Investigating the gender differences on bicycle-vehicle conflicts at urban intersections using an ordered logit...

Area of Science:

• Traffic Safety
• Human Factors in Transportation
• Urban Planning

Background:

• Traditional crash-based models for cyclist injury risk are limited, especially for evaluating human factors and dangerous behaviors.
• Surrogate safety measures using video data and statistical analysis offer an alternative to reliance on crash data for studying bicycle-vehicle interactions.

Purpose of the Study:

• To investigate bicycle-vehicle conflict severity at urban intersections with cycle tracks.
• To evaluate the impact of cyclist gender, speed, and environmental factors on cyclist risk.
• To apply advanced statistical modeling to video-based road user interaction data.

Main Methods:

• Utilized a segmented ordered logit model to analyze post-encroachment time (PET) between cyclists and vehicles.
• Collected video data from seven intersections in Montreal, Canada, extracting 1514 road user interactions.
• Classified PET into normal interactions, conflicts, and dangerous conflicts, with independent variables for cyclist, vehicle, and environmental attributes.

Main Results:

• An ordered model effectively analyzed traffic conflicts and identified key contributing factors.
• Male cyclists were found to be less likely than female cyclists to be involved in conflicts, all else being equal.
• Bicycle and vehicle speed, and conflict timing relative to the red light phase, significantly influenced conflict severity.

Conclusions:

• Video-based analysis and ordered logit models are effective for studying traffic conflicts and understanding influencing factors.
• Exogenous segmentation aids in comparing different population segments, revealing gender differences in cyclist risk.
• Findings highlight gender disparities in cycling safety at intersections and identify critical factors for intervention.