Guest Authors: Dylan Roskams-Edris (Open Science Consultant and Invited Scholar with the Center for Genomics and Policy,), JB Poline (Associate Professor, Neurology and Neurosurgery, McGill University), and Nikola Stikov (Associate Professor, Department of Electrical Engineering, Polytechnique Montreal).
First off, what is Plan S and why does it matter?
In September 2018 a group of European funding agencies, along with some private funders, assessed how scientific information is disseminated – namely that university researchers receive public funding, perform experiments, collect data, write articles about the results, review each other’s articles, and then the publicly funded institutions they work for foot the ever increasing bill to buy access to all this work – and decided enough was enough; publicly funded science should be publicly accessible.
They formed cOAlition S and announced Plan S which, boiled down, says that after 2020 any work funded in whole or in part by the members of the coalition must be published in journals that allow immediate, free, and public access to articles. They also published a set of guidelines and best practices that journals publishing funded research must follow. Those interested in the details of Plan S are encouraged to read about them here.
Only 5 months after Plan S was revealed, the University of California system – one of the largest and most well-respected public academic systems in the world – announced that they were ending their relationship with the publisher Elsevier. The split was due primarily to Elsevier’s refusal to make all UC published research open access on the UC’s terms.
These efforts are a great start.
For those actively working to open global research efforts, however, efforts like Plan S are only the beginning. The push for open science is much larger than making scholarly articles open access. Open science demands that the data, software, materials, and know-how that enables science are also shared – or better yet, openly published.
Sharing, Publishing, and Open Access Articles
It is worth pausing here for a brief aside about the distinction between open sharing, open publishing of research resources, and open access publishing of articles. All of these are important but for open science to be successful the distinction between them has to be clear.
Open sharing consists of making research resources available in a way they can be freely accessed and used. Sharing datasets in a repository or data sharing platform like Dryad, or code used for data analysis and visualization via a service like Github, are good examples. Sharing in this way rapidly disseminates resources and makes them available for use and adaptation by others as quickly as possible. Open publishing of research resources, however, involves the filtration of these resources through other researchers. These peer researchers make sure that the shared resource – whether it is data, code, single figures, or any of the plethora of resources developed throughout the scientific process – is in a form that is standard and easily usable by others, as well as presenting those resources in a curated form on a website or repository. Open access publishing of articles is the primary target of efforts like Plan S and relates to publishing scholarly articles in such a way that they are freely accessible and usable.
The Canadian Open Neuroscience Platform (CONP), along with myriad other organizations, are developing the resources needed to enable open sharing, open publishing of research resources, and open publishing of articles. By doing so the CONP is helping open science and reduce the current inequalities in access to all of the tools and research outputs science needs to thrive.
Assessing the situation here in Canada, the current open access policies of Canada’s primary trio of scientific funding agencies take a weaker stance than Plan S and, as we’ve already argued, efforts like Plan S are only a start. Even if Canada were to suddenly marshal its forces of scientific-policy and follow the lead of cOAlition S we would still be well behind where open science should be. Scholarly papers are just the tip of the iceberg of scientific discovery. Any given publication is the result of months of work involving data collection, coming up with hypotheses, planning experiments, carefully recording progress in notebooks, developing software, analyzing results, writing drafts, and working with editors. The iceberg metaphor is a good one; the publication, like the burg’s tip, is what is seen, but floats atop a much larger body of work that supports it.
Open science in its full and rich instantiation is about sharing and publishing all as many scientific resources as possible to enable collaboration throughout the research process. It grows from the simple idea that working together is more effective than working in silos; that many hands make light work and many eyes spot more errors. In order to work together scientists need to be able to freely discover each other’s work and use each other’s resources, not just have access to the polished end result. Doing so will improve replication, catch errors in methodology early on, and prevent redundancy, all while reducing unnecessary barriers to discovery and outlays from the public coffers.
The internet and modern computing technology make this dissemination possible. What we need now is for funders, institutions, and scientists themselves to commit to encouraging and enabling the sharing their data, code, notebooks, experimental protocols, draft manuscripts, and all the other pivotal elements of discovery.
Building on Plan S
Plan S does mention the sharing of scientific data. On their Principles and Implementation page they encourage the sharing of research resources underlying publications in repositories that make the resource “as open as possible and as closed as necessary”, as well as highlighting resource sharing as a key topic of negotiations between journals, funders, and institutions. Doing so, however, requires repositories where data and code that can be stored, data sharing platforms where that data can easily linked to publications, as well as organized curation efforts to make sure the linked research resources are usable by others.
The CONP is building exactly these resources for the neuroscience community. It is collaboratively developing the user portal, online publishing, informational resources, and governance and policy mechanisms that will enable neuroscientists to disseminate each component of their work as openly as possible, and to keep them open. The CONP scientists and stakeholders believe that these resources will catalyze collaboration and accelerate neuroscientific discovery.
What’s Needed Now to Go Further
But more is needed.
Opening science requires the collective effort of funders, data sharing platforms, academic institutions, and individual scientists. Science doesn’t have to be opened all at once, but steps down the open road must be taken, and must be taken now. The CONP will provide tools and guidance, but scientific culture shift requires a concerted community effort.
Some first steps needed to enable the open publishing of all research resources include: (1) forging agreements and partnerships between journals and open science platforms to make it easy for scientists to share their data, publish it in a curated form, and link it to publications, (2) promotion and tenure policies at academic institutions that value the sharing and publishing of data on par with producing articles, (3) funding agencies that require (and enforce) sharing and publishing data, code, and materials associated with publications as a condition of receiving a grant, and (4) a commitment from scientists themselves to change the culture of science towards openly sharing and publishing as many of their resources as they can.