BEYOND HUMAN FACTORS : EXAMINING THE UNDERLYING DETERMINANTS OF RECREATIONAL BOATING ACCIDENTS WITH SPATIAL ANALYSIS AND MODELING

Abstract

Recreational boating has grown in popularity in recent decades, accompanied with increased accidents resulting in property damage and personal injury. Some 5,000 recreational boating accidents are reported annually, ranking recreational boating as a leading cause of transportation accidents, second only to automotive. Recent research suggests that recreational boating accidents stem from multiple factors. In contrast, public perception and public policy overwhelmingly attribute boating accidents to human error, e.g., operator drug or alcohol use or lack of experience. This dissertation offers a comprehensive perspective on recreational boating accidents by exploring human, technological, and environmental factors that most influence these accidents. This level of inclusiveness is absent from previous research. The conceptual model developed in this dissertation is derived from general accident theory that integrates spatial and temporal qualities of recreational boating (and boating accidents) from satellite imagery, on-the-water boater surveys, and federal boating accident data. Data were assembled for two distinctive research sites, Sandusky, OH and Tampa, FL. Analyses of these data depended, in part, upon various forms of spatial statistics, e.g., hot spot analyses. The boating accident model developed here uses the multivariate negative binomial model to analyze accident count data aggregated to 0.25 mi² grid cells. The result is a synthetic model with improved parameter estimates and predictive capability compared to previous boating accident research. Key risk factors contained in the final model clearly represent human (operator experience), technological (boat speed and length), and environmental (boat density and channel character) dimensions. This research has important societal impact, i.e., to public officials faced with the allocation of limited resources. In particular, this research emphasizes the concentrated nature of boating risk in time (seasonality, day of week, time of day) and in space (shoals, channels, fixed facilities). These features should guide the timing and the placement of mobile law enforcement capacity as well as the location of operation centers near high risk boating sites. Finally, this work emphasizes the need for investigations of additional sites and the importance of including remotely sensed data to complement survey data in studies of recreational boating accidents.