Designing environmental uncertainty information for experts and non‐experts: Does data presentation affect users’ decisions and interpretations?

• Published in: Meteorological applications 2020-01, Vol.27 (1), p.n/a
• Main Authors: Mulder, Kelsey J., Lickiss, Matthew, Black, Alison, Charlton‐Perez, Andrew J., McCloy, Rachel, Young, Joseph S.
• Format: Article
• Language: English
• Subjects: communication; Decision making; Decisions; Experts; Graphics; Ice cover; Ice thickness; Mental task performance; Probability theory; Representations; Risk communication; Shipping; Uncertainty
• ISSN: 1350-4827; 1469-8080
• DOI: 10.1002/met.1821

Uncertainty information in natural hazard forecasts is increasingly being communicated explicitly. This study was designed to determine whether different ways of communicating uncertainty graphically affects the decisions and interpretations of forecasts and if expertise was a factor in the decisions and interpretations from forecasts explicitly showing uncertainty. In a hypothetical decision‐making task regarding ice thickness and shipping, 138 experts and non‐experts received ice‐thickness forecasts in four different presentations expressing uncertainty: worded probability, spaghetti plot, fan plot and box plot. These forecasts contained no measures of central tendency. There was no consistent difference in decision or best‐guess forecast (deterministic ice thickness forecast based on the forecast representation) between the different forecast representations. However, participants interpreted different amounts of uncertainty across the different forecast representations. Experts made significantly more economically rational decisions than non‐experts, interpreted lower best‐guess forecasts and inferred significantly more uncertainty than non‐experts. These results suggest that care be taken in choosing how uncertainty is represented in forecasts, especially between expert and non‐expert audiences. This research sought to determine if the way in which uncertainty information was presented affected expert and non‐expert interpretations of and decisions made using that information. The results suggest that participants interpreted different amounts of uncertainty across the forecast representations. Experts made significantly more economically rational decisions than non‐experts and inferred more uncertainty.