Retention Timeline for Recordings and Transcripts: IRB-Friendly Plan (Template)

An IRB-friendly retention timeline is a simple plan that says what research files you will keep (recordings, transcripts, consent forms, analysis files), how long you will keep them, who can access them, and how you will securely delete them. The safest approach is to separate timelines for identifiable vs. de-identified materials and to document every destruction step in an audit trail. This guide gives you a practical fill-in template you can paste into your IRB protocol or data management plan.

Primary keyword: retention timeline for recordings and transcripts

Key takeaways

• Split retention by file type and identifiability: audio/video, identifiable transcripts, de-identified transcripts, consent forms, and analysis files.
• Choose timelines that match your purpose: verification, publication, reanalysis, and sponsor or institutional requirements.
• Plan for exceptions with a legal hold rule so you can pause destruction when needed.
• Create an audit trail that records what you destroyed, when, how, and by whom.

What an IRB typically wants from a retention plan

Most IRBs focus on risk, access, and clarity. They want to know how long identifiable data will exist, who can see it, and how you will reduce exposure over time.

A strong plan answers these questions in plain language.

• What will you keep (and in what format).
• Where you will store it (approved systems, encryption, backups).
• Who can access it (roles, least-privilege access).
• When you will de-identify and destroy each item.
• How you will document destruction and handle legal holds.

If your study uses audio or video, the retention timeline matters because recordings can carry biometric identifiers (voice, face). They also tend to live in more places than you expect, such as transcription vendor portals, local downloads, and cloud backups.

Common retention rationales (and what they imply for timelines)

Retention is not “keep everything forever.” It is a reasoned choice based on what you need the data for, balanced against privacy risk.

Rationale 1: Data verification and quality control

Teams often keep recordings long enough to check transcript accuracy, resolve unclear sections, and confirm coding decisions. This usually supports a short retention window for raw audio once the transcript is verified.

• Keep audio through transcription + QA + spot checks.
• Keep identifiable transcripts only while you still need to remove identifiers and verify meanings.
• Keep de-identified transcripts longer for coding and analysis.

Rationale 2: Ongoing analysis and publication

If you will publish findings or write a dissertation, you may need to retain analysis files and de-identified transcripts long enough to support writing, peer review, and any corrections.

• Define what “analysis support” means (codebook, memos, coded excerpts).
• Prefer retaining de-identified data rather than identifiable source files.

Rationale 3: Future research use or data sharing

If you plan to reuse data later, your consent language and IRB approval should match that plan. Future use usually pushes you toward retaining de-identified transcripts and documentation (like codebooks) longer, while still limiting identifiable materials.

• Confirm whether consent allows reuse or sharing.
• Write a clear boundary: what stays internal, what could be shared, and in what form.

Rationale 4: Sponsor, institutional, or funder requirements

Some studies have required retention periods set by a sponsor, university policy, or contract. In that case, your plan should cite that requirement and align your timelines to it.

If you need to reference a rule, point to the exact policy used by your institution or sponsor rather than guessing a number.

Fill-in retention timeline template (IRB-friendly)

Use the template below as-is, or customize it. Keep it consistent with your consent form, protocol, and any vendor agreements.

Project overview

• Study title: [Fill in]
• IRB protocol number: [Fill in]
• Data steward (role/name): [Fill in]
• Approved storage locations: [Fill in: e.g., encrypted institutional drive, approved cloud]
• Access roles: [Fill in: PI, Co-I, RA; least-privilege]
• De-identification point: [Fill in: after transcription QA, before coding, etc.]

Retention schedule (separate timelines)

Fill in a timeline for each file type. If you do not keep an item, write “Not retained” and explain why.

