Public access for public research
Modern science was founded on the free exchange of ideas. In the 17th century, the first scientific journals emerged and began to replace the informal network of correspondences that had previously linked Europe’s scientists. This innovative step allowed for faster and more transparent dissemination of scientific discoveries, and the journal eventually became the gold standard for communicating with the wider scientific community.
But the premier mechanism for documenting and spreading scientific knowledge has failed to live up to the ideals of openness and transparency which made it so promising. Despite the proliferation of peer-reviewed journals, many scientific disciplines are now facing a crisis of reproducibility. A recent survey conducted by Nature found that 70% of scientists have tried and failed to reproduce another scientist’s experiments. 52% of those surveyed agreed that there was a “significant crisis” of reproducibility, while 38% thought that there was a “slight” crisis. Only 3% of researchers thought that there was no crisis.
“As data-intensive science becomes more common, the scientific paper’s ability to fully describe scientific discoveries diminishes.”
It seems that these researchers’ fears are well founded. Last year, Science published the results of its Reproducibility Project. Collaborators attempted to replicate the results of 100 experimental and correlation studies published in three high-ranking psychology journals. Only 39 studies were successfully replicated, and even then, the effects tended to be smaller or weaker than those found in the original study. While the question of reproducibility in science is complex, it’s likely that the structure of the traditional scientific journal contributes significantly to the problem. As data-intensive science becomes more common, the scientific paper’s ability to fully describe scientific discoveries diminishes. Added to this is the fact that the presence of publication bias coupled with the pressure on researchers to publish in traditional “high-impact” journals may skew the literature towards seemingly significant positive results.
“If science is meant to be an open, transparent, and collaborative endeavour, then it’s time we re-examine our approach.”
There are other issues beyond reproducibility which call into the question the current model. The rising costs of journal subscriptions has prompted both Harvard and Cornell to cut major titles from their libraries, and many scientists can’t access the papers they need in order to conduct a given research project. If science is meant to be an open, transparent, and collaborative endeavour, then it’s time we re-examine our approach.
Open Science
There is a growing trend among some researchers toward an “open science” model. Taking inspiration from the open source movement in computer programming, open science takes research out of the walled gardens of traditional journals and intellectual property regimes and makes it available to anyone with a computer. Some open access journals have already seen incredible success. PLOS ONE, a peer-reviewed open access journal founded in 2003, is now the world’s largest scientific journal as measured by number of articles (165,000 to date).
Open science initiatives can also go well beyond simply providing an alternative to traditional subscription based journals. In 2003, a massive collaborative effort between the NIH, the FDA, industry, universities, and NGOs was undertaken in order to determine the biological markers that show the progression of Alzheimer’s disease. Research methods and data formats were standardized, any potential patent applications were disavowed, and every single finding was made immediately available online. The Alzheimer’s Disease Neuroimaging Initiative generated a flood of papers on early diagnosis and new potential treatments. The entire data set has now been downloaded 3,200 times, and the set of brain scan images has been downloaded nearly a million times.
More recently, the Montreal Neurological Institute made the bold decision to become the first research institution in the world to fully embrace open science. The Neuro will make all of its results and data freely accessible, as well as as its software, algorithms, and even tissue samples — it also won’t pursue patents on any of its discoveries. The Neuro hopes that removing artificial barriers will help accelerate research in an area where progress has so far been slow.
Criticisms and the Ethical Argument
However, not everyone in the scientific community is excited about the open science movement. Some critics argue that the pay-to-publish business model of open access journals is not economically sustainable, and will only damage the market as traditional journals experience declining revenues; others believe that the open access model will not generate enough profit to support investment in new technology. There are also concerns over the impact on the careers of young scientists who choose to publish in open access journals, rather than traditional ones.
“We ought to find ways of supporting scientists who strive to make their work accessible to everyone.”
Despite these concerns, there is a compelling ethical argument in favour of open science. Much scientific research in Canada is publicly funded, at least in part. A large portion of journals’ subscription revenue comes from public funds via university libraries, and the experts who conduct peer review free of charge also draw their salaries from the public purse. Why then are the results of these publicly funded efforts locked behind paywalls? And the concern over the impact that eschewing traditional journals may have on the careers of up and coming scientists is not really an argument against open science, but an observation on the difficulty of changing a culture. We ought to find ways of supporting scientists who strive to make their work accessible to everyone.
Perhaps the most significant criticism levelled at open access journals is the claim that they do not have a rigorous peer review process and are thus of low quality. Critics point to episodes such as the spoof paper that was accepted by more than half of the 304 open access journals to which it was submitted. It’s interesting to note, however, that this “sting” only targeted open-access journals — we have no way of knowing how the paper would have fared with traditional journals. What’s more, a different spoof paper with “egregiously bad” science was accepted by Science in 2011.
Conclusion
There are legitimate questions about how to best approach open science. The level of openness which can be pursued may vary across research areas, and shunning patent applications may not work in certain disciplines (such as engineering). But these are questions that can only be answered by pursuing open science initiatives.
Open science shows a great deal of promise, and seems to be truer to the fundamental ideals of science than the system we have now. Initiatives such as the one at the Montreal Neurological Institute show us that there is another model we can follow that may accelerate research innovation in areas that could have real payoffs for society at large.
