Basic Best Practices for Documenting Research Data
Metadata standards and best practices for documenting research data can vary by discipline. Refer to the Tools & Resources section to find discipline-specific metadata standards. Below is a list of general best practices which apply to almost all research datasets. For assistance in choosing a metadata standard or identifying a strategy to document your research data, contact Matt Carruthers, Metadata Projects Librarian, University of Michigan Library.
Important considerations when documenting your research data
Below are some things to keep in mind when documenting your research data to help others understand and potentially reuse the datasets.
- Remember to document any coding schemes or variables used in the data.
- Remember to document any conventions used for naming files in the datasets as well as how the files were organized in the research project.
- Remember to document how the data was collected or generated, including any processes, software, or equipment used.
Basic elements to document about your data
- Title - The name of the research project or dataset.
- Creator - Names of people who created the dataset, including organizational affiliation.
- Dates - The date range during which the data was collected, processed, and/or modified, as well as the dates to which the data pertain.
- Methodology - The process by which the data was created or captured, including any code, software, equipment, or protocols used.
- Subjects or Keywords - Words or phrases which describe the type or content of the data, the location in which it was collected, as well as the discipline or domain to which it pertains.
- Funders - The agencies or organizations which funded the research that produced the dataset.
- Rights Statement - Information regarding conditions governing access to or use of the data, as well as who holds the intellectual property rights for the data.
- Unique Identifiers - Any name, number, or alpha-numeric text string used to uniquely identify the project or dataset, including grant numbers or internal reference numbers.
Introductory Readings and Training
- Best Practices for Metadata for social science data by the Inter-university Consortium for Political and Social Research
- Guide to writing "readme" style metadata by the Cornell Research Data Management Service Group
- Understanding Metadata by the National Information Standards Organization
- Metadata Basics by the Dublin Core Metadata Initiative
- Metadata Best Practices by DataONE
- MANTRA Research Data Management Training module on metadata, documentation, and citation
- Practical guidance for anyone working with research data by the UK Data Service
- Create & Manage Data: Documenting Your Data by the UK Data Archive
- Metadata: Working Level by the Australian National Data Service
