Constructing a Bayesian network model for improving safety behavior of employees at workplaces. (January 2017)

• Record Type: Journal Article
• Authors: Mohammadfam, Iraj; Ghasemi, Fakhradin; Kalatpour, Omid; Moghimbeigi, Abbas
• Is Part Of: Applied ergonomics. Volume 58(2016)
• Page Start: 35
• Page End: 47
• Publication Date: 2017-01
• Subjects: Bayesian network -- Human behavior -- Accident prevention -- Safety climate; Human engineering -- Periodicals; 620.82
• DOI: 10.1016/j.apergo.2016.05.006
• Languages: English
• ISSNs: 0003-6870

Abstract

Abstract: Introduction: Unsafe behavior increases the risk of accident at workplaces and needs to be managed properly. The aim of the present study was to provide a model for managing and improving safety behavior of employees using the Bayesian networks approach. Methods: The study was conducted in several power plant construction projects in Iran. The data were collected using a questionnaire composed of nine factors, including management commitment, supporting environment, safety management system, employees' participation, safety knowledge, safety attitude, motivation, resource allocation, and work pressure. In order for measuring the score of each factor assigned by a responder, a measurement model was constructed for each of them. The Bayesian network was constructed using experts' opinions and Dempster-Shafer theory. Using belief updating, the best intervention strategies for improving safety behavior also were selected. Results: The result of the present study demonstrated that the majority of employees do not tend to consider safety rules, regulation, procedures and norms in their behavior at the workplace. Safety attitude, safety knowledge, and supporting environment were the best predictor of safety behavior. Moreover, it was determined that instantaneous improvement of supporting environment and employee participation is the best strategy to reach a high proportion of safety behavior at the workplace. Conclusion: The lack of a comprehensive model that can be used for explaining safety behavior was one of the most problematic issues of the study. Furthermore, it can be concluded that belief updating is a unique feature of Bayesian networks that is very useful in comparing various intervention strategies and selecting the best one form them. Highlights: This study used Bayesian networks for modeling and assessing safety behavior of employees. Bayesian networks are a promising tool for planning interventions for improving safety behavior of employees at workplaces. Safety attitude, safety knowledge, and supportive environment are best predictors of employees' safety behavior. Providing a supportive environment and facilitating employees' participation are crucial for improving safety behavior.