Based on the historical travel data provided by commercial driver monitoring systems, this study presents the analysis of contributing factors affecting hazardous materials (HAZMAT) truck safety. A comprehensive dataset of HAZMAT truck travel records (n=4,487), including driver characteristics, roadway characteristics, environmental characteristics, and temporal characteristics was developed for the study period of July 1^st to July 31^st, 2020. A random parameter ordered probit model was established for investigating the HAZMAT truck travel risk levels, which were measured by three warning frequency categories: high risk (more than one warning), moderate risk (one warning), and low risk (zero warning). The results show that speed, acceleration, lighting conditions (dusk/dawn, dark-lighted and dark-unlighted), visibility (less than 500 m and between 500 m and 1000 m), weather (rain and fog), and time of day (0:00-6:00) have statistically significant effects on the increase of high-risk probability.

HAZMAT truck;Risk factor;Random parameter ordered probit model