Applying Bayesian hierarchical models to examine motorcycle crashes at signalized intersections.

Area of Science:

• Traffic safety engineering
• Accident analysis
• Road safety

Background:

• Motorcycles are disproportionately involved in road traffic crashes.
• Signalized intersections pose significant risks to motorcyclists.

Purpose of the Study:

• Identify causal factors contributing to motorcycle crashes at four-legged and T signalized intersections.
• Determine the effectiveness of red light cameras in mitigating these crashes.

Main Methods:

• Utilized time-series cross-section panel data.
• Employed and compared various Hierarchical Poisson models.
• Selected the most suitable model with autoregressive lag-1 dependence specification.

Main Results:

• Increased lanes at four-legged intersections elevate crash risk due to greater exposure.
• Wide medians and uncontrolled left-turn lanes exacerbate hazards at four-legged intersections.
• Exclusive right-turn lanes and uncontrolled left-turn lanes increase crashes at T intersections.
• High-speed roadways are associated with more motorcycle crashes.
• Red light cameras significantly reduce motorcycle crashes at both intersection types.

Conclusions:

• Intersection design elements like lane configuration and median presence significantly influence motorcycle safety.
• Motorcycle vulnerability on high-speed roads necessitates specific safety considerations.
• Red light cameras enhance motorcycle safety by promoting disciplined behavior and reducing conflict exposure.