December 7, 2018
National Institutes of Health
Office of Science Policy
6705 Rockledge Drive, Suite 750
Bethesda, MD 20892
To whom it may concern:
On behalf of the over 3,000 scientists who are members of the Population Association of America (PAA) and the over 40 federally supported population research centers at U.S. based research institutions comprising the Association of Population Centers (APC), we are pleased to respond to the “Request for Information (RFI) on Proposed Provisions for a Draft Data Management and Sharing Policy for NIH Funded or Supported Research (NOT-OD-19-014).”
Population scientists include demographers, economists, and sociologists who conduct research on population trends and the individual, societal, and environmental implications of population change. They rely on discretionary grant support from the National Institutes of Health (NIH) and accurate and timely data from the federal statistical agencies to produce research findings and conduct research training activities. Population scientists also have unique expertise in data collection, dissemination, and archiving strategies. Thus, the draft NIH data management and sharing policy is especially central to our organizations.
In sum, our organizations support the data management and sharing principles expressed by the NIH. We feel strongly that “data should be made as widely and freely available while also safeguarding the privacy of participants and protecting confidential and proprietary data.” To this end, we believe data sharing should be consistent with the FAIR (Findable, Accessible, Interoperable, and Re-usable) data principles. Further, we believe that data collected as part of any NIH award should be shared regardless of the award’s size. We also support the use of centralized archives for long-term dissemination and support, as well as the development of archives to handle analysis and dissemination of restricted data, such as “Data Sharing for Demographic Research” (DSDR) program, which is funded by the National Institute of Child Health and Human Development (NICHD).
To inform further refinements to the policy, we offer several recommendations.
Recommendation #1: Develop policy for sharing and archiving data extracts
We encourage NIH to articulate how data extracts, analysis files, constructed variables, etc… that are derived from primary sources of data (surveys and other sources), and are used in specific analyses and publications, will be shared. There should be a policy and opportunities for these extracts to be archived in appropriate repositories to facilitate future research, including replication studies. In addition, archiving extracts should be consistent with policies of reuse established by primary data collectors.
Recommendation #2: Address management of paradata
NIH should encourage the systematic collection, documentation, and dissemination of paradata—i.e., data about the data collection process. These data can help users better understand and interpret primary data and support survey methods research that benefits future and ongoing data collection activities.
Recommendation #3: Reward data collection and sharing
The RFI focuses largely on compliance and enforcement. We recommend stipulating enhanced incentives to ensure greater compliance. For example, citations benefit data collectors, offering them recognition and reward, but requires establishing new norms about citing data files. The final policy should provide clear citation guidance, including recommendations for how to cite secondary data that are created and shared with the research community. Both primary and secondary data that are eligible for citation should receive an NIH data catalog record analogous to a PMID or PMCID (in addition to be cataloged using DOIs or other persistent identifiers). Data collectors should also be required to provide clear guidance to users who cite their data–especially for complex, multi-part, and long-running surveys.
Recommendation #4: Address costs of data sharing
Sharing data properly and widely incurs costs usually towards the end of a project after the data have been collected and processed, and funds are exhausted. As part of its data management and sharing policy, we encourage NIH to consider options that could help offset costs associated with data sharing. These strategies could include: holding a fraction of funds in “escrow” for release at end of project for data sharing; award separate supplements to cover data sharing costs; and/or separate data sharing and archiving grants similar to an R03 program (PAR-16-149) that the NICHD has successfully implemented.
Recommendation #5: Timing
Ideally, a timeline for data sharing should be identified, and should ideally occur before the end of the grant—though this is not always possible. We encourage NIH to implement its new data management and sharing policy quickly and efficiently rather than slowly phasing in the policy.
Thank you for considering our recommendations as you develop a data management and sharing policy for NIH funded or supported research. We are pleased to offer our organizations as resources as the agency develops it final policy.
Sincerely,
Wendy Manning, Ph.D., President, Population Association of America
Steven Ruggles, Ph.D., President, Association of Population Centers
Click here to download a PDF of the letter.
#latest-news