Causal inference of construction safety management measures towards workers' safety behaviors: A multidimensional perspective
Zhitian Zhang,Heng Li,Hongling Guo,Yue Wu,Zhubang Luo
DOI: https://doi.org/10.1016/j.ssci.2024.106432
IF: 6.392
2024-01-22
Safety Science
Abstract:Workers' unsafe behaviors are a major contributor to construction accidents. To improve safety behaviors, managers have developed comprehensive management measures. However, due to the complexity of workers' unsafe behaviors, it is difficult to understand the specific impact mechanism of management measures on different types of unsafe behaviors, thus making it difficult to strengthen targeted management efforts. This research aims to uncover the impact mechanism of management measures on workers' safety behaviors from a multidimensional perspective through a novel method of causal inference. A seven-week questionnaire survey was conducted to evaluate the multidimensional performance of managers' and workers' safety behaviors, getting a total of 5,608 valid questionnaires. Change rates and heatmaps were employed to illustrate the improvements and correlations of different management dimensions. Then, the Structural Causal Model-based causal inference was performed to quantitatively analyze the causal effect. The results show that the overall safety behaviors are improved significantly, and the correlations among management measure dimensions are strengthened after the interventions. Furthermore, it indicates that Objectives and Assessments, Safety Organization and Hazard and Emergency Management should be emphasized in the practice of safety management, as they have a continuous impact on workers' safety behaviors. Meanwhile, while some management dimensions are improved, their impact on certain unsafe behaviors is still restricted in terms of stability and pertinence. These research findings contribute to the body of knowledge of construction safety management and benefit relevant practices by revealing the impact mechanism of management measures on workers' safety behaviors in a detailed and comprehensive level.
engineering, industrial,operations research & management science