Statistical Analysis of Flight Crew Safety Behavior Using Fuzzy Set

Despite advancements in aviation technology, pilot-related human errors continue to contribute to accidents. This study examines flight crew safety behavior using both 5-point Likert-scale responses and subjective evaluations converted into fuzzy numbers. Seven predictors—management commitment, colleague commitment, organizational support, staff and equipment, collaboration and involvement, flight quality assurance, and just culture—were analyzed through correlation analysis, SEM, and path analysis with crisp data. While SEM showed excellent model fit, several predictors were not significant in the path analysis. To address these limitations, fuzzy statistical analysis was applied. Subjective single-item responses were transformed into triangular fuzzy numbers, and fuzzy regression using Lasso and least absolute deviation revealed significant relationships not detected in crisp data. Differences by airline type and flight experience were also identified, indicating lower fairness perception among low-cost carrier pilots and varying cooperative behavior depending on flight hours. Overall, the study demonstrates that combining crisp and fuzzy methods provides a more comprehensive understanding of pilots’ safety behavior and offers useful implications for aviation safety management.