data-management-protocols

# Data Management Protocols for Qualitative Research A data management plan (DMP) specifies how you **collect, organize, secure, retain, and optionally share** qualitative materials. Funders and IRBs increasingly expect DMPs even when data cannot be fully open. ## DMP components (typical) - Data types (audio, transcripts, field notes, screenshots, photos). - File naming/versioning conventions. - Storage locations, backups, encryption, access roles. - Anonymization/de-identification strategy. - Retention period and secure destruction procedures. - Sharing constraints and derived sharing products (e.g., redacted excerpts). ## File naming conventions Use consistent patterns: `YYYY-MM-DD_siteID_participantPseudonym_interview01_audio.wav`. Avoid real names in filenames. Maintain a **separate encrypted key** linking pseudonyms to identifiers if needed for longitudinal contact (IRB-permitted). ## Storage solutions Prefer institutional encrypted storage over personal laptops. If cloud storage is used, verify **BAA** or equivalent for health data contexts and institutional approval. Keep **3-2-1 backups** where feasible (three copies, two media types, one offsite). ## Anonymization techniques Remove direct identifiers; generalize places and dates when small communities enable re-identification; paraphrase highly distinctive stories in publications when necessary. Document **what was altered** for auditability. ## De-identification procedures Distinguish **de-identified** vs **anonymous** data. Qualitative audio often cannot be truly anonymous without destruction—plan accordingly in consent (what participants agree you may retain/share). ## Retention policies Align with IRB approvals, institutional rules, and legal holds. After retention ends, use **secure wiping** (not simple delete) for sensitive files. ## FAIR principles (adapted for qualitative data) FAIR (Findable, Accessible, Interoperable, Reusable) originated for digital objects; qualitative adaptation emphasizes rich **metadata** (who, where, when, how collected), stable identifiers where sharing occurs, and **ethical reuse conditions**. ## Ethical constraints on sharing Consent may forbid sharing; community harm risk may forbid open archives. Consider **controlled access repositories** or sharing only derived categories with illustrative redacted excerpts. ## DMP templates Use funder-specific templates (NSF, NIH) as shells, then add qualitative specifics: transcript versioning, memo logs, software exports (NVivo/Atlas.ti), and team collaboration rules. ## Team workflows Define: who transcribes; QC process; where “gold” transcripts live; how coding exports are versioned; how Slack/email fragments are purged if they contain identifiers. ## Checklist - [ ] Naming, folder structure, and versioning documented. - [ ] Encryption and access control match sensitivity. - [ ] Anonymization map maintained separately from public files. - [ ] Retention/destruction schedule approved and actionable. - [ ] Sharing options align with consent and community risk. ## References (starting points) - Corti, L., et al. *Managing and sharing research data* — qualitative considerations. - UK Data Service guidance on anonymization of qualitative data. - NIH/NSF data management policy pages (update with current agency language when applying).