ABSTRACT

Construction workers are exposed to psychosocial pressures that may influence safety behavior, yet empirical findings on the relationship between work stress and safety behavior remain inconsistent. This study examines the association between work stress and safety behavior and tests whether education level and marital status moderate that relationship. A quantitative explanatory design was applied to 35 workers selected purposively from a population of 100 workers on a warehouse construction project in Sidoarjo. Data were collected using a questionnaire and analyzed using Hayes' PROCESS moderation regression. The instruments showed acceptable reliability (Cronbach's alpha = 0.855 for work stress and 0.891 for safety behavior). The education-moderated model produced R² = 0.067, while the marital-status-moderated model produced R² = 0.0856. Work stress was not significantly associated with safety behavior, and neither education level nor marital status showed a statistically significant interaction effect. These results should be interpreted cautiously because the small, single-project sample limits statistical power and generalizability. The findings indicate that demographic moderation was not detected in this sample and highlight the need to examine organizational and psychological factors, including safety climate, leadership, knowledge, and motivation, in broader samples of construction workers.