Joint modeling of collision type and driver injury severities in truck-involved crashes: A copula-based multivariate approach

Despite the vital role of trucks in freight transport, crashes in which they are involved often result in severe consequences. It is therefore crucial to identify the factors that influence the risk of truck-related crashes and resulting injuries to formulate effective fleet management and transport policy strategies. However, previous studies have often overlooked two major issues: interdependent correlations in driver injury severity within the same crash, and the endogenous influence of collision type on driver injury severity. Overlooking these issues can result in biased parameter estimation. Therefore, this study develops a copula-based multivariate model to jointly examine the relationships among collision type and driver injury severities in truck-involved crashes. Based on crash data from North Carolina for the period 2016–2019, the proposed model simultaneously measures the associations between potential risk factors and three dependent variables: collision type, car driver injury severity, and truck driver injury severity. The results indicate remarkable differences in how various factors affect the injury severity of truck drivers versus car drivers. The model also successfully accounts for the endogenous effect of collision type on driver injury severity using a recursive structure, thereby improving the precision of parameter estimation. Furthermore, the findings show that the proposed copula-based model effectively captures the interdependent correlations with superior interpretability, compared to conventional models. Based on these findings, decision-makers in the trucking industry can develop more effective countermeasures to reduce the injury risk of all drivers in truck-involved crashes. This can, in turn, enhance the overall safety and operation performance of freight logistics.