• 1) Audio/Video recordings (raw)
  • Contains identifiers? Yes (voice/face) / No
  • Where stored: [Fill in]
  • Who can access: [Fill in]
  • Retention start: Date of recording / Date of upload / Other: [Fill in]
  • Retention end (destroy by): [Fill in: e.g., after transcript verification + within X days]
  • Reason for keeping: [Fill in: transcription accuracy checks, dispute resolution]
  • Destruction method: [Fill in: secure deletion method per institutional IT guidance]
  • Backup handling: [Fill in: include backups, synced folders, vendor portals]
• 2) Identifiable transcripts (names/places/employers, etc.)
  • Where stored: [Fill in]
  • Who can access: [Fill in]
  • Retention start: Date transcript received / Other: [Fill in]
  • Retention end (destroy by): [Fill in: e.g., after de-identification + QA + within X days]
  • Reason for keeping: [Fill in: create a clean de-identified version, confirm context]
  • De-identification process: [Fill in: remove direct identifiers, generalize indirect identifiers]
  • Destruction method: [Fill in]
• 3) De-identified transcripts (working dataset)
  • Where stored: [Fill in]
  • Who can access: [Fill in]
  • Retention start: Date de-identified version created / Other: [Fill in]
  • Retention end (destroy by): [Fill in: end of study + X years/months, or per sponsor policy]
  • Reason for keeping: [Fill in: analysis, publication, auditability]
  • Sharing plan (if any): [Fill in: none / internal only / controlled access]
  • Destruction method: [Fill in]
• 4) Consent forms and recruitment materials
  • Format: Paper / Digital / Both
  • Where stored: [Fill in: locked cabinet, encrypted drive]
  • Who can access: [Fill in]
  • Retention start: Date signed / Other: [Fill in]
  • Retention end (destroy by): [Fill in: per institutional policy or sponsor requirement]
  • Reason for keeping: [Fill in: documentation of consent, compliance]
  • Destruction method: [Fill in: shredding for paper, secure deletion for digital]
• 5) Analysis files (codes, codebooks, memos, datasets, outputs)
  • Contains identifiers? Yes / No / Sometimes
  • Where stored: [Fill in]
  • Who can access: [Fill in]
  • Retention start: Date created / Other: [Fill in]
  • Retention end (destroy by): [Fill in]
  • Reason for keeping: [Fill in: reproduce analysis, respond to questions]
  • Rule for identifiers in analysis: [Fill in: prohibit direct identifiers, use participant IDs]
  • Destruction method: [Fill in]

Crosswalk: linkages and keys

If you use participant IDs, you may also have a linkage file (a “key”) that connects IDs to identities. Many IRBs treat this as the most sensitive file you have.

• Linkage file kept? Yes / No
• Where stored: [Fill in, separate from study data]
• Who can access: [Fill in, usually PI only]
• Destroy by: [Fill in, often earlier than de-identified transcripts]
• Reason: [Fill in, e.g., follow-up contact, incentive distribution]

Legal hold and “stop destruction” rules (what to include)

A retention plan should say what happens when you cannot delete data on schedule. A legal hold is a formal instruction to preserve information because of actual or anticipated litigation, investigation, or audit.

Even outside litigation, you may need a hold for things like a research misconduct inquiry or a sponsor audit. Your wording can stay simple, but it should be explicit.

Sample legal hold clause (fill-in)

• Legal hold trigger: [Fill in: notice from legal counsel, IRB, sponsor, institution]
• Scope: [Fill in: which datasets, date ranges, custodians]
• Actions: Suspend scheduled destruction for in-scope items; preserve originals; restrict access as needed.
• Release: Resume destruction only after written release from [Fill in authority].
• Documentation: Log hold start date, scope, and release date in the audit trail.

If your files live in automatic backup systems, include a sentence about how a hold affects backups. You can say you will coordinate with institutional IT to preserve needed backups during the hold period.

How to document destruction with an audit trail

An audit trail is a record that shows you followed your plan. It can be a spreadsheet, a ticket in your lab’s system, or a folder of signed destruction forms, as long as it is consistent and protected.

What to log (minimum fields)

• Item type: audio, identifiable transcript, de-identified transcript, consent form, analysis file, linkage key
• Unique ID: participant ID or file ID (avoid names in the log if possible)
• Location: system/path or platform name
• Destruction date/time: [Fill in]
• Destroyed by: name/role
• Method: secure delete, shredding, vendor deletion confirmation
• Reason: per retention schedule / participant request (if applicable) / other
• Verification: second-person check, system log, or confirmation receipt
• Notes: legal hold applied? exceptions?

