Framework for Applying Highway Safety Manual Methods during Project Alternatives Analysis

Abstract

The Highway Safety Manual (HSM) provides methods for transportation agencies to quantify and compare safety performance of project alternatives for estimated average crash frequency (CF). When applicable, HSM promotes use of the empirical Bayes (EB) method to combine predicted average- and observed CF to obtain a more reliable estimate of expected average CF. If the EB method does not apply to one or more alternatives, HSM recommends not using EB within the Part C predictive method and relying solely on predicted crashes for comparing alternatives. This has led agencies to avoid using the EB method, including for “future no-build” scenarios. Other than EB, HSM does not recommended an approach for accounting for historic crashes under future scenarios. This may result in missing critical insights from analysis sites that may influence future crashes. Our objective was to consolidate best practices and research findings to develop a reliable, consistent approach supporting project alternatives analysis using HSM-recommended methodologies. This approach provides a method for considering historic crashes when facility types may differ and EB is not applied. We provide a five-step alternatives analysis approach prioritizing methods based on reliability and consistency: it establishes an estimated baseline average annual CF for the no-build condition in the design year, then recommends identifying appropriate crash modification factors (CMFs), including pseudo-CMFs based on the HSM predictive method. The results highlight the importance of calibrating HSM safety performance functions (SPFs) or developing jurisdiction-specific SPFs. This research demonstrated a need for consistency when applying EB based on external- and pseudo-CMFs, particularly when evaluating whether historic crash data are applicable to project alternatives that do not change HSM facility type.