Quantifying Safety Culture Maturity: Empirical Validation of a Psychometric–Behavioral Framework and Its Association with Industrial Incident Rates

Safety culture maturity evaluations frequently rely on perceptual surveys or qualitative maturity frameworks, while objective indicators of behavioral fidelity and incident outcomes are often assessed separately.This study introduced and empirically tested a five-dimensional Safety Culture Maturity Index (SCMI) that combines psychometric and behavioral data.The study involved 1,240 operational staff from 18 facilities in the petrochemical, construction, and heavy manufacturing sectors, utilizing 24-month data on facility-level incidents and exposures.The measurement model included leadership commitment (LC), safety voice and psychological safety (SV), procedural compliance (PC), risk competence (RC), and just and fair reporting (JR).Confirmatory factor analysis (CFA) revealed robust standardized item loadings (0.742-0.835), with all item z-values exceeding 26.85.Internal consistency was strong across all five constructs (α=0.854-0.884;ω=0.855-0.886),with composite reliability (CR) ranging from 0.855 to 0.885 and average variance extracted (AVE) from 0.596 to 0.658.The Fornell-Larcker diagnostics confirmed discriminant validity through the Heterotrait-Monotrait ratio (HTMT≤0.62).Facility SCMI scores varied from 1.91 to 4.60, with petrochemical facilities having the highest average (3.86), followed by manufacturing (3.55), and construction (3.01).A facility-level Poisson loglinear model, adjusted for exposure hours and using heteroscedasticity-robust standard errors, identified a significant negative association between SCMI and incident counts (β=-1.149,p<.001), corresponding to an incident rate ratio of 0.317 for each unit increase in SCMI within the petrochemical reference sector.The SCMI × industry interaction was significant for construction (β=0.497,p<.001) and manufacturing (β=0.328,p=.046), suggesting that the strength of the SCMI-incident relationship differs across industrial