Citation

Ahmed SS, Alnawmasi N, Anastasopoulos PC, Mannering F. Anal. Meth. Accid. Res. 2022; 34: e100213.

Copyright

(Copyright © 2022, Elsevier Publishing)

DOI

10.1016/j.amar.2022.100213

PMID

unavailable

Abstract

Over the last few decades, interstate speed limits in different US states have been increased from 70 mi/h to 75 mi/h or 80 mi/h, even to 85 mi/h in some instances. The implication of such speed limit increases on crash likelihoods and resulting injury severities is a key concern. To understand the impact of speed limit increases on no-injury and injury crash rates (including fatalities) multi-year segment-specific freeway crash data from the state of Kansas (including both pre-, and post-speed limit increase crash information) are modeled using a correlated random parameters bivariate tobit model. To address possible temporal variations in the effects of explanatory variables across years, year-specific models were estimated. Model estimation results indicate that several traffic, segment geometry, and pavement-specific characteristics affect no-injury and injury crash rates. From the year-specific model estimation results it was determined that the effects of the factors affecting pre- and post-speed limit crash rates did change significantly over time. However, such changes were also observed in the pre-speed limit increase years, as well as the post-speed limit increase years. While findings do suggest a small but statistically significant increase in injury crash rates after speed limits were raised, temporal changes in the effects of factors contributing to no-injury and injury crash rates make it difficult to isolate the true impacts of increased speed limits.

Language: en

Keywords

Bivariate Tobit; Correlated random parameters; Increasing speed limits; Injury severity rates; Out-of-sample prediction; Temporal instability