The Use of Cream in the Investigation of Human Error in Aviation Events: a Case Study

To investigate Human Factors issues of the Traffic Alert and Collision Avoidance System (TCAS), a part-task simulation has been conducted. The analysis method employed was based on the CREAM principles and included the following steps: Predictive analysis, Simulations, Data analysis, Retrospective analysis. The project called ARCADES , presented in this paper, concerns the errors identification and research of causes. The collective work based on CREAM is composed of four steps alternating between work in pairs and plenary session: the events identification and classification, consensus on the events to be analyzed, the retrospective analysis, and the mutual analysis. The work group was composed of pilots, controllers, investigators and Human Factors experts. The analysis was based on the data collected during the simulations: video recordings, questionnaires, audio recording of debriefing. Concerning the identification step, results show the interest of the clear distinction between causes and consequences, as it exists in CREAM. Indeed, it enables to bind the analysts to objectivity in the identification and the calling of the events. Inter-pairs variability remains important for the three scenarios studied. Participants agreed on the interest of the method concerning the search of causes. The method enables to pursue analyses more deeply and more exhaustively, results show that the method guides towards systemic causes. Results also emphasize the complementarity of different points of view and the interest of a collective approach