Related Experiment Videos
Ejection decisions by strike pilots: an extreme value interpretation.
R John Irwin1, Kathleen S N Callaghan
1Department of Psychology, The University of Auckland, New Zealand.
Aviation, Space, and Environmental Medicine
|January 21, 2006
Summary
Pilot ejection decisions are explained by the extreme value model, which better fits asymmetrical Receiver Operating Characteristic (ROC) curves than traditional models. This finding clarifies why pilot ejection decisions show varied response patterns.
Area of Science:
- Human Factors
- Decision Science
- Aviation Psychology
Background:
- Receiver Operating Characteristic (ROC) curves for pilot ejection decisions exhibit asymmetry.
- Existing Gaussian models with unequal variances fit these curves but lack clear interpretation.
- Understanding this asymmetry is crucial for aviation safety and decision-making analysis.
Purpose of the Study:
- To explain the asymmetry observed in pilot ejection decision ROC curves.
- To introduce and validate the extreme maximum value model for pilot ejection decisions.
- To provide a more interpretable model for pilot emergency response.
Main Methods:
- Utilized the extreme maximum value model to describe pilot decision-making processes.
- Applied information criteria (Akaike and Bayesian) for model comparison.
- Fitted the extreme value model to published asymmetrical ROC curves.
Main Results:
- The extreme maximum value model, with one parameter, provided a superior fit to asymmetrical ROC curves.
- This model demonstrated a better fit compared to the two-parameter unequal variance Gaussian model.
- Model comparison favored the extreme value model based on information criteria.
Conclusions:
- Pilot ejection decisions in emergencies are based on exceeding a criterion from any single information source.
- The extreme value model accurately describes this decision policy and explains ROC curve asymmetry.
- This provides a credible theoretical framework for analyzing pilot emergency response patterns.