Destruction documentation template (copy/paste)

• Project: [Study title + IRB #]
• Record destroyed: [File type + ID]
• System/location: [Fill in]
• Destruction method: [Fill in]
• Destroyed by (role/name): [Fill in]
• Verified by (role/name): [Fill in, optional]
• Date/time: [Fill in]
• Confirmation evidence: [Fill in: system log screenshot stored at…, ticket #…, vendor email saved at…]

For paper consent forms, document shredding and the chain of custody. For digital files, rely on institutional IT guidance for secure deletion and keep a record of the action taken.

Pitfalls that can derail an IRB-friendly retention timeline

Most problems happen because files multiply. A good plan names the common “extra copies” and assigns responsibility for cleaning them up.

Pitfall 1: Multiple uncontrolled copies

• Downloads on personal laptops.
• Sync folders (desktop cloud drives) that replicate to multiple devices.
• Email attachments and chat uploads.

Fix this by stating where files may live and banning other locations. You can also require team members to delete local copies after upload to the approved store.

Pitfall 2: Keeping identifiable transcripts “just in case”

Identifiable transcripts often linger because they are convenient for searching. If you need search, build it into your de-identified workflow and delete the identifiable version on schedule.

Pitfall 3: Forgetting vendor platforms and shared portals

If you use outside help, include vendor retention and deletion in your plan. Ask for a clear statement of when files are removed from portals, and save that confirmation in your audit trail.

If you need help cleaning up transcripts, consider transcription proofreading services so you can verify accuracy quickly and shorten how long you keep raw audio.

Pitfall 4: Timelines that conflict with consent language

If your consent form says “we will delete recordings after transcription,” do not write an IRB plan that keeps recordings for years. Update the consent language or adjust the retention schedule so they match.

Choosing the right approach: minimal-risk vs. long-horizon retention

If you feel stuck, choose one of these patterns and adapt it to your study. The goal is consistency and clear reasoning, not a perfect number.

Option A: Minimal-risk, minimal-retention model

• Destroy audio/video soon after transcript verification.
• Destroy identifiable transcripts right after de-identification and QA.
• Retain de-identified transcripts and analysis files through publication and required retention windows.
• Retain consent forms per institutional or sponsor requirements.

Option B: Long-horizon model (only if justified and consented)

• Keep de-identified transcripts and core analysis documentation longer for future research.
• Limit audio/video retention and linkage keys unless you truly need them.
• Use stricter access control and periodic reviews to confirm you still need retention.

If your timeline depends on meeting accessibility goals for recorded content, you may also need captions or subtitles. In that case, plan how long you will keep caption files and where they live, and consider closed caption services for consistent formats.

Common questions

How long should we keep audio recordings after transcription?

Keep them long enough to verify the transcript and resolve disputes, then delete them if you do not need them for a stated purpose. Your consent form and IRB protocol should match that timing.

Do we need separate retention timelines for identifiable and de-identified transcripts?

Yes, if you create both. Separate timelines help you minimize the period when identifiable text exists and reduce risk while still supporting analysis.

What counts as “de-identified” in a transcript?

At minimum, remove direct identifiers such as names and contact details. You should also review indirect identifiers that could point to someone in a small group, and generalize them when needed.

Where should we store the linkage key that maps participant IDs to names?

Store it separately from the research data with tighter access, often limited to the PI or a small number of authorized staff. Set an earlier destruction date if you only need it for incentives or follow-up.

How do we handle a participant request to delete their data?

Follow what your consent form promises and what your IRB approval allows. Document the request and your action in the audit trail, and note any limits if data has already been de-identified or aggregated.

What is a legal hold, and who can issue it?

A legal hold is an instruction to stop normal deletion because of an investigation, audit, or legal matter. Your institution’s legal counsel, compliance office, sponsor, or IRB may have authority depending on the situation.

How do we prove we actually destroyed the files?

Use a destruction log plus evidence such as system logs, ticket numbers, or vendor deletion confirmations. For paper, keep a signed shredding record and chain-of-custody notes.

If you’re collecting interviews, focus groups, or recorded research sessions, accurate transcripts make de-identification and retention much easier. GoTranscript can support your workflow with professional transcription services, so you can move from recordings to verified text and follow your retention plan with more confidence.