Investigating factors affecting bus/minibus accident severity in a developing country for different subgroup datasets characterised by time, pavement, and light conditions

Developing countries are primarily associated with poor roadway and lighting infrastructure challenges, which
has a considerable effect on their traffic accident fatality rates. These rates are further increased as bus/minibus
drivers indulge in risky driving, mainly during weekends when traffic and police surveillance is low to maximise
profits. Although these factors have been mentioned in the literature as key indicators influencing accident
severity of buses/minibuses, there is currently no study that explored the complex mechanisms underpinning the
simultaneous effect of pavement and light conditions on the generation of accident severity outcomes while
considering weekly temporal stability of the accident-risk factors. This study seeks to investigate the variations in
the effect of contributing factors on the severity of bus/minibus accidents in Ghana across various combinations
of pavement and light conditions and to identify the exact effects of weekdays and weekends on severity outcomes
using a random parameter ordered logit model with heterogeneity in the means to account for unobserved
heterogeneity in the police-reported data. Preliminary analysis demonstrated that accident-risk factors used in
the models were temporally unstable, warranting the division of the data into both weekend and weekday time periods.
A wide variety of factors such as sideswipes, median presence, merging, and overtaking had significantly
varying effects on bus/minibus accident severities under different combinations of pavement and light conditions
for both weekdays and weekends. Insights drawn from this study, together with the policy recommendations
provided, can be employed by engineers and policymakers to improve traffic safety in developing nations.