Calibration and Development of Safety Performance Functions for Alabama

Abstract

One critical component of the Highway Safety Manual (HSM) statistical methods is the safety performance function (SPF). SPFs are essentially regression models that correlate quantitatively the expected number of crashes with traffic exposure and geometric characteristics of the road. As part of a project performed by the University of Alabama to facilitate implementation of the new HSM procedures in the state, this study aims to evaluate the applicability of HSM predictive methods to Alabama data and to develop state-specific statistical models for two facility types: two-lane, two-way rural roads and four-lane divided highways. This study first calibrates HSM base SPFs by using two approaches: the method recommended by the HSM and a newly proposed approach that treats the estimation of calibration factors as a special case of a negative binomial regression. In addition, new forms of state-specific SPFs are further investigated by using Poisson-gamma regression techniques. Four new functional forms are studied in this project. The prediction capabilities of the two calibrated models and the four newly developed state-specific SPFs are evaluated with a validation data set. Five performance measures are considered for model evaluation. The study is able to identify a particular state-specific SPF that fits the Alabama data well and outperforms other models, including the calibrated SPFs. The best model describes the mean crash frequency as a function of annual average daily traffic, segment length, lane width, year, and speed limit. The study finds that the HSM-recommended method for calibration factor estimation also performs well.