A methodology to assess the causation relationship of seafarers’ unsafe acts for ship grounding accidents based on Bayesian SEM

Abstract

Unsafe acts have been recognized as the substantial factors leading to maritime accidents. An obstacle to analyzing seafarers' unsafe acts involved in maritime accidents is the lack of data. Bayesian approach in structural equation modelling (SEM) estimation has attracted extensive attentions across other fields, due to its advantages that can handle complicated models and small sample size. Therefore, this study proposes an innovative methodology to address the multifactor risk coupling and causation involved in seafarers' unsafe act contributing to ship grounding accidents. Based on the principle of Bayesian SEM, the advantages of Grounded theory and Human Factors Analysis and Classification System (HFACS) are integrated into this methodology. Specifically, Grounded theory and HFACS model are applied to identify and structure seafarers' unsafe acts and their contributory factors, facilitating the development of the conceptual model. Then, by employing Bayesian SEM, the causation paths and effects of seafarers' unsafe acts are examined. The results reveal that organizational influences play a significant role in the causation mechanism of seafarers' unsafe acts, and preconditions for unsafe acts are the only variables that exert direct impact on unsafe acts. Moreover, unsafe supervision and preconditions for unsafe acts are served as intermediate factors affecting seafarers' unsafe acts. In addition, this results indicate the feasibility of Bayesian SEM in analyzing seafarers’ unsafe acts.