The purpose of this study was to investigate statistical relationships between the major factors of rural highway conditions and the accident experience. The objective was to utilize multivariate statistical techniques including cluster, factor, discriminant, and regression analyses to relate traffic accidents to the traffic and roadway design characteristics of rural intersections and rural segments. The end result of the statistical analysis was a set of discriminant analysis models which provide the best linear combination of variables to distinguish between high accident frequency and low accident frequency locations for three different intersection types and three segment types.