Appraisal-based emotion identification in urban search and rescue exercises using Roseman's model of emotion elicitation

This dataset contains coded observational data collected from three Urban Search and Rescue (USAR) simulation exercises conducted in Australia and New Zealand between March and October 2024. Two of the exercises were official INSARAG External Reclassifications (IERs), and one was a full-scale mock exercise. Each exercise simulated an earthquake in a foreign country and involved continuous operations over a three-day period with approximately 100 participants, including commanders, rescue workers, and support staff. The dataset is comprised of 28 excerpts representing emotionally significant events observed during daylight operations (08:00–18:00), spanning approximately 90 hours of direct observation in total. Each excerpt is accompanied by a structured analysis applying five appraisal dimensions from Roseman’s model of emotion elicitation (goal consistency, motive consistency, agency, certainty, and control potential) culminating in the identification of a specific discrete emotion. The data are anonymised and adhere to UTAS Human Research Ethics Committee protocols.