Statistics consistently indicate that red-light running has become a significant safety problem throughout the United States. Comprehensive guidelines for treating red-light running at problem intersections have been developed. Unfortunately, these guidelines do not include a tool or technique for quantitatively determining if a problem exists and if a countermeasure is truly effective. The objective of this work is to describe the development and calibration of such a tool. The calibrated prediction model developed for this research indicates that red-light running increases with flow rate, speed, and dense platoons arriving at the end of the phase. It was also found that red-light running decreases with increasing cycle length and cross-street width, and when back plates are used on the signal heads. Uses for the calibrated model are described.