Analysis of pilots' cognitive bias based on information processing model

In this paper, we conducted a study on the mechanism of cognitive bias affecting pilots. The purpose of this study is to reduce aviation accidents caused by pilots' cognitive biases. A questionnaire based on information processing model was designed, build up from three stages: information input, information processing and information output. Through questionnaires, the active civil aviation pilots in China were investigated, and the pilots' cognitive biases were quantitatively analyzed by Structural Equation Model (SEM). The results showed that: At the stage of information input, the factors affecting the maximum and minimum are the halo effect and the representative deviation respectively, and the path coefficients are 0.68 and 0.41; At the stage of information processing, the factors affecting the maximum and minimum are the idiosyncratic deviation and the framework respectively, and the path coefficients are 0.45 and 0.20; At the stage of information output, the factors affecting the maximum and minimum are the hindsight and self-attribution deviations respectively, and the path coefficients are 0.80 and 0.33. By analyzing the phenomenon of cognitive bias of pilots during flight, it is helpful to reduce errors of judgement and decision-making, and to ensure aviation safety.