The Bill and Melinda Gates Foundation has announced the world’s strongest policy in support of open research and open data. If strictly enforced, it would prevent Gates-funded researchers from publishing in well-known journals such as Nature and Science.

On 20 November, the medical charity, based in Seattle, Washington, announced that from January 2015, researchers it funds must make open their resulting papers and underlying data sets immediately upon publication — and must make that research available for commercial reuse. “We believe that published research resulting from our funding should be promptly and broadly disseminated,” the foundation states. It says it will pay the necessary publication fees (which often amount to thousands of dollars per article).

The foundation is allowing two years’ grace: until 2017, researchers may apply a 12-month delay before their articles and data are made free. At first glance, this suggests that authors may still — for now — publish in journals that do not offer immediate open-access (OA) publishing, such as Science and Nature. These journals permit researchers to archive their peer-reviewed manuscripts elsewhere online, usually after a delay of 6–12 months after publication.

Allowing 1 year’s delay makes the charity’s OA policy similar to those of other medical funders, such as the UK Wellcome Trust or the US National Institutes of Health (NIH). But the charity’s intention to close off this option by 2017 might put pressure on paywalled journals to create an OA publishing route.

However, the Gates Foundation’s policy has a second, more onerous twist that appears to put it directly in conflict with many non-OA journals now, rather than in 2017. Once made open, papers must be published under a licence that legally allows unrestricted reuse — including for commercial purposes. This might include ‘mining’ the text with computer software to draw conclusions and mix it with other work, distributing translations of the text or selling republished versions.  In the parlance of Creative Commons, a non-profit organization based in Mountain View, California, this is the CC-BY licence (where ‘BY’ indicates that credit must be given to the author of the original work).

This demand goes further than any other funding agency has dared. The Wellcome Trust, for example, demands a CC-BY licence when it is paying for a paper’s publication — but does not require it for the archived version of a manuscript published in a paywalled journal. Indeed, many researchers dislike the thought of allowing such liberal reuse of their work, surveys have suggested. But Gates Foundation spokeswoman Amy Enright says that “author-archived articles (even those made available after a 12-month delay) will need to be available after the 12-month period on terms and conditions equivalent to those in a CC-BY licence.”

Most non-OA publishers do not permit authors to apply a CC-BY licence to their archived, open, manuscripts. Nature, for example, states that openly archived manuscripts may not be reused for commercial purposes. So do the American Association for the Advancement of Science, Elsevier and Wiley and many other publishers (in relation to their non-OA journals).

“It’s a major change. It would be major if publishers that didn’t previously use CC-BY start to use it, even for the subset of authors funded by the Gates Foundation. It would be major if publishers that didn’t previously allow immediate or unembargoed OA start to allow it, again even for that subset of authors. And of course it would be major if some publishers refused to publish Gates-funded authors,” says Peter Suber, director of the Office for Scholarly Communication at Harvard University in Cambridge, Massachusetts.

“You could say that Gates-funded authors can’t publish in journals that refuse to use CC-BY. Or you could say that those journals can’t publish Gates-funded authors. It may look like a standoff but I think it’s the start of a negotiation,” Suber adds — noting that when the NIH’s policy was announced in 2008, many publishers did not want to accommodate all its terms, but now all do.

That said, the Gates Foundation does not leave as large a footprint in the research literature as the NIH. It funded only 2,802 research articles in 2012 and 2013, Enright notes; 30% of these were published in OA journals. (Much of the charity’s funding goes to development projects, rather than to research which will be published in journals.)

The Gates Foundation also is not clear on how it will enforce its mandate; many researchers are still resistant to the idea of open data, for instance. (And most OA mandates are not in fact strictly enforced; only recently have the NIH and the Wellcome Trust begun to crack down.) But Enright says that the charity will be tracking what happens and will write to non-compliant researchers if need be. “We believe that the foundation’s Open Access Policy is in alignment with current practice and trends in research funded in the public interest.  Hence, we expect that the policy will be readily understood, adopted and complied with by the researchers we fund,” she says.

More than half of all peer-reviewed research articles published from 2007 to 2012 are now free to download somewhere on the Internet, according to a report produced for the European Commission, published today. That is a step up from the situation last year, when only one year  – 2011 – reached the 50% free mark. But the report also underlines how availability dips in the most recent year, because many papers are only made free after a delay.

