Evaluating the pandemic’s impact on speeding-related injury severity on Oregon’s rural roads: a mixed logit analysis with temporal constraints

Abstract

During the COVID-19 pandemic, Oregon experienced fewer crashes and a 10.77% decline in vehicle miles traveled, yet fatal crashes increased. This study examines factors influencing injury severity in speeding-related single-vehicle passenger car crashes on rural Oregon roads and their temporal instability during the pandemic. We analyzed police-reported data from March 2019 to 2022 using a partially temporally constrained mixed logit model with heterogeneity in means, distinguishing severe injury, minor injury, and property damage only. Likelihood ratio tests assessed model validity. Invalid or revoked licenses, non-Oregon residence, and tree involvement showed consistent effects across periods, whereas effects associated with female drivers, alcohol involvement, ditch involvement, and reckless driving varied. Out-of-sample predictions from the base-year model revealed probability differences in subsequent years, supporting pandemic-related changes in injury outcomes. These findings suggest that transportation planners and policymakers may need to adapt countermeasures to address evolving crash risks during pandemics and other atypical events.