Research Data Management (RDM) describes the active organization and maintenance of data throughout its entire lifecycle. This is process is usually formalized in a data management plan (DMP).
A Data Management Plan is built around the particular needs of each project, but there are concepts and considerations in building the plan that are similar across projects. Typically, sections of a DMP include, but are not limited to: collecting, storing, documenting, preserving, and sharing data -- with additional considerations paid to data handling, responsibilities and resources, and ethics and legal compliance.
Learn more about DMPs here.
Research data encompasses the materials used as evidence to support your scholarship and validate research findings. It includes both digital and non-digital materials.
Research data comes in many forms:
- Observational: e.g., Sensor readings, survey results.
- Experimental: e.g., Gene sequences, chromatograms.
- Simulational: e.g., Climate models, economic models.
- Derived: e.g., Text mining corpora, 3D models.
- Humanities/Creative: e.g., Interview transcripts, field notes, musical scores.
Not every file on your computer requires preservation. At bottom, we should ask which files act as evidence for our research conclusions. Administrative records, for example, support the project’s logistics but do not act as evidence for your research conclusions. These usually do not need to be preserved in a research repository. For more, see the Planning page.
The following benefits are associated with effective research data management.
You will be able to:
- Meet funding agency requirements.
- Write more competitive grant applications.
- Get explicit credit for your data and increase its impact and visibility.
- Encourage the discovery and use of your data to explore new research questions.
- Improve your data's accuracy, completeness, and usability.
- Ensure long-term preservation of data for future researchers.
- Comply with ethics and privacy policies.
Beyond satisfying grant requirements (such as the Tri-Agency RDM Policy), good management practices align your work with FAIR Principles.
"FAIR" does not automatically mean "Open"; it means that data is managed professionally such that it is:
- Effective RDM prevents data loss and reduces the timesink of searching for data.
- Making research data available helps to address the crisis of reproducibility.
- Sharing research data is associated with higher citation rates.
Watch this "Data Management Horror Story" by NYU Health Sciences Library.
Click on any phase below to view more information.
Determine resources, storage costs, and ethical constraints early.
Establish file naming and metadata protocols before files pile up.
Assess data risk levels, implement the 3-2-1 backup rule, and secure active data.
Select significant data for long-term retention, license it, and deposit it in a repository.
Treat data as a primary research output with proper citations to allow impact tracking.