The American Association of State Highway and Transportation Officials (AASHTO) Highway Safety Manual (HSM) includes methods for crash prediction for a variety of roadway facility types and site conditions. The current methods account for traffic volumes and other roadway characteristics but not for vehicle mix or the distribution of vehicle types. Recent studies suggest that heavy traffic adversely impacts roadway safety performance and consideration of vehicle mix would improve predictive methods for estimating crash frequency and severity. 

Under NCHRP Project 22-49, the University of Central Florida was asked to develop and test a statistically valid predictive methodology to quantify the effect of vehicle mix on crash frequency and severity for several facility types. The research team developed spreadsheet tools for practitioners to quantify the effect of vehicle mix on safety performance and analyzed two alternative statistical approaches. The research team calibrated the models using existing HSM methods to serve as a benchmark for the proposed alternative approaches. The team used predictive metrics to compare model performance for the estimation and validation of datasets and selected a single preferred model system for each facility group. This project included analysis of two new facilities, rural arterial 3-lane, and rural arterial 5-lane roadway segments, not included in the current HSM but exhibited a significant number of crashes across the states.

To aid practitioners with crash frequency and severity analysis using the new models, a final report NCHRP 1103: The Effect of Vehicle Mix on Crash Frequency and Crash Severity,  NCHRP Web-Only Document 393:User’s Guide for Quantifying the Effects of Vehicle Mix on Crash Frequency and Crash Severity, a PowerPoint presentation,  and a suite of crash frequency and severity prediction Excel spreadsheet tools are accessible by searching the National Academies Press website.