Analysis of Truck Accidents Using Structural Equation Models

Many researchers and practitioners in the trucking industry have expressed concerns with the implementation of Compliance, Safety and Accountability (CSA), which is a safety management initiative created by the Federal Motor Carrier Safety Administration (FMCSA). CSA assigns Behavior Analysis Safety Improvement Categories (BASIC) scores to carriers, based on driver and vehicle performance. This technique of assigning BASIC scores to carriers is a cause of concern for many stakeholders in the industry. This article presents a structural equation model (SEM) to understand the relationship between the BASIC scores and crash records of the carriers. Exogenous variables are Unsafe Driving, Fatigue Driving, Driver Fitness, Controlled Substances/Alcohol, and Vehicle Maintenance. Crash-Per-Million-Miles is the endogenous variable in the SEM. We observed existence of a latent variable driver characteristic, which in turn influences three of the exogenous variables. Based on the SEM model output we concluded that four of the exogenous variables–Unsafe Driving, Fatigue Driving, Driver Fitness, and Controlled Substances/Alcohol–have significant influence on the endogenous variable.