The Perception of Visual UncertaintyRepresentation by Non-Experts

Area of Science:

• Geospatial analysis
• Cognitive science
• Data visualization

Background:

• Accurate interpretation of uncertainty is crucial in scientific and technical fields.
• Visual representations are commonly used to convey uncertainty, but their effectiveness can vary.
• Understanding how non-experts process visual uncertainty information is key to improving communication.

Purpose of the Study:

• To evaluate how non-experts judge point probability based on different visual representations of uncertainty.
• To investigate the influence of visualization type on the internal models observers form of uncertainty distributions.
• To explore the relationship between participants' numeracy and their interpretation of visual uncertainty.

Main Methods:

• Conducted an experiment with 140 non-expert participants assessing point probability judgments.
• Utilized seven distinct visualizations to represent the positional uncertainty of an earth layer boundary.
• Analyzed how visualization type and participant numeracy affected the internal models of uncertainty.

Main Results:

• Most observers modeled uncertainty distributions similarly to a normal distribution across all visualization types.
• The specific visual form of uncertainty significantly influenced these internal models.
• Participant numeracy correlated with the constructed internal models of uncertainty.

Conclusions:

• The visual representation of uncertainty demonstrably affects perceived certainty among non-experts.
• The presence or absence of a center line in visualizations did not impact the internal uncertainty models.
• Findings suggest careful consideration of visualization design is needed for effective uncertainty communication.