This study aims to investigate the role of the of truck volumes of various classes on the severity of non-truck crashes, with which the correlations between space-time-varying heterogeneities are considered. Fixed-, uncorrelated and correlated random-parameter binary logit models are established, based on crash and traffic data of five freeway segments in Shandong of China during 2016 to 2019. In total, 4,008 crashes on 5 freeways are extracted from the database and stratified into space-time panels. Results indicate that the proposed correlated random-parameter model is the optimal model among the three. Correlation in the heterogeneous effects between super-large truck volume and average speed is significant. The increase in mid-sized truck volume is associated with the increase in the crash severity of non-truck involved crashes, which gets weakened along the increasing total traffic volume.

CICTP