Wide access to scientific research results and to data supporting these results is an important tool for accelerating scientific progress. To this end, the National Academy of Sciences (NAS) recently launched a study on how to move “toward an open science enterprise.” Given NIH’s role in promoting the first federal public access policy and creating tools such as PubMed Central for promoting free access to the journal literature, this topic will be of great interest to the NIH Data Science community.
An ad hoc committee under the NAS Board on Research Data and Information (BRDI) will lead this 18-month study exploring the challenges of broadening access to the results of scientific research, described as “open science.”
Open science is defined, for the purposes of this study, as public access (i.e., no charge for access beyond the cost of an internet connection) to scholarly articles resulting from research projects, the data supporting the results contained in those articles, computer code, algorithms, and other digital products of publicly funded scientific research, so the products of this research are findable, accessible, interoperable, and reusable (FAIR) with limited exceptions for privacy, proprietary business claims, and national security.
The committee’s efforts will culminate in a report with findings and recommendations for the scientific community and research enterprise.
The ad hoc committee is chaired by Dr. Alexa McCray, co-director of the Center for Biomedical Informatics at Harvard Medical School, and most of its members are academic researchers. The John and Laura Arnold Foundation is sponsoring the study.
A kickoff meeting was held on July 20, 2017 at the NAS headquarters in Washington, DC. Invited speakers discussed issues for the committee to consider, including scientific norms, incentives/disincentives to share information, differences between scientific disciplines, and trends in open science.
Heather Joseph, Executive Director of the Scholarly Publishing and Academic Resources Coalition (SPARC), provided an historical view, reminding attendees of “first principles” from founding documents such as the Budapest Open Access Initiative and the Berlin Declaration on Open Access. “Open” is not objectively better than “closed,” she observed, but “open” can enable improvements to the utility and usability of knowledge, accelerating research, and sharing knowledge.
Dr. Brian Nosek, Executive Director of the Center for Open Science at the University of Virginia, talked about the drivers necessary to achieve open science. These include the technology to enable change, the training to enact change, the incentives to embrace open access, and the meta science to evaluate it. He emphasized change would involve all facets of the ecosystem, including universities, publishers, funders, and societies.
Speaking about enhancing reproducibility as an important aspect of open science, Dr. Victoria Stodden, of the University of Illinois at Urbana-Champaign, noted there are many types of reproducibility, including empirical reproducibility, statistical reproducibility, and computational reproducibility. To collect appropriate data to achieve reproducible science, she said we must first agree upon the kind of reproducibility being envisioned.
Dr. James Kurose, Assistant Director of the National Science Foundation (NSF) for the Computer and Information Science and Engineering, shared some observations from the NSF experience requiring open science of their investigators. He emphasized the heterogeneity of the research data collected and the need for agency guiding documents to be written in a way that recognizes the variety of data to which those guidelines might apply. For example, in crafting guidance for data management plans, there is an inherent tension between being specific enough to accommodate particular data types and yet broad enough to have general applicability. He also mentioned that an inter-agency public access working group, led by NSF and NIH, is discussing various open science issues relevant to the federal government.
Dr. Marcia McNutt, President of the NAS, provided an insightful summary of how far notions of sharing have evolved over the past three decades. As an example, she pointed toward increased acceptance of pre-prints in the biological sciences. Previously, these documents—draft manuscripts made available prior to formal peer review—were primarily uitilized in the physical sciences. Now NIH allows grant applicants to cite pre-prints and acknowledges their role in speeding the dissemination of knowledge. (See NIH Guide NOT-OD-17-050.) Dr. McNutt also noted top journals can significantly influence researchers to share data (and not hoard their data) by tying publication to making that data open and available.
The next public meeting for this NAS study on moving “Toward an Open Scientific Enterprise” will be held September 18, 2017. Like the NIH Open Data Science Symposium (held December 2016), where over 1,200 people attended from across various sectors, this NAS study signals interest in open science is growing and holds much promise for accelerating the pace of research. We look forward to following the study’s progress and will keep you posted on further developments in open science at NIH.