How is Highway Safety Manual methodology applied for safety analysis?

HSM methodology uses Empirical Bayes (EB) combining observed crash history with predicted crashes from Safety Performance Functions (SPFs): Expected crashes = w * SPF prediction + (1-w) * observed, where w depends on SPF reliability. SPFs are regression models relating crashes to exposure (ADT) and site characteristics. Crash Modification Factors (CMFs) quantify expected change from treatments. Process: network screening to identify sites, diagnosis to understand crash patterns, countermeasure selection, and economic evaluation (B/C ratio). HSM provides predictive method for new designs where crash history unavailable.