Running consent for field recording, from approach to archive

Field recording — streets, markets, transit stations, community events — is the hardest consent setting there is. There is no appointment, no waiting room, and no second chance: the moment passes and the person is gone. The process has to be simple enough to run in a busy place and strict enough to hold up in a review months later.

Everything below assumes a trained collector working from a defined script, not a volunteer with a recorder and good intentions.

The approach, in order

The sequence matters as much as the content, because it decides whether the person is agreeing or just not objecting:

• Introduce yourself and the organization first, and show an ID badge without being asked.
• State the purpose in one plain sentence: what the recording is for and what happens to it.
• Ask whether they are willing to hear the details. If not, thank them and move on — no persuasion, no follow-up.
• Move to a quieter spot before completing consent; a noisy street is not a place where anyone can read or hear terms.
• Record only after consent is complete.

The form, and reading it aloud

The form carries the minimum: what is recorded, the purpose, who will receive it, how long it is kept, the right to stop or withdraw, and how to reach the organization. Practices that matter in the field:

• Read the form aloud in full, every time, even when literacy is clearly not an issue — uniformity is easier to defend than judgment calls.
• Offer the form in the participant's language. If no collector on site speaks it, a second collector or a recorded audio version handles the consent, never a rough on-the-spot translation.
• Accept a signature, a thumbprint, or a timestamped verbal consent recorded at the start of the audio, per the participant's preference and local practice.
• Have the collector sign an attestation that the process was followed, including any deviation.

Refusals: counting, not converting

A refusal is a result, not a failure, and the process should treat it that way. Stop immediately, record nothing about the person, and log the refusal as a count with a location code and time — no description, no photo, no note about who they were.

Two rules prevent the common abuses. Collectors do not circle back to the same person in a later session; a repeated approach turns a no into pressure. And the refusal rate is treated as a health metric for the campaign: a public-space effort with almost no refusals usually means the approach is too aggressive, or the participants did not understand what they were agreeing to. Never record first and ask afterwards, not even for a few seconds to test the microphone — the pre-consent recording is the finding that sinks an entire batch.

The bystander problem

Background voices are the classic field-recording failure: the participant consented, the crowd did not. Rules that keep it manageable:

• Prefer locations and times where a clean take is possible; a quieter corner is a consent safeguard, not just an audio improvement.
• If an identifiable bystander conversation lands in the audio, either stop and re-record or mark the file for review and remove the segment before delivery.
• Where bystanders cannot be avoided — markets, transit — set a written delivery rule for what gets discarded, so the decision is made once, in advance, rather than per file by whoever is editing.
• Never keep bystander segments "just in case." They are data subjects with no consent record, and their presence in the archive is a liability with no upside.

After the session: the archive side

Every consent form gets an ID that appears in the provenance log for the files it covers. Forms are scanned the same day, while the session is still identifiable and the collector can resolve ambiguities. The consent version in force is logged with the batch, not with the project.

Withdrawals route to one named owner who locates the files by consent ID, removes them from storage and any pending delivery, and confirms the deletion in writing. In jurisdictions that require all parties to consent to a recording, this process satisfies the requirement by design, because the audio begins with the consent on the record. The whole process is judged months later, by people who were not there — the record is all they will see, and the record is built in the field, in minutes, one participant at a time.