Clarifying unsafe behaviors of construction workers through a complex network of unsafe behavior chains

Numerous studies indicate that unsafe behaviors of construction workers (UBCW) are the direct causes of most construction safety incidents. To unveil the interconnections among unsafe behaviors and the underlying mechanisms of different accident types, this study extracts the unsafe behavior chains from accidents. Based on the theory of complex networks, we constructs both the whole network of UBCW and four subnetworks corresponding to different accident types by using unsafe behavior chains. The findings reveal that the degree distributions of all networks conform to truncated power-law distributions, highlighting the critical role of hub nodes in accident occurrences. Moreover, the whole network exhibits small-world characteristics, underscoring the need to address the propagation and clustering effects of unsafe behaviors, reassess risk management approaches, and devise targeted control strategies and measures. By analyzing the topological structures of the networks, this paper ranks the importance of different behaviors across various metrics and proposes tailored accident prevention recommendations through comparative analysis of different accident types. This research holds significant theoretical and practical implications for refining accident causality models and enhancing safety management practices at construction sites. However, the study has limitations, notably future studies should focus on the absence of specific construction accident databases and the unaddressed link between accident severity and unsafe behaviors.