In recent highway safety studies, there has been increasing attention to the accuracy and timeliness of crash data, but there have been limited studies on data aggregation. This thesis tests the effects of the underlying preprocessing of crash data for identifying high crash locations. The first portion of this thesis shows a sensitivity analysis of assigning crashes to intersections by spatial proximity and crash data attributes. In addition the sensitivity of using fixed length segmentation was tested along with the influence of using crash rate for identifying high crash locations.;The results indicate minimal effects of using different spatial proximities in identifying high crash locations. The use of spatial proximity and crash data attributes in assigning crashes to intersections may have implications on benefit/cost analyses. Predetermined fixed length segmentation and the use of crash rate may both have impacts on identifying high crash locations.