Analysis of Injury Severity of Drivers Involved in Single-Vehicle and Two-Vehicle Crashes on Ontario Highways Using Heteroscedastic Ordered Logit Models

摘要

The objective of this study is to analyze driver’s injury severity in single-vehicle and two-vehicle crashes and compare the effects of explanatory variables between various types of crashes. The study identified factors affecting injury severity and their effects on severity levels using 5-year crash records for provincial highways in Ontario, Canada. Considering non-uniform variations in unobserved effects of explanatory variables on injury severity among observations called “heteroscedasticity”, heteroscedastic ordered logit (HOL) models were developed for singlevehicle and two-vehicle crashes separately. The results show that there exists heteroscedasticity for some variables in both single-vehicle and two-vehicle crash models. The results also show that some factors have opposite effects between single-vehicle and two-vehicle crashes, and between car-car crashes and truck-truck crashes. The study demonstrates that HOL models using separate crash data sets classified by vehicle type can better capture the associations of variables with driver’s injury severity.