Loading…

Safety Analysis of Freeway Segments with Random Parameters

The purpose of this study was to analyze the effect of geometric features on freeway crashes while accounting for the effect of unobserved factors likely to influence crash occurrence. The investigation was also motivated by the reality of the distribution of crashes in space not being limited to on...

Full description

Saved in:
Bibliographic Details
Published in:Transportation research record 2015-01, Vol.2515 (1), p.78-85
Main Authors: Mulokozi, Eneliko, Teng, Hualiang (Harry)
Format: Article
Language:English
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The purpose of this study was to analyze the effect of geometric features on freeway crashes while accounting for the effect of unobserved factors likely to influence crash occurrence. The investigation was also motivated by the reality of the distribution of crashes in space not being limited to only the influence areas of the divergence and convergence segments and the weaving segments. Areas beyond the influence areas were observed to have crashes, and those areas included, data in the weaving and nonweaving segments can be clustered to quantify the variability of unobserved factors through the variance of random parameters using multilevel count models. The model results indicated that 13.9% of the variation in crash frequency was unaccounted for; this finding indicated the existence of unobserved factors influencing the occurrence of crashes. It was also revealed that weaving segments had the highest between-segment variance compared with nonweaving segments. In addition, it was revealed that more vehicles and short segments increased crash frequency while a wider right shoulder decreased crash frequency. It was also found that weaving segments decreased crash frequency compared with nonweaving segments. These results indicate that by allowing parameters to vary across segments, it is possible to capture and quantify unobserved factors. Ignoring these factors results in biased coefficients in a multilevel setting because the estimate of the standard errors will be incorrect.
ISSN:0361-1981
2169-4052
DOI:10.3141/2515-11