Psychological distress is a growing concern in the construction industry, yet its relationship with work performance remains poorly investigated. Most existing studies assume a linear and negative association, overlooking the potential complexity of this relationship. This study aims to investigate the relationship between psychological distress and work performance, with a focus on the Vietnamese construction context.
A total of 422 valid responses were collected using a snowball sampling approach. Several machine learning (ML) algorithms were applied to explore the relationship between psychological distress and work performance.
Gaussian Process Regression was the most robust approach (RMSE = 0.3892, MAE = 0.2842, MASE = 1.2172, R2 = 0.8481). The analysis revealed an inverted U-shaped relationship consistent with the Yerkes–Dodson law. Specifically, work performance improved under mild psychological distress but declined beyond a threshold of 8.75.
The results underscore the importance of understanding and managing distress to remain productive. The study highlights that work performance improves with mild distress, but begins to decline sharply when psychological distress passes a certain level. These findings suggest that employees should not try to eliminate all distress, but instead focus on keeping distress at a healthy level.
