Tracing the Factors Influencing Learning from Incidents Behavior in the Construction Industry: Based on a Complex Systems Perspective

Learning from incidents (LFI) is an effective behavior to prevent risk and uncertainty in the construction industry, while being subject to multiple systems working together. Existing research on LFI behavior is still at the exploratory and conceptual stage and ignores the role of multiple systems and the multidirectional causal relationships between factors, creating a research gap where factors in complex systems cannot be accurately identified. This research is conducted to understand the structural causes of LFI behavior and the interaction mechanisms between factors. The hybrid DEMATEL-ISM method was used to explore the critical factors in a multisystem context influencing LFI behavior and to analyze the relationship between the factors. A causal map representing degrees of centrality and causality and a hierarchical model representing causal relationships were constructed to understand the structural causes that influence LFI. It is found that the factors influencing LFI behavior can be attributed to the organizational, individual, work, external environment, and knowledge systems. The five systems we constructed can serve as a future direction for studying complex systems that LFI urgently needs to address. Furthermore, based on the results from the hybrid method, we found that the critical factors of LFI behavior have intricate systemic and cross-systemic interactions among them. The five systems classified in this study eliminate the research gap of factor fragmentation in a multisystem context. Meanwhile, the influencing factors of LFI and their interrelationships are conducive for practitioners to take adequate measures to enhance organizational learning and improve safety performance in the construction industry.