Considering the deadly impacts of commercial motor vehicle (CMV) crashes on road safety, many previous studies have put in efforts to find the determinants of CMV crash injury severities. However, studies examining the impact of the different number of lanes on CMV crash
injury severities are lacking. Besides, the importance of addressing unobserved heterogeneity (variations in effects of variables across observations by unobserved factors) and temporal instability (unstable effects of variables over time) have been increasingly emphasized in crash
data analysis. Thus, this thesis aims to bridge the research gap by investigating crash injury severities of CMV drivers for the different number of lanes while accounting for unobserved heterogeneity and temporal instability.
This thesis consists of two manuscripts. The first manuscript investigated injury severities of CMV drivers based on the number of lanes by developing four separate models (2, 4, 6, and 8/10 lanes, in both directions). Using Oregon crash data between 2013 and 2019, the random
parameters logit approach with heterogeneity in means and variances was applied to capture potential unobserved heterogeneity. The model outcomes and likelihood ratio test results indicated significant differences in the injury severity determinants by the number of lanes.
On the basis of the findings from the first manuscript, the second manuscript examined temporal instability in injury severity determinants of CMV drivers for each number of lanes. Using the same data source and methodology as the first manuscript, the data of 2-lane, 4-lane, and 6-lane roads were further split into four temporal subgroups (2013–2014, 2015–2016, 2017–2018, and 2019), which ended up with a total of twelve separate models (4 time periods × 3 numbers of lanes). The model outcomes and likelihood ratio test results showed that the injury
severity determinants are significantly different not only for the different number of lanes but also for the different time periods.
Combining the findings from both manuscripts, this thesis provides a better understanding of CMV crash injury severities in terms of the number of lanes. From a research standpoint, this thesis extended the existing literature by focusing on the impact of the number of lanes on the
CMV driver’s injury severities while capturing unobserved heterogeneity and temporal instability. From a practical standpoint, this thesis identified a wide range of injury severity determinants in CMV crashes by the number of lanes, with special emphasis on the determinants for severe
injuries. Road agencies, CMV companies, and safety practitioners may use these findings to build up effective safety strategies to reduce the risks of CMV crashes.