2019 CSCE Annual Conference - Laval (Greater Montreal) Conference
Dr. John Gambatese, Oregon State University
Dr. Ahmed Al-bayati, Western Carolina University (Presenter)
The interactions between construction workers have been studied in many ways. However, network-based interactions are barely investigated. Additionally, the few attempts to study the influence of social network on construction safety research have not used inferential aspects. Accordingly, the present study investigates the application of Exponential Random Graph Models (ERGMs) in construction safety research. In this study, an exploration of ERGMs is introduced theoretically by explaining the basics of the modeling, and practically by introducing a case study to enhance the understanding of how ERGMs can be used. The research methods include a literature review and an analysis of visual data from a construction field. The results show that both approaches show it is possible to use ERGMs in safety research. Additionally, the findings suggest that there is a higher probability of having an unsafe condition (i.e., higher risk) in construction sites when there is higher connectivity. The contribution of this study is the introduction of a new inferential procedure for analyzing worker interactions on construction sites concerning safety.