Evaluation of the Characteristics of Drivers with Multiple Crashes

Over the last few decades, the traffic safety community has paid. limited attention to the characteristics of drivers who have driving records with convictions and crashes. Such drivers could be generally considered as high-risk drivers. Analysis of the Kentucky Crash Database reveals that more than 10 percent of the drivers involved in a crash had been in at least two crashes during the 1989-1999 period. This finding indicates the necessity of further research efforts to improve CUlTent understanding of the human elTorissues involved in such repetitive crashes as well as the factors that may have led to those errors. On the other hand, crash predictors obtained by analyzing these multiple crashes are more reliable than those of single crashes due to the fact that involvement in a single crash does not necessarily reflect the driver's behavior as a whole because a crash is an event that can happen by chance. The objectives of this research are to establish more reliable crash predictors by analyzing those drivers' characteristics involved in multiple crashes and then to highlight critical areas for future development of potential counter-measures aiming to improve safety. A crash database over a 10- year period will be used for this analysis because crash frequencies are rather unstable between time periods, especially when they are short (1). Thus, it is expected that this research will improve the reliability of predictors beyond the current understandings by examining driver characteristics for several crashes over a longer period.