Formulation and Validation of Operating Speed-Based Design Consistency Models by Bootstrapping

Current U.S. policy for designing rural two-lane highways is based on design speed to ensure consistency among consecutive highway segments. The design speed concept, however, does not ensure that a consistent alignment will be achieved. A recent FHWA-sponsored project (Horizontal Alignment Design Consistency for Rural Two-Lane Highways) led to three operating speed-based geometric design consistency models, which have not yet been validated. Traditionally, the validation of such models involves the collection of additional data. The statistical technique known as “bootstrapping” was used to formulate and validate the operating speed-based geometric design consistency models by using the existing FHWA database. Bootstrapping involves random sampling with replacement from the existing database, which becomes the population. One-half of the original data collected are used in formulating the models. The remaining half of the data are subsequently used for validation. The models resulting from bootstrapping were statistically equivalent to the models developed in the FHWA study. In addition, the model validation indicated that the bootstrapping technique used to validate the operating speed models is a viable alternative means of validation. It was concluded that bootstrapping is a very useful tool that can be exploited in many related areas in the transportation field, especially because of the large amounts of data typically required in developing and validating empirical models.