✓✓ Double Check DIALOGUE CODING
Saved on this device
ANNOTATION WORKSPACE

Dialogue

Dialogue
1 Select utterances—2 Label action—3 Score
TRANSCRIPTClick anywhere on a row to select
No utterances selectedSelect one turn or an entire exchange.

Import your work

Choose a case folder, a dialogue CSV, or a Double Check backup. Files are read on your device.

Rater forms supply the label set. Existing scores are not imported as new annotations.

Export annotations

A consistent coding workflow

  1. Select evidence. Click one or more complete utterances. Select only User turns, or include Character turns for an exchange. Non-adjacent selections and overlapping actions are allowed.
  2. Create an action. Actions are numbered from 1, ordered by their first selected utterance. Actions that start on the same turn keep creation order.
  3. Choose one label and score. Each occurrence gets exactly one label and a score from 1 to 3. Use Edit turns to revise evidence.
  4. Review and export. Switch to By label to review repeated actions. Export JSON to back up all your work.

This version scores each occurrence. The original coding instructions instead request one overall score per label; grouped exports do not calculate that overall score. “No occurrences” means no evidence has been annotated yet.

Enter creates the selected action. Escape clears a selection. Each row also supports Tab + Space. Imported files and annotations stay in this browser; they are not synced to GitHub.

Add an action label