Differences in road-traffic crash rates during construction and non-contruction times on arterial streets: A comparative statistical analysis

Abstract

Different studies over the past 30 years have shown an increase in the rates of crash frequencies during road construction time, but this trend is not reported as valid for all cities. A question is raised whether higher rates are observed in arterial roads in Bogotá, Colombia. It is possible to apply descriptive statistics and hypothesis tests to prove them and identify the variables that affect the accident rate during road construction. Our research aims to verify the incidence of high impact construction zones, in the crash rate at the arterial road network of Bogotá. We use descriptive statistics and inferential statistical tests to analyze whether crash rates are statistically higher during construction time than during nonconstruction time at the same highway sections considering different crash severities (damages-only, injuries, and fatalities). Our database considered 871 road links that make up 68 artery corridors for the city of Bogota, and 5.450 road construction zones, from 2015 to 2019. An analysis by corridor was performed, in which we identified seven patterns in the behavior of accident rates influenced by the presence of a road intervention in the corridor. Within the patterns, it was evident that some corridors reported an increase in the accident rate during the time of construction, while others showed a decrease in the same comparison. We used the Wilcoxon test to establish the statistical significance of our conclusions, with a significance level of 10%. We also found that those construction interventions that do not require excavation of more than half a meter there was a decrease in the accident severity, as damages-only crashes, diminished during the construction time, while for those interventions that include excavation greater than half a meter there was an increase in overall accident rates during construction time.