The National Science Foundation’s Announcement
On May 10, 2010, the National Science Foundation (NSF) disclosed a new development in its data sharing policy: the agency announced that, by fall 2010, it intended to require data management plans (DMPs) as part of all proposals responding to NSF grant funding solicitations. Chapter II of the current Grant Proposal Guide (NSF 11-1 January 2011), under “Special Information and Supplementary Documentation,” provides some detail on what a data management plan should encompass.
The Context: Providing Access to Federally Funded Research Data
Like many federal agencies that provide research funding, the NSF has had a data sharing policy in place for many years, in compliance with the Office of Management and Budget’s (OMB) Circular A-110. Approved in 1999 and enacted the following year, the OMB’s Circular A-110 requires that data generated from federally funded research be made publicly available via the process instituted by the Freedom of Information Act. A more recent development is the 2009 Open Government Initiative, charging federal agencies to take a committed stance toward effecting transparency, participation, and collaboration in their activities.
For many years the NSF has long expected investigators to publish, with proper authorship attribution, significant research findings and to share – within an acceptable period of time and at moderate expense – data sets, samples, collections, and other research-related materials generated by funded projects. In the last five years the agency has made further inroads in this area. For example, its solicitation for the Sustainable Data Preservation Access and Network Partners (DataNet), launched in 2007, marks a continuing effort to address the challenges of keeping digital data accessible, usable, and reusable.
The Gist of the New Requirement
As the NSF states in its current Grant Proposal Guide, data management plans are intended to document how research data will be described, accessed, archived, shared, re-used, and re-distributed over the length of the funded project and beyond. The guidance provided by the NSF is general in nature, since specifics for the plan will depend on what the agency refers to as the “community of interest” – essentially the domain or discipline to which the project proposal is relevant. Not to exceed two pages, data management plans are to be part of the supplementary materials in NSF grant proposals. Basic components to such a plan might be as follows:
Types of data, samples, physical collections, software, curriculum materials, and other materials that are generated for the duration of the project
Standards for data and metadata format and content
Access and sharing policies, with stipulations, as needed, for privacy, confidentiality, security, intellectual property, or other rights or requirements
Policies for re-use, re-distribution, and creation of derivatives
Plans for archiving data, samples, and other research outcomes, as well as for maintaining their access
The list above represents a suggestion. A DMP may include all of the above, or some of it.
The NSF also notes that a DMP may consist of a statement that no plan is necessary, as long as “clear justification” for such an omission supports the statement.
Certain directorates within the NSF, however, provide explicit guidelines and advice on forming DMPs. See our Resources on Data Management Planning.
Additional guidance may be found at “Data Management & Frequently Asked Questions (FAQs)” on the NSF website.
Why the Change to NSF’s Policy?
With the May 2010 announcement, the NSF essentially modified its policy implementation for data sharing. Various factors informed this decision, including the increasingly data-centric, collaborative nature of science research and, by extension, the need for more openness and communication about research findings. It also signals an explicit recognition of the realities of research in the “Digital Age.” As assistant director for the NSF’s Computer & Information Science & Engineering directorate, Jeannette Wing, stated in the May 2010 press release, “Digital data are both the products of research and the foundation for new scientific insights and discoveries that drive innovation.” Most important, the agency’s new requirement bespeaks its design for a more encompassing, programmatic take on data sharing and data curation.
Authors: Patricia Hswe and Ann Holt