Classifying urban crashes for countermeasure development.

Abstract Efforts to reduce urban crash rates have been hampered by a lack of information about motor vehicle crash types. The present study is based on a systematic sample of 4,526 police crash reports from four urban areas. The sample was weighted to give each area equal representation. Diagrams and narrative descriptions from each report were reviewed, and the most common crash types based on precrash driver/ vehicle behavior, were identified. Fourteen crash types were defined, and five of these were found to account for 76% of all crash events and 83% of injury crashes. Although the rank order of the five types differed from city to city, they accounted for the vast majority (69%–81%) of the crashes in each. Potential countermeasures are discussed based on the predominant crash types identified in this study. For example, ran traffic control crashes, the most common type, might be reduced by changes in signal timing, providing all-red signal intervals, increasing sign visibility, and increasing sight distances.