Edit Data

The Edit Data page allows authorised administrators to make limited corrections to collected data without breaking the audit trail.

What you can do

Action	Description
Soft-delete a subject	Mark a participant’s entire record as deleted. The data is retained in the database but excluded from exports and counts by default.
Edit individual responses	Correct the value of a specific question response for a specific participant.

All changes are logged with a timestamp, the username of the administrator who made the change, the original value, and the new value.

Participant table

The table shows participants with:

• Participant ID: unique identifier assigned by the tablet
• Facility: which facility the participant was enrolled at
• Completed: whether the interview was completed
• Status: active or deleted
• Actions: edit or delete

Filtering

Use the Facility dropdown to filter by facility. Check Show deleted to include soft-deleted records in the list.

Soft-deleting a participant

Click the delete icon next to a participant and confirm. The participant’s status changes to deleted. Deleted participants:

• Are excluded from exports by default (unless Show deleted is enabled and you re-export)
• Are retained in the database with all their responses intact
• Can be restored by an administrator

Editing a response

Click the edit icon next to a participant to see their responses. Click the pencil icon next to any response to change its value. Enter the corrected value and provide a reason for the edit. The original value is preserved in the audit log.

Edit log

Click Download Edit Log to download a CSV of all edits and deletions, including:

• Timestamp
• Administrator username
• Participant ID
• Field edited
• Original value
• New value
• Reason given

The edit log is the authoritative record of all post-collection corrections.

When to use Edit Data

Use Edit Data sparingly and only for genuine data quality corrections, for example, a transcription error in a numeric response, or a test participant whose record should be excluded. Programmatic corrections to entire variables should be done in your analysis scripts, not in the database.