Writing Data Management Plans for Grant Proposals

What Is a Data Management Plan?

A data management plan (DMP) describes how you will collect, organize, store, share, and preserve the data generated by your research project. Most federal funding agencies now require a DMP as part of the grant application, and many foundations are following suit.

The DMP serves two purposes. First, it demonstrates to reviewers that you have thought carefully about data stewardship — an indicator of overall scientific rigor. Second, it fulfills the funder's commitment to open science and public access. Taxpayer-funded research is increasingly expected to produce publicly accessible data, and your DMP is the plan for making that happen.

Core Elements of a DMP

While specific requirements vary by agency, most data management plans address the same fundamental questions.

Types of Data

Describe the data your project will generate. This includes experimental measurements, survey responses, observational records, images, models, software, code, and any other digital research products. Be specific about formats, expected volume, and the relationship between different data types.

If your project involves human subjects data, specify what data will be collected and how you will handle sensitive information. Genomic data, clinical records, and survey responses all require different management approaches.

Data Standards and Formats

Identify the file formats, naming conventions, and metadata standards you will use. Whenever possible, use widely accepted, non-proprietary formats that ensure long-term accessibility. For example, CSV files are more broadly accessible than proprietary spreadsheet formats, and TIFF images are preferable to format-specific files for archival purposes.

Describe the metadata you will attach to your data — the information that allows others to understand and reuse what you collected. Metadata standards exist for many disciplines. Using an established standard rather than inventing your own makes your data more discoverable and interoperable.

Storage and Backup

Explain where your data will be stored during the project and how it will be protected. Describe your backup strategy, including frequency of backups, storage locations (local and off-site), and redundancy. For sensitive data, describe the security measures you will implement — encryption, access controls, and physical security of storage media.

Most institutions provide research data storage infrastructure. Reference your institutional resources and explain how they meet the needs of your project.

Data Sharing

Describe how and when your data will be made available to others. Specify the repository where data will be deposited, any embargo periods, and the terms of access. If your data will be shared through a discipline-specific repository (such as GenBank for genomic data, ICPSR for social science data, or Dryad for general research data), name the repository and describe its access policies.

If portions of your data cannot be shared — due to privacy protections, proprietary concerns, or other restrictions — explain the reasons clearly and describe what steps you will take to maximize sharing within those constraints.

Data Preservation

Describe your plan for long-term preservation of the data after the project ends. How long will data be maintained? Where will it be archived? Many repositories provide long-term preservation services, but you should specify which repository you will use and confirm that it meets the funder's requirements for retention period.

Federal agencies typically require that data be preserved for a minimum number of years after the end of the grant — often three to ten years, depending on the agency and data type.

Agency-Specific Requirements

NIH Data Management and Sharing Policy

NIH's Data Management and Sharing Policy requires all NIH-funded research to include a DMS Plan. The plan must describe the data to be managed and shared, related tools and software, standards, data preservation and access, and oversight. NIH expects that scientific data be shared at the time of publication or by the end of the award, whichever comes first.

NIH also requires that costs associated with data management and sharing be included in the budget. This means you can — and should — budget for data curation, repository fees, and staff time associated with data sharing.

NSF Data Management Plan

NSF requires a two-page DMP as a supplementary document in all proposals. The plan must describe the types of data, standards for format and metadata, policies for access and sharing, plans for archiving and preservation, and policies for re-use and redistribution. NSF reviews the DMP as part of the merit review process, and a weak plan can count against you.

Other Agencies

DoE, NOAA, USDA, and other federal agencies have their own DMP requirements. Check the specific solicitation for required elements and page limits. When in doubt, consult the agency's data management guidance documents, which are typically available on the agency website.

Writing Tips

Be specific, not generic. A DMP that could apply to any research project is not useful. Describe your particular data, your particular methods, and your particular plan. Generic statements like "data will be stored securely" do not demonstrate thoughtful planning.

Consult your institutional resources. Most research universities have data management specialists in their libraries or research offices who can help you develop your DMP. They know the current agency requirements and can recommend appropriate repositories and standards for your discipline.

Budget for data management. Data curation, quality control, metadata creation, and repository fees all cost money. Include these costs in your budget and justify them in your budget narrative. Reviewers view funded data management as a sign of a serious, well-planned project.

Plan for the full data lifecycle. Think beyond the active research phase. What happens to your data when the project ends, when you change institutions, or when a graduate student graduates? A complete DMP addresses the entire lifecycle of the data.

Common Pitfalls

• Treating the DMP as a checkbox exercise rather than a genuine plan
• Not specifying a repository for data deposit
• Failing to address human subjects data protections when applicable
• Ignoring costs associated with data management in the budget
• Using proprietary formats that may not be readable in the future

A strong data management plan reflects the same rigor and forethought that characterize a strong research proposal. It assures reviewers that the data your project generates will be responsibly managed, broadly accessible, and preserved for future use.