Accident Analysis and Prevention 40 (2008) 1674–1682 Contents lists available at ScienceDirect Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap Analysis of left-turn crash injury severity by conflicting pattern using partial proportional odds models Xuesong Wang ∗ , Mohamed Abdel-Aty Department of Civil & Environmental Engineering, University of Central Florida, Orlando, FL 32816-2450, United States article info Article history: Received 19 October 2007 Received in revised form 24 May 2008 Accepted 2 June 2008 Keywords: Signalized intersection Left-turn crash Conflicting pattern Crash injury severity Partial proportional odds model Significant factor abstract The purpose of this study is to examine left-turn crash injury severity. Left-turning traffic colliding with opposing through traffic and with near-side through traffic are the two most frequently occurring con- flicting patterns among left-turn crashes (Patterns 5 and 8 in the paper, respectively), and they are prone to be severe. Ordered probability models with either logit or probit function is commonly applied in crash injury severity analyses; however, its critical assumption that the slope coefficients do not vary over dif- ferent alternatives except the cut-off points is usually too restrictive. Partial proportional odds models are generalizations of ordered probability models, for which some of the beta coefficients can differ across alternatives, were applied to investigate Patterns 5 and 8, and the total left-turn crash injuries. The results show that partial proportional odds models consistently perform better than ordered probability models. By focusing on specific conflicting patterns, locating crashes to the exact crash sites and relating approach variables to crash injury in the analysis, researchers are able to investigate how these variables affect left-turn crash injuries. For example, opposing through traffic and near-side crossing through traffic in the hour of collision were identified significant for Patterns 5 and 8 crash injuries, respectively. Protected left-turn phasing is significantly correlated with Pattern 5 crash injury. Many other variables in driver attributes, vehicular characteristics, roadway geometry design, environmental factors, and crash charac- teristics were identified. Specifically, the use of the partial proportional formulation allows a much better identification of the increasing effect of alcohol and/or drug use on crash injury severity, which previously was masked using the conventional ordered probability models. © 2008 Elsevier Ltd. All rights reserved. 1. Introduction Intersections are among the most dangerous locations of a road- way network. In the state of Florida, 43.1% of fatalities and serious injuries occurred at or were influenced by intersections (Florida Department of Transportation, 2006). In the U.S., although only around 10% of all intersections are signalized, in 2005, nearly 30% (2744) of intersection fatalities occurred at signalized intersections (Rice, 2007). Left-turn crashes occur frequently and they account for a high percentage of total crashes at signalized intersections. They are prone to be severe, possibly due to the relatively high con- flicting speeds of involved vehicles and the angle of impact. In a sample of signalized intersections collected in Orange and Hillsbor- ough counties in Florida, 64.2% of left-turn crashes involved injury, ∗ Corresponding author at: Department of Civil & Environmental Engineering, University of Central Florida, Engr II, Room 301B, Orlando, FL 32816-2450, United States. Tel.: +1 407 8234902; fax: +1 407 8234676. E-mail address: xuewang@mail.ucf.edu (X. Wang). whereas the percentage of injury crashes was only 50.1% for all other crashes. From 2002, a series of crash frequency studies have been conduced in Florida to identify the crash profiles for the major inter- section types (Abdel-Aty and Wang, 2006; Wang and Abdel-Aty, 2006, 2007, 2008; Wang et al., 2006). In one study, Wang and Abdel- Aty (2008) investigated conflicting flows, intersection geometric design features, and traffic control and operational features on left- turn crash occurrence. Left-turn crashes were classified into distinct conflicting patterns (i.e., left-turn traffic colliding with opposing through traffic, or with near-side through traffic, etc.), and then the crash frequencies of different patterns were modeled. The studies indicate there are obvious differences in the factors which corre- lated with different left-turn collisions. However, crash frequency studies model accumulated crash counts, which ignores the dif- ference of severe and minor crashes. Therefore, they are unable to investigate how specific features affect crash injury severity. The left-turn crashes at signalized intersections result in a huge cost to society in terms of death, injury, lost productivity, and prop- erty damage. However, how the different factors affect left-turn 0001-4575/$ – see front matter © 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.aap.2008.06.001