• Federal Highway Administration >
• Publications >
• Research Publications >
• 98133 >
• Ch04.Cfm >
• Accident Models for Two-Lane Rural Roads: Segment and Intersections

Accident Models for Two-Lane Rural Roads: Segment and Intersections

4. Analysis

To analyze the data acquired for the segments and intersections, a variety of new variables were developed based on the originally collected variables. It has already been noted that the traffic variables used for modeling the Minnesota intersections were obtained from the original variables by applying growth factors from nearby segments. There was significant variation in the number and size of vertical and horizontal curves from segment to segment and from one intersection to the next. Thus aggregate variables were developed for vertical and horizontal alignment to summarize alignment data and permit direct comparison of one observation with another. Other variables were developed for such items as exposure, driveway density, and intersection density. A speed variable was developed from the multiple speed variables collected.

For both the new variables and the old, univariate statistics were compiled showing their distributions in each data set. In preparation for the modeling effort, bivariate comparisons were also done to reveal correlations between variables and to clarify relationships among variables.

In this chapter we discuss the new variables and exhibit and review the univariate and bivariate statistics for both old and new variables. See the Index of Variables, at the beginning of this report, for a comprehensive listing of variables used in the modeling.

New Variables

Univariate Statistics

Bivariate Statistics

Summary