“A substantial part of the material openly available is relatively old, or as some would say, outdated,” writes Science-Metrix, a consultancy in Montreal, Canada, who conducted the study, one of a series of reports on open access policies and open data.

The study (which has not been formally peer-reviewed) forms part of the European Commission’s efforts to track the evolution of open access. Science-Metrix uses automated software to search online for hundreds of thousands of papers from the Scopus database.

The company finds that the proportion of new papers published directly in open-access journals reached almost 13% in 2012. The bulk of the Internet’s free papers are available through other means – made open by publishers after a delay, or by authors archiving their manuscripts online. But their proportion of the total seems to have stuck at around 40% for the past few years. That apparent lack of impetus is partly because of a ‘backfilling’ effect, whereby the past is made to look more open as authors upload versions of older paywalled papers into online repositories, the report says. During this last year, for instance, close to 14,000 papers originally published in 1996 were made available for free.

“The fundamental problem highlighted by the Science-Metrix findings is timing,” writes Stevan Harnad, an open-access advocate and cognitive scientist at the University of Quebec in Montreal, Canada. “Over 50% of all articles published between 2007 and 2012 are freely available today. But the trouble is that their percentage in the most critical years, namely, the 1-2 years following publication, is far lower than that. This is partly because of publisher open access embargoes, partly because of author fears and sluggishness, but mostly because not enough strong, effective open access mandates have as yet been adopted by institutions and funders.”

The report’s conclusions are only estimates, as the automated software does not pick up every free paper, and this incompleteness must be adjusted for in the figures (typically adding around 5-6% to the total, a margin calculated by testing the software on a smaller, hand-checked sample of papers). And many of the articles, although free to read, do not meet formal definitions of open access – for example, they do not include details on whether readers can freely reuse the material. Éric Archambault, the founder and president of Science-Metrix, says it is still hard to track different kinds of open manuscripts, and when they became free to read.

The proportion of free papers also differs by country and by subject. Biomedical research (71% estimated free between 2011 and 2013) is far more open than chemistry (39%), for example. The study suggests that from 2008-2013, the world’s average was 54%, with Brazil (76%) and the Netherlands (74%) particularly high. The United Kingdom, where the nation’s main public funder, Research Councils UK, has set a 45% target for 2013-14, has already reached 64% in previous years, the report suggests.

The study comes during Open Access week, which is seeing events around the world promoting the ideas of open access to research. Yesterday saw the launch of the ‘Open Access Button’ in London – a website and app that allows users to find free research. If no free copy is available, the app promises to email authors asking them to upload a free version of their paper – with an explanation direct from the user who needs the manuscript. “We are trying to make open access personal – setting up a conversation between the author and the person who wants access,” says Joe McArthur, who co-founded the project and works at the Right to Research Coalition, an advocacy group in London.

It’s a common complaint among academics: today’s researchers are publishing too much, too fast. But just how fast is the mass of scientific output actually growing?

Many would throw up their hands and declare the question impossible. It’s clearly wrong to cite the growth of academic databases, such as Thomson Reuters Web of Science, which has increased its coverage by around 3% per year (barring occasions when the database incorporates a flood of new journals). That dramatically undercounts the true expansion: no database captures everything.

Bibliometric analysts Lutz Bornmann, at the Max Planck Society in Munich, Germany and Ruediger Mutz, at the Swiss Federal Institute of Technology in Zurich, think they have a better answer. It is impossible to know for sure, but the real rate is closer to 8-9% each year, they argue. That equates to a doubling of global scientific output roughly every nine years.

In a study to be published in the Journal of the Association for Information Science and Technology, and uploaded to the online server arXiv, Bornmann and Mutz find that global scientific output has probably kept up this dizzying rate of increase since the end of World War II. Other researchers say the study seems sound, although it is hedged with caveats.

“We identified three growth phases in the development of science, which each led to growth rates tripling in comparison with the previous phase: from less than 1% up to the middle of the 18th century, to 2 to 3% up to the period between the two world wars and 8 to 9% to 2012,” they write.

Continue reading →