| description abstract | Despite its significance, the Indonesian construction industry has poor safety performance. Improving the safety climate has been seen as a way to improve safety in the industry. Research on safety climate in this context has identified a range of safety climate factors that require improvements. However, construction organizations face difficulties in implementing improvement recommendations due to resource constraints. In order to help construction organizations in their efforts to improve the safety climate, this research demonstrates the use of supervised classification approaches to identify specific safety climate factors that construction organizations should focus on. Data were collected from 311 construction practitioners in Indonesia using a 22-item safety climate survey. Supervised classification methods, comprising ensemble methods, Support Vector Machine, Naïve Bayes, and Nearest Neighbor, were used. The analysis identified 14 safety climate items that can represent the original dataset with high accuracy (93%). These 14 items can be considered crucial items that should be prioritized in the Indonesian construction industry. These items revealed that, due to the high power distance culture in Indonesia, top-down approaches, such as giving clear instructions, providing training, and reminding people often about safety, are effective for engaging employees to focus on and participate in safety. The findings also suggest that understanding cultural context is important to determine effective strategies to improve safety. This research has also demonstrated the potential application of supervised classification approaches to help decision makers improve safety by focusing on crucial factors within a context. | |
