NSF Encourages Open Science Principles

Several people in leadership positions for the NSF’s research directorates and the Offices of Integrative Activities and International Science & Engineering recently came together to write the latest Dear Colleague Letter, Effective Practices for Data. This letter was written to describe effective research data management practices and encourage NSF-funded researchers to implement them.

NSF is specifically encouraging the researchers that it funds to use persistent identifiers for their data, and create data management plans (DMPs) that are machine-readable. These two practices respond to the trend toward open science principles. Open science principles pave the way for more publicly available data, publications, and other research materials. Greater access to research data and outputs allows for new knowledge to be more widely circulated and perhaps built upon by other researchers and scholars, increasing further discovery.

Encouraging researchers to use persistent identifiers for their data and creating DMPs that are machine-readable is a development on their requirement that any researcher applying for funding submit a DMP, outlining how they will manage their data and any resulting publications throughout the course of the research project.

Assigning a persistent identifier to data makes the data more easily discoverable by others, provides information about the dataset, and helps data citations be more persistent and unchanging in publications or other forms of sharing. Persistent identifiers are globally unique and resolvable; they allow for linking, such as within publications or research presentations. DOIs (Digital Object Identifiers) are a common persistent identifier, and global information trackers, like Scholix, use persistent identifiers in publications or citations to increase data and publication sharing.

To implement this practice, researchers should include the citation to the dataset in the body of the article, and list it in the reference list.

To make DMPs machine-readable, NSF recommends using online tools such as DMPTool. Creating machine-readable DMPs allows a computer program to interpret the content of the DMP; for example, it could indicate the need to prepare a data repository for the deposit of a large or complex dataset. Using online DMP-writing tools allows researchers to generate both a PDF version that they would submit to the funding agency when applying for a grant, as well as a machine-readable version to share with their chosen repository or their home institution. UW-Madison has an institutional membership with DMPTool, giving all students, faculty, and researchers access to it, and we encourage UW-affiliated researchers to use it to write DMPs and request feedback from RDS consultants about their DMPs.

For detailed guidance on writing a successful DMP and using DMPTool to create DMPs, refer to our DMP resources. And, if you have any questions about implementing these practices, don’t hesitate to contact us.