Icons and examples build accurate mental models faster than text alone
Aliases: privacy icons and examples · data-flow illustration · visual privacy notice
What it is
A multimodal privacy explanation combines concise text with semantically defined icons, a data-flow depiction, or concrete examples so people can identify relations among data, actions, participants, and consequences quickly. Its advantage comes from making abstract propositions comparable and situated, not from images being inherently truthful or universal.
Why it happens
Consistent icons provide visual indexes that make recurring location, third-party, or retention concepts easy to find while scanning. Examples with concrete actors and boundaries supply reasoning instances for abstract rules. But an unlearned icon mapping or an example showing only the mildest case can produce fast, overbroad generalization.
Studying it
Holding propositions constant, compare text, text plus icons, text plus examples, and all three. Measure target-location time, flow reconstruction, transfer to new cases, exception recognition, and delayed recall. Ask participants to interpret icons independently to expose unintended mappings. Speed alone is unsafe: decorative cues can exchange surface efficiency for faster error.
Where it stops holding
Icons cannot independently carry novel, contested, or legally precise meaning and need readable labels. An example instantiates a rule rather than limiting its scope, so include a consequential boundary or counterexample. Graphics dependent on color, cultural convention, or vision require textual alternatives and cross-cultural validation.
Applying it
- Maintain a small consistent icon vocabulary for data class, source, recipient, retention, and control, labeling every first occurrence.
- Write one typical and one boundary example using real product objects, marking them as examples rather than the complete scope.
- Use arrows and regions for movement while retaining text labels for actor, purpose, condition, and time.
- Require unprompted icon interpretation, flow reconstruction, and novel-case transfer; fail designs that improve speed while reducing accuracy.