on this, there is a good annotated bibliography on the topic.