42 U.S.C. § 19053 — Research reproducibility and replicability

(a) In general. Consistent with existing federal laws on privacy, intellectual property, and security, the Director must help the public gain access to research products — data, software, and code — developed through Foundation-funded projects. (b) Data management plans. Every proposal for Foundation funding must include a machine-readable plan explaining how the applicant will archive its data, software, and code and keep them publicly accessible. To support this, the Director must: provide training and workshops to teach researchers and students how to write and review strong data management plans; make sure program officers and merit review panels have the training and resources to judge how good a data management plan is; and make sure program officers and panels treat these plans as an essential part of a proposal, where appropriate. (c) Open repositories. The Director must: consult with other federal research agencies, as appropriate, and ask the scientific community for input to build and widely share a set of criteria for trustworthy open repositories that Foundation-funded researchers can use, taking into account the different needs of different fields and the need to protect sensitive information; work with stakeholders to find major gaps in the repositories that already meet those criteria and look at ways to support building new or better ones; make competitive awards to colleges or nonprofits, including groups of them, to build, upgrade, and maintain open data repositories that meet those criteria; work with stakeholders and build on existing models to create one public, web-based place where people can find repositories holding data, software, and code from Foundation-supported projects; work with stakeholders to set up the policies, procedures, and resources needed to make sure, as much as practical, that data behind published findings from Foundation-funded projects gets deposited in a qualifying repository by the time of publication; give researchers incentives to deposit data, software, and code in qualifying repositories; and work with the scientific publishing community and other federal agencies to support voluntary, agreed-upon standards for archiving and sharing data. (d) Research, development, and education. The Director must make competitive awards to colleges or nonprofits, including groups of them, to: fund research and development of open-source, sustainable, easy-to-use tools and infrastructure that support reproducibility across many fields of study; fund research on computational reproducibility, including how reproducible computational results stay as new computing hardware, tools, and methods are developed; and fund education and training for students, faculty, and researchers on computational methods and tools that improve the quality and sharing of data, code, and related information, to produce research that others can reproduce.