Résumé
This dissertation expands upon the Master's Thesis written by this author. It begins with an in-depth literature review of applicable social theories. These set the context for studying the phenomena of speeding, automobile collisions, drivers ' perceptions of risk and their own skill, and the method by which the criminal justice system, through the police services, attempts to manage road safety. Following the development of several social theories, this dissertation reviews the more pragmatic and scientific literature. This review examines surveys and experiments in situ. The two reviews present a solid theoretical and practical base from which this dissertation launches into an exploration of traffic collision data. The data is organized within a Geographic Information System (GIS) framework The data is scrutinized using several exploratory data analysis (EDA) techniques then logistic regression is utilized to search for multivariate relationships. Time series analysis is applied to traffic flow data so that a weekly average of time flow may be used to examine collisions in a temporal fashion. Breaking apart the data into this time series then arranging it spatially in the GIS allows the researcher to display and examine traffic collisions in the spatial and temporal dimensions. This presents a unique view of the risk characteristics presented at intersections and interchanges throughout the city. Furthermore, the citywide holistic approach to investigating collision data across the entire roadwork also adds to the understanding of collision occurrence. A methodology for determining hazardous locations is developed and refined so that selected sites can be scrutinized in greater detail. Suggestions are offered for continued research into this area.