Search

A safety analysis of a specific location requires the knowledge of "base crash rates," also known in the literature as "expected values," for identifying crash patterns at the study location. Base crash rate models were developed for the following nine crash types: (a) Injury (b) PDO (c) Total (d) Wet (e) Night (f) Rear end (g) Sideswipe (h) Fixed object and (i) Left turn crashes. These models were developed for intersection legs (also called intersection approaches) as the crash behavior of different legs within the same intersection were likely to be different. Base crash rates for the following eight leg types were modeled based on the type of control: (a) Signalized legs at four way signalized intersections (b) Signalized legs at three way signalized intersections (c) No control legs at one way stop controlled intersections (d) Stop controlled legs at one way stop controlled intersections (e) No control legs at two way stop controlled intersections (f) Stop controlled legs at two way stop controlled intersections (g) No control legs at intersections with flashing beacons and (h) Stop controlled legs at intersections with flashing beacons. This study focused on intersections formed by two or more State or US routes. To evaluate the complex interaction among the dependent and independent variables, Automatic Interaction Detection (AID) technique was used. After the completion of AID analysis, stepwise multiple regression technique was used to develop mathematical models for the split groups. It is recommended that the AID and regression models developed in this study (Appendices II and III) be used for estimating statewide base crash rates in Ohio.