Abstract
As part of the continuous process of improving highway safety, the engineer relies on information provided by accident record systems. The study described in this paper sought to determine the reliability of this system in new mexico. Techniques employed in the study included internal consistency checks, comparison with other record systems, and matching actual and reported crash site data. The extent of omitted and inacurrate data having primary relevance to engineering analysis was found to exceed acceptable limits. Incorrect locational information was the most serious problem. The recommended solutions to this problem consist of a modified accident report form and improved contact with enforcement officials. (author/publisher)