The project involved a police department of 124 sworn personnel in a medium-sized city in the south central United States that was dissatisfied with its use of a trait rating scale to assess patrol officers. A summary of how the rating scales were developed focuses on task analysis, developing scales and scale anchors, validation, and testing. The BARS instrument was also analyzed for problems of lenience, central tendency errors, and the halo effect -- difficulties commonly associated with personnel evaluation techniques. The results showed that the BARS system had several advantages over conventional trait rating scales. First, the anchors chosen were compiled from a careful analysis of the patrol officer's job. Second, these job-related factors once identified were then placed into dimensions which actually provided the rater with a series of behavioral descriptors for each factor which could be used to assess and rate each officer. Moreover, strong relationships were found between the composite BARS ratings and supervisory and peer rankings. Tables and over 30 references.