Safety data can be collected from a number of sources. Excellent discussions of these may be found in Ogden, 1996, Chapter 4, and Pfefer, et al, 1998. Ogden identifies the following sources of safety data:
For each reported crash
Accident file
Vehicle file
Roadway file
Commercial vehicle file
Citation/conviction file
Emergency medical services file
Ogden also identifies supplementary sources of data to consider, especially when the more traditional sources are not available (e.g., crash reports, roadway inventory):
Local knowledge (public or government employees)
Maintenance reports
Interviews with road users
Special surveys (including support systems such as law enforcement and adjudication agencies)
Traffic conflict surveysCoroner’s reports
Site investigations
Details on the conduct of these types of studies may be found in several references. For example: see Robertson, et al (1994); FHWA (1981); Bowman (1986); and FHWA (December 1981), the latter of which lists and describes19 types of studies.
A review of these lists leads to the quick conclusion that the collection and use of safety data requires the involvement of a number of agencies and offices working in coordination. This rarely happens. So, the analyst is left either with a daunting task to assemble it all, or will have to work without some of it. Even crash reports, the mainstay of highway safety analysis, are not always available for low-type roads in a jurisdiction. If the analysis is of a major portion of a system in a jurisdiction, it may be infeasible to collect supplementary data. The analyst may have to rely solely upon archived data. . Since studies covering facilities in an entire jurisdiction are often the initial step in successively more detailed definitions of the problem, it makes economic sense to require supplementary data collection when the study of a potential problem is narrowed to a limited portion of the system, or a relatively few locations. When studies are made of a limited number of sites, consideration may be given to taking a structured approach of the type used under the rubric of "safety audits." The types of analyses that may be needed, whether in the context of a safety audit, or otherwise, are likely to require the involvement of several disciplines, since one is considering factors involving highways, humans, environment, and vehicles. Therefore, the involvement of more than one discipline may be desirable, to carry a truly thorough study. For example, even though the emphasis area being studied deals with driver behavior, it is not considered sufficient to involve only traffic law-enforcement personnel. The inclusion of persons with human factors and highway engineering backgrounds is also desirable. In general, consideration should be given to the inclusion of the following disciplines when developing a program:
Traffic Safety Specialists
Highway engineering
Vehicle engineering
Systems engineering
Environment
Traffic law enforcement
Human factors
Public Information
Education
Medicine
Statistics
For a presentation of users and uses of safety data, see Pfefer et al (1998), pp 38-44.
Exposure data are one of the more important types needed in safety analyses. Frequency information does not provide a sufficient "story." Numerous research studies have shown the strong correlation between traffic volume and crashes. Therefore, a large frequency of crashes may be associated with a high volume of traffic, rather than other factors of interest. Accounting for exposure, therefore, is an important part of a safety analysis. Traffic volume is the most commonly used example of exposure data. It is used to arrive at crash rates (e.g., crashes per million vehicle-miles, or intersection crashes per 100,000 entering vehicles). Other measures of exposure are sometimes used. An example is the number of conflicts of specified type in a time period. The primary source of volume data in most jurisdictions is a roadway-inventory file. However, these data often have severe limitations.
For drivers, exposure is often measured in terms of annual miles driven. Alternatives include information about the percentage of driving done, by type of driver (to be compared with percentage of drivers of the specified type that are involved in the crashes being studied). The latter is especially useful when analyzing crashes by time periods within the day (e.g. nighttime crashes involving elderly drivers).
Sources of exposure data are identified and discussed in Campbell, et al (1997).
B. Bowman, Local Highway Safety Studies – User’s Guide, National Highway Institute, FHWA, July 1986.
Federal Highway Administration, Highway Safety Engineering Studies – Procedural
Guide, FHWA Report No. FHWA-TS-81-220, November 1981.
K.L. Campbell, et al, Sources of Exposure Data for Safety Analyses, FHWA Report No. FHWA-RD-97-025, November 1997, 102pp.
K. W. Ogden, Safer Roads: A Guide to Road Safety Engineering, Avebury Technical Ashgate Publishing Ltd., Aldershot, England, 1996, 516pp.
National Highway Traffic Safety Administration, The Art of Appropriate Evaluation, Report No. DOT HS 808-894, May 1999. ONT
R. Pfefer, et al, Highway Safety Data: Costs, Quality, and Strategies for Improvement – Final Report, FHWA Report No. FHWA-RD-96-192, 95pp.