Working School & college skill
Manage research data
Manage research data across collection, naming, documentation, quality, access, security, sharing, retention, and disposal.
Govern raw evidence and transformations from the first record.
When to use
- Use for qualitative, quantitative, laboratory, field, archival, or computational research data.
- Do not move restricted, personal, Indigenous, export-controlled, licensed, or security-sensitive data into unapproved tools.
Procedure
- Inventory data types, sources, formats, volumes, sensitivity, ownership, licenses, consent limits, and retention duties.
- Define stable identifiers, directory structure, naming, versioning, metadata, codebooks, and provenance.
- Separate immutable raw data from cleaned, derived, analytic, and published outputs.
- Record collection, import, transformation, exclusion, correction, and quality-control steps.
- Assign role-based access, authentication, encryption, transfer, backup, recovery, and incident procedures.
- Define de-identification or pseudonymization and keep re-identification keys separately controlled.
- Plan reproducible environments, code versions, dependencies, and checksums.
- Specify what can be shared, with whom, under which agreement, at what disclosure risk, and in which repository.
- Test restore, access revocation, metadata completeness, and a representative reproduction.
- Execute retention review and verified disposal only with authority.
Worked example
A team stores interview audio and coded excerpts in one shared folder. The plan moves raw audio to restricted storage, separates the identity key, versions the codebook, logs transformations, and creates a de-identified analytic dataset whose disclosure risk is reviewed before sharing.
--- name: manage-research-data category: school description: Manage research data across collection, naming, documentation, quality, access, security, sharing, retention, and disposal. Use when planning or operating a study's data lifecycle. --- # manage-research-data Govern raw evidence and transformations from the first record. ## When to use - Use for qualitative, quantitative, laboratory, field, archival, or computational research data. - Do not move restricted, personal, Indigenous, export-controlled, licensed, or security-sensitive data into unapproved tools. ## Procedure 1. Inventory data types, sources, formats, volumes, sensitivity, ownership, licenses, consent limits, and retention duties. 2. Define stable identifiers, directory structure, naming, versioning, metadata, codebooks, and provenance. 3. Separate immutable raw data from cleaned, derived, analytic, and published outputs. 4. Record collection, import, transformation, exclusion, correction, and quality-control steps. 5. Assign role-based access, authentication, encryption, transfer, backup, recovery, and incident procedures. 6. Define de-identification or pseudonymization and keep re-identification keys separately controlled. 7. Plan reproducible environments, code versions, dependencies, and checksums. 8. Specify what can be shared, with whom, under which agreement, at what disclosure risk, and in which repository. 9. Test restore, access revocation, metadata completeness, and a representative reproduction. 10. Execute retention review and verified disposal only with authority. ## Worked example A team stores interview audio and coded excerpts in one shared folder. The plan moves raw audio to restricted storage, separates the identity key, versions the codebook, logs transformations, and creates a de-identified analytic dataset whose disclosure risk is reviewed before sharing. ## Done - A data-management plan records inventory, provenance, identifiers, versions, quality controls, access, sharing, retention, and disposal - Backup restore, permission, checksum, metadata, and representative reproduction tests are completed and recorded