|
Copyright
|
(Copyright © 2006, Transportation Research Board, National Research Council, National Academy of Sciences USA, Publisher SAGE Publishing) |
|
Abstract
|
|
To date, there has been little examination of the area of roadside safety on nonfreeway urban roads. To understand better the design of safe roadsides in urban environments, this study used negative binomial regression models to examine the safety effects of three roadside design strategies: widening paved shoulders, widening fixed-object offsets, and providing livable-street treatments. The model results indicated that of the three strategies, only the livable-streets variable was consistently and negatively associated with reductions in roadside and midblock crashes. Wider shoulders were found to increase roadside and midblock crashes, while unpaved fixed-object offsets had a mixed safety effect by decreasing roadside crashes but having a slightly positive effect on midblock crashes. To understand better the reasons for these findings, this study then examined roadside crash site locations for tree and utility pole crashes. It found that the majority (between 65% and 83%) did not involve random midblock encroachments, as currently assumed, but instead involved objects located behind both driveways and side streets along higher-speed urban arterials. Collectively, these findings suggest that most urban roadside crashes are not the result of random error but are instead systematically encoded into the design of the roadway. The study concluded by distinguishing between random and systematic driver errors and by discussing strategies for eliminating systematic error while minimizing the consequences of random error. |
|