Applying the Ordered Probit Model to Injury Severity in Truck-Passenger Car Rear-End Collisions

Abstract

Collisions between heavy trucks and passenger cars are a major concern because of the severity of injuries. This research has two objectives. One is to examine the impact of various factors on injuries to passenger car occupants involved in such collisions. Due to the complex interaction of factors influencing injury levels in truck-car collisions, the ordered probit model is used to identify specific variables significantly influencing levels of injury in two-vehicle rear-end involvements on divided roadways. Another objective is to demonstrate the use of the ordered probit in this complex highway safety problem. A set of vehicle, occupant, roadway, and environmental factors expected to influence injury severity was developed. Given two-vehicle passenger car-truck rear-end collisions, the variables that increase passenger vehicle occupant injury severity include darkness; high speed differentials; high speed limits; grades, especially when they are wet; being in a car struck to the rear (as opposed to being in a car striking a truck to the rear); driving while drunk; and being female. The interaction effects of cars being struck to the rear with high speed differentials and car rollovers were significant. Variables decreasing severity include snowy or icy roads, congested roads, being in a station wagon struck to the rear (as opposed to a sedan), and using a child restraint. With injuries ordered in five classes from no injury to fatalities, the marginal effects of each factor on the likelihood of each injury class are reported.Collisions between passenger cars and heavy trucks are of major concern to the traveling public and the trucking industry in the United States. This is largely due to the severity of injuries occurring in these events. The Insurance Institute for Highway Safety Fatality Facts 1996 edition reports that in 1995, tractor trailers had a higher fatal crash involvement rate [2.9 per 161 million km (100 million mi)] than passenger cars [1.9 per 161 million km (100 million mi)]. In 1995, though large trucks accounted for 3 percent of registered vehicles and 7 percent of vehicle kilometers (or miles) traveled, they were in crashes involving 12 percent of all motor vehicle deaths. Because of the importance of this problem and the complex nature of truckcar crashes, this study uses the ordered probit method for understanding the specific factors that influence injury severity in rear-end crashes involving these two vehicle types. This technique is particularly appropriate for analyzing ordinal categorical injury data, but has not been widely used in transportation safety.This paper seeks to understand how the personal attributes of passenger car occupants and drivers in combination with the vehicle, roadway, and environmental conditions influence the severity of an individual's injury in a rear-end collision involving a heavy truck. To test hypotheses empirically, the Federal Highway Administration's Highway Safety Information System (HSIS) database is used for

Road Traffic Congestion and Crash Severity: Econometric Analysis Using Ordered Response Models

“…Our results suggest that wet road surface conditions reduce the level of M25 crash severity compared with dry surface conditions. This finding is consistent with other studies (e.g., Quddus et al, 2002;Duncan et al, 1998;Shankar and Mannering, 1996). Quddus et al (2002) argued that this is likely to be an effect of reduced speed levels.…”

Identification of Factors Leading to High Severity of Crashes in Rural Areas

“…A review of the studies indicated that many researchers used ordered probit models to examine injury severities in highway crashes (O'Donnell and Connor 1996, Abdel-Aty 2003, and Duncan et al 1999. However, use of such models to examine severity of rural highway crashes is rare.…”

Injury severity analysis of accidents involving young male drivers in Great Britain

“…Injury severity probabilities for accidents involving young male drivers are also predicted using the estimated ordered probit models as suggested by O'Donnell and Connor (1996), Duncan et al (1998), andQuddus et al (2002).…”