Impact of Bus Routes on Crash Frequency in Metropolitan Areas

Although research in transportation safety is abundant, very few studies have examined the relationship between public transportation systems and safety performance. Most studies on the subject have focused on the impact of infrastructure countermeasures related to bus rapid transit systems. However, the impact of city-street buses on safety performance remains unknown. This research explores the pseudo-causal impact of the presence of bus routes and bus traffic on observed crash frequencies by developing safety performance functions (SPFs) that include the presence of a bus route and estimated weekly bus traffic as input variables. The SPFs were developed using the propensity score–potential outcomes (PS-PO) framework to reduce unobserved biases that might exist between segments that have and do not have bus routes. The results suggest that PS-PO reduced standardized biases significantly, allowing stronger causal inferences to be obtained. The results revealed that the presence of a bus route was associated with a 27% increase in expected crash frequency after controlling for other infrastructure-related variables. Weekly bus traffic was also found to be a significant predictor of overall crash frequency, with a 1% increase in ] weekly bus traffic associated with an expected increase in crash frequency of 0.016%. A non-parametric approach is also presented for comparison with the results from the SPFs; this confirmed the findings from the parametric method used.