Connecting the dots: A new model for understanding how unsafe behaviors happen in construction

What is it about?

We designed a model based on a Bayesian network structure that can learn from existing data from any sector of the construction industry and automatically determine the multiple factors that can lead to human error. Using our model, it is easy to see that human error is not an isolated occurrence in the workplace, but rather the end result of one of many potential cascades of events. Thus, we present a useful tool to shift the attention of managers to certain aspects of the workplace that may need tweaking, thereby preventing human errors and potentially saving lives.

Featured Image

Why is it important?

The construction industry is one of the most dangerous ones to work in because most human errors directly translate to accidents. Previous studies have established that workers usually make errors because of the conditions of the working environment, and many models for analyzing these factors have been built. However, the mechanisms and interactions that link the workplace conditions with human error are not clear, especially since existing models rely on subjective data. How can we build a model that is not only more objective, but also clearly indicates all intermediate factors that arise in the working environment, which ultimately lead to human error?