All that glitters is not gold: the fallacy of open access evangelism
All-or-nothing open access evangelism perpetuates the problems of scientific publishing. Writing in the Guardian, one advocate has even suggested that publishing behind a pay-wall is immoral. That form of evangelism is wrong – for now – and may do more harm than good.
Yes, there are clear advantages to gold open access. Chief among these is that everyone gets unfettered access to research. There is no doubt that access is fundamental to the way science works. Yet there is a trend towards a simplistic kind of open access evangelism that seems to be gaining traction in mainstream academia, and it has me worried.
Problem 1. It is keeping costs of doing research unsustainable
In its current form, paid gold open access only shifts the cost of access from the libraries to the researchers without making it affordable. Where researchers and libraries are both publicly funded, money flows from the same founts to the same drains, just through different pipes. For hybrid policies (subscription-based journals with paid open access options), the public can sometimes pay twice.
Traditional publishers appear comfortable with gold and hybrid open access. Elsevier manages approximately two thousand journals and nearly 80% have open access options, costing between USD$500 and USD$5000 per article. The average cost of publishing an article in open access is a little over USD$900, and for biomedical disciplines the costs are far higher than that. There are low- or no-cost gold open access alternatives but they are not as popular and do not confer the same levels of prestige.
If advocates (and working groups in the UK) continue to promote the idea that gold open access is the appropriate direction for right now, publishing will remain too expensive for researchers with already restricted budgets, and the public will continue to fund ridiculously profitable publishing groups and shonky operators.
The obvious short-term solution, often proposed from within the open access community, is to focus on green open access (authors may freely distribute a version of the article), and dramatically improve the rate of self-archiving. Per article, library subscriptions may still be more expensive than open access fees, but a critical mass of self-archiving is a necessary step in the process.
The gap between the number of pay-walled articles that could be made available and the number that actually are, is a shameful indictment of academia. In a soon-to-be-formally-published article, Björk et al. have showed that 81% of published articles could have been self-archived, yet the uptake is around 12%. The blame for the gap is aimed directly at academics.
The civil disobedience of #icanhazpdf (Twitter users requesting and receiving articles from academics with institutional access) is nearly always against institutional policy. No one would recommend the practice, even if tools like TOR and anonymous public dropbox accounts could safely preserve anonymity. But it does provide a hint for what could be done with the 50 million articles already out there. Just as the legitimate use of Spotify replaced the illegal use of Napster for sharing music, we may soon see a tool that will be viewed as both the saviour and destroyer of academic publishing.
Problem 2. It permits the widespread ruination of quality and rigour in research
It is not just the largest publishing houses that are embracing pay-to-publish open access models. There appears to be an endless supply of researchers willing to engage with nearly 9000 open access journals. The number of journals has doubled since 2009, with more than three new journals established every day. While there are paragons of quality among the top tiers of open access journals, many have instead duped researchers and watered down the global research endeavour with misconduct, plagiarism, and inadequate peer-review.
Hindawi, a borderline predatory publisher, is reported to have a higher profit margin (52%) than Elsevier (36%), suggesting that publishing offers exceptional returns on investment, better than mining, pharmaceuticals, perhaps only beaten by illegal drug/human trafficking. Only a small fraction of open access journals have impact factors. Yes, there are gems among the bulk of gold open access journals (I’ve seen some), but many of these journals and the crap published in them only function to fill the research arena with work that adds very little to science, to innovation, and to improving society.
As a consequence of the publish-or-perish mentality and ease of publishing, scientific research has jumped way over the line from a curated library (little redundant information and sustained relevance) to the kind of vast streams of quickly-forgotten information that gives meaning to the phrase “sipping from the fire hose”.
This is where gold open access and creative commons (with attribution) licenses will eventually become vital. If academic publishing does reach a state in which access to all research is free, immediate and unrestricted, it will likely signal the biggest shift in science since the digital revolution. Much of the new research synthesis ideas involving algorithmic filtering, reanalysis, and meta-analysis are yet to be developed. Where I work, some of us have started to wander around the edges of these possibilities.
We may even be able to do away with journals completely. Michael Eisen, co-founder of PLoS, has argued that publishers add very little to the work done by researchers. Mathematicians have moved to take publishers out of the equation with the Episciences Project. In the future, scientists might open up their lab books and hard disks so that data and models can be freely shared, searched, recycled and linked together like vast open source software communities have been doing for years.
To get there, science needs better ways to attribute and praise individuals for discrete chunks of research. This is where altmetrics are expected to extend citation-based metrics to detail the full range of impact that research (not just publications) can have on scholarship and society. Using citations and journal impact factors to find good science is like trying to fish with explosives.
Lessons for researchers and research policy developers
While we wait for the future of research dissemination to emerge, there are simple ways in which academics can make sure they act in the best interests of the scientific community and the public.
Besides respecting alternative measurements of impact, funders should continue to mandate self-archiving through institutional repositories. Informaticians should investigate tools and motivations for sharing in line with the decentralized #icanhazapdf, the NIH manuscript submission system, Figshare, the Synaptic Leap, or the loophole that ResearchGate uses to encourage uploads to ‘personal’ pages.
Australians are relatively lucky when it comes to self-archiving. We have a mandate from National Health and Medical Research Council, and the Australian Research Council, that all publicly-funded research must be shared, and we have seen experts in open access discussing the cost-effectiveness of gold and green open access early and clearly.
As a bare minimum, academics must make their contact details public to entertain requests for inaccessible articles, check journal policies on open access prior to submission, avoid the temptation of predatory open access journals, and most importantly:
dramatically improve the woeful record in self-archiving.
There is no direct route to an academic publishing future where publicly-funded research outputs are both libre and gratis. A sustainable trajectory requires a diversity of affordable ways to disseminate research widely, but it will only work if we can retain our grip on the processes that ensure the rigour and quality being eroded.
[Image credit: The ENIAC]