As shown in figure 1, the work was conducted in two phases. Phase I involved the development of a detailed experimental design. This was reviewed and subsequently approved by FHWA with minor modifications. In Phase II, the design was then implemented. This report describes both of these phases in detail.
Figure 1. Project workflow.
As shown in figure 1, the Phase I effort involved multiple steps to establish goals for the evaluation, collection of background information that would help shape the evaluation plan, choice of the jurisdictions to be involved in the evaluation, and definition of the data collection and statistical techniques to be used. Following are the tasks in the Phase I experimental design:
- Conduct literature review-The study team specified, obtained, and produced detailed critical reviews of key U.S. and international studies. The intent of this review was not only to summarize what is known concerning program effectiveness, but also to identify the critical experimental design factors that can overcome problems found in past studies.
- Determine study questions-The study team worked with the oversight panel established by FHWA to define and prioritize the study questions to be answered by the experimental design. The primary effort was in a workshop held early in the contract period, and the preliminary results of the literature review were used in the decisions.
- Determine RLC-related data availability-The project team conducted telephone interviews with 15 local agencies that have significant RLC programs in place. The most important questions to be answered concerned not only the details of the RLC program (e.g., number of intersections, violation definition, level of fine), but also the quality and availability of data related to before-and after-program crashes, signal data for treated and untreated intersections, traffic flows, and public information efforts.
- Specify a multijurisdictional experimental design to estimate RLC safety effects-Based on the outputs from the preceding tasks, the study team defined the multijurisdictional experimental plan, including data needs, the most appropriate local agencies to work with, data analysis techniques, and the specification of which of the desirable study questions might be answerable.
- Specify an experimental plan for examining the effects of RLC on economic costs of related crashes-Because RLCs are likely to decrease more severe angle crashes and increase less severe rear end crashes, the study team also defined a study method which can combine both crash frequency and severity in terms of economic costs.
As the task list shows, there were actually two experimental plans or methodologies developed in this Phase I effort. The first is a national design in which data from multiple jurisdictions across the United States were used to both define the crash-related effectiveness of RLC programs and to explore factors that could make such programs more beneficial in terms of crash reduction (e.g., signalization variables such as clearance intervals and cycle lengths, and public information programs and signage related to RLC programs). The primary outcome variable of interest here is RLC-related change in the frequencies of different types of crashes. The second experimental plan involved the development of a database and methodology to examine, in economic terms, costs and benefits of RLCs, thus simultaneously analyzing changes in crash severity and crash frequency. The following sections describe the development of this economic analysis methodology initiated in Phase I and revised and completed in Phase II; The full development process is described in the Phase II section.
After FHWA agreed with the Phase I evaluation design, funding was provided for the implementation of the evaluation efforts. The project team then performed the following steps:
- Collect and code data from jurisdictions-Project staff established contacts with and visited the seven jurisdictions chosen for participation. Staff collected data related to intersection geometry, traffic flow, signalization parameters, and other descriptive data for treatment, signalized reference, and unsignalized reference intersections in each jurisdiction. The data, almost always in noncomputerized form, were then collected for data extraction, coding, and analysis file preparation. Because crash data for a multiyear before-period were not available in six of the seven jurisdictions, crash data were located and extracted from State data files.
- Develop new crash cost estimates-The final economic analysis methodology required the development of human capital and comprehensive cost estimates for each crash severity level in each pertinent crash type (e.g., the cost of a "moderate injury" in a rear end collision at a signalized intersection). Because FHWA cost data had not been updated since 1994, and because even those older data were not specific to crash and location types, a subcontractor with extensive experience in highway safety economics joined the team to develop these estimates.
- Analyze data-Following the national-study design and the crash-cost methodologies developed in Phase I, the project team analyzed both RLC-related crash frequency and crash cost in each of the seven jurisdictions. The effectiveness estimates from all jurisdictions were then combined to develop a national estimate of the effectiveness of RLC system.
- Prepare reports-Based on the analysis results, project staff then prepared this FHWA final report, an FHWA technical summary, and two Transportation Research Board (TRB) journal articles (one on the crash effects analysis, the other on the analysis of economic effects and the identification of factors associated with the greatest economic benefits).(2,3)
The narrative following the literature review will describe each of the steps in both Phase I and Phase II in more detail.
Previous | Table of Contents | Next
FHWA-HRT-05-048
