METHOD

The study was designed to identify and compare roadway characteristics at two sites. The site where one or more vehicle occupants died when the vehicle struck a roadside object was termed the crash site. A site located 1.6 km (1 mile) upstream, which the vehicle had likely passed before reaching the crash site, was termed the comparison site. The differences noted between the roadway characteristics of these sites can be used to identify other sites where fatalities are likely to occur. Comparison of characteristics of crash and comparison sites with characteristics of other Georgia roadways provides additional criteria for selecting sites for modification. During a 14-month period ending in April 1975, virtu-ally all locations where fatal collisions into roadside objects had occurred in the 108 contiguous counties in n01· th and central Georgia were studied (Figure 1). This area includes a variety of land usage (rural, suburban, urban), roadway types, and topography. The Georgia State Patrol routinely mailed fatal accident reports to the research team. Those accidents in which the fixed object had not been struck or had not been a significant factor in the fatality were excluded.

Engineering measurements were made by a threeperson team in a 0.3-km (0.2-mile) section at each site. Measurements were referenced from an object most likely to receive impact or cause vehicle occupant death. Crash sites were determined by locating a point along the roadway edge immediately adjacent to a selected object. The comparison site was located 1.6 km (1 mile) upstream from the crash site (Figure 2). The choice of turns at T-or Y-intersections was randomly selected in the