A central problem in the use of children's behavior inventories involves the verity of endorsements made by reporters. Therefore, procedures for constructing, administering, and scoring a deviant behavior inventory were developed so as to minimize and control over- and under-reporting error. To minimize false negatives, comprehensive coverage, a not sure response category, a re-sort procedure, and use of 5 reporters were included. To control false positives, clarity and specificity of working, a detailed inquiry, and independent validity judgments were included. In 20 deviant behavior inventory assessments of school children a considerable amount of both types of error was demonstrated when measured against the independent validity judgments, suggesting caution in the use of similar behavioral assessments which neglect to control for such error. (28 ref.)