Statistical Analysis of Safety Incidents and the Implications of Imperfect Reporting

Quantitative models of safety incidence allow managers to connect risk factors and elements of the safety management system (SMS) (e.g., risk-management practices, inspection protocols, processes for reporting unsafe conditions) to outcomes. The results provide evidence that can be used to allocate resources to those incident-prevention efforts with the largest benefit/cost ratios. However, conventional methods of statistical and regression analysis of safety incidents do not account for the fact that not all incidents are reported. By relying on conventional methods, it is possible that resources are being misallocated, and that the industry is missing opportunities for improvement. This paper specifies nonconventional regression models that explicitly account for imperfect reporting. The methods are applied to an operator-provided data set of safety incidents. The analysis indicates which variables contribute to changes in incidence, which contribute to imperfect reporting, and which contribute to both. While all of the results of the case study are not generalizable, they clearly demonstrate the incremental value of a more complete model of safety incidence and reporting behavior.