Safety impacts of pavement surface roughness at two-lane and multi-lane highways: accounting for heterogeneity and seemingly unrelated correlation across crash severities

Abstract

Highway agencies continue to show interest in measuring pavement condition effects on safety. This paper estimates univariate negative binomial (UNB) and random-parameters seemingly-unrelated negative binomial (RPSUNB) regression models. The latter account for unobserved heterogeneity and correlation in crash frequencies across the crash severity levels. The analysis was carried out for two-lane and multi-lane highways, and the results suggest that at the latter, the pavement condition generally has a far more significant safety impact compared to the former. This could be due to risk compensation effects where drivers offset the safety hazard associated with inherently less safe situations by driving more carefully. It was determined that compared to UNB, RPSUNB models have superior efficacy in addressing seemingly unrelated correlations among the crash severity levels.