Modeling the traffic injury severity for drivers and passengers with different seat positions in electric vehicles: a hierarchical data-matching approach

This study examines the association between occupant seating position and injury outcomes in Electric Vehicles (EVs) and Internal Combustion Engine Vehicles (ICEVs) to inform vehicle safety design. Using Pennsylvania crash records from 2021–2024, we applied matched-pair sampling combined with a Bayesian hierarchical ordered logit model to quantify injury risks across seating positions. The analytical sample comprised 3,806 EV occupant records from 1,559 vehicles and 42,481 ICEV occupant records from 17,398 vehicles. In addition, airbag deployment status was used to define crash-context subgroups, allowing an exploratory analysis across these contexts. Seating position was associated with occupant injury severity. Compared with drivers, EV right-front occupants had higher probabilities of possible injury, suspected minor injury, and serious injury by 0.82 (95% CrI: 0.47–1.20), 3.31 (95% CrI: 1.91–4.81), and 0.37 (95% CrI: 0.20–0.55) percentage points, respectively. For EV rear-seat occupants, the corresponding increases were 1.03 (95% CrI: 0.59–1.47), 4.23 (95% CrI: 2.40–6.12), and 0.48 (95% CrI: 0.26–0.74) percentage points. ICEVs showed a similar pattern, with right-front and rear-seat occupants having higher probabilities of possible injury, suspected minor injury, and serious injury than drivers. Second-difference results indicated that seating-position effects were greater in EVs than in ICEVs. For right-front seating, the effects were 0.54 (95% CrI: 0.19–0.93), 2.07 (95% CrI: 0.63–3.60), and 0.20 (95% CrI: 0.02–0.38) percentage points greater in EVs than in ICEVs for possible injury, suspected minor injury, and serious injury. For rear seating, the corresponding differences were 0.69 (95% CrI: 0.25–1.15), 2.73 (95% CrI: 0.87–4.65), and 0.27 (95% CrI: 0.04–0.53) percentage points. Stratified analyses further showed that the second differences in seating-position effects between EVs and ICEVs were evident when airbags were deployed. For right-front seating, the second differences were 0.59 (95% CrI: 0.18–1.09) percentage points for possible injuries, 5.65 (95% CrI: 1.70–9.75) percentage points for suspected minor injuries, and 1.29 (95% CrI: 0.38–2.29) percentage points for serious injuries. For rear seating, the corresponding second differences were 0.65 (95% CrI: 0.15–1.22), 7.24 (95% CrI: 1.77–12.50), and 1.75 (95% CrI: 0.34–3.42) percentage points for possible, suspected minor, and serious injuries, respectively. In contrast, no significant differences in seating-position effects between EVs and ICEVs were observed when airbags were not deployed. These findings provide evidence-based support for enhancing occupant protection strategies, particularly for rear-seat passengers in EVs. The results highlight the need for targeted improvements in rear-seat safety design.