Impacts of traffic interventions on road safety : an application of causal models
File(s)
Author(s)
Li, Haojie
Type
Thesis or dissertation
Abstract
This thesis is concerned with the causal relationship between traffic interventions and road safety. It focuses on two issues that have been overlooked in the existing empirical literature: the establishment of a causal link between traffic interventions and road traffic accidents, and the application and development of formal causal approaches, which have not yet been applied in the field of road safety.
In the past decades substantial studies have been conducted to investigate the risk factors contributing to road accidents. It has been shown that the frequency and severity of road accidents are associated with various factors, including traffic characteristics, road environment and demographic characteristics. However, the existence of a causal link between traffic interventions and road accidents remains unclear due to the complex character of traffic interventions. Meanwhile, the lack of formal causal models makes it difficult fully to address issues such as confounding effects and regression to the mean bias.
This thesis begins by reviewing and discussing different types of traffic interventions in order to demonstrate the chains through which traffic interventions are related to road safety. To address the shortcomings in empirical literature, three models for causal inferences are discussed: the difference-in-difference method, the propensity score matching method and Bayesian methods.
These formal causal approaches are then applied to three empirical studies: the London congestion charging scheme, speed limit enforcement cameras, and the road network design. The conventional models are also employed and compared with formal causal models.
In the past decades substantial studies have been conducted to investigate the risk factors contributing to road accidents. It has been shown that the frequency and severity of road accidents are associated with various factors, including traffic characteristics, road environment and demographic characteristics. However, the existence of a causal link between traffic interventions and road accidents remains unclear due to the complex character of traffic interventions. Meanwhile, the lack of formal causal models makes it difficult fully to address issues such as confounding effects and regression to the mean bias.
This thesis begins by reviewing and discussing different types of traffic interventions in order to demonstrate the chains through which traffic interventions are related to road safety. To address the shortcomings in empirical literature, three models for causal inferences are discussed: the difference-in-difference method, the propensity score matching method and Bayesian methods.
These formal causal approaches are then applied to three empirical studies: the London congestion charging scheme, speed limit enforcement cameras, and the road network design. The conventional models are also employed and compared with formal causal models.
Version
Open Access
Date Issued
2013-08
Date Awarded
2013-12
Advisor
Graham, Daniel
Majumdar, Arnab
Sponsor
China Scholarship Council
Lloyds Register Foundation
Publisher Department
Civil and Environmental Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)