Examining the safety impacts of narrow lane widths on urban/suburban arterials: Estimation of a panel data random parameters negative binomial model

ABSTRACT

Design practices in the United States generally allow substantial flexibility in the design of cross-sectional elements of urban and suburban arterial roadways. The operational and safety performance of such facilities is influenced by a variety of design controls, which influence the selection of important cross-sectional components, such as the number of lanes, lane widths, and the presence of shoulders, medians, or on-street parking. This study assesses the effects of lane width on the safety performance of urban and suburban arterial segments in the State of Nebraska. Ten years of traffic crash, volume, and roadway geometry data are analyzed to discern differences in the frequency of sideswipe-same direction and rear-end collisions across four functional classes of urban arterials. Crashes generally decreased as lane width increased, with 10-ft. lanes experiencing 18.1% more rear-end crashes and 27% more same-direction sideswipe collisions. In general, 12-ft. lanes showed fewer crashes than 11-ft. lanes. However, there was significant variability across the sample as 11-ft. lanes sometimes demonstrated fewer crashes, as well. These differences may be reflective of geometric characteristics that were not available in the dataset or of general differences in driver behavior across segments with varying cross-sectional characteristics.

KEYWORDS:

• lane width
• cross-section
• urban arterial
• random parameters