OBJECTIVES: (1) To determine the frequency of occurrence of oppositional defiant behaviors in the general population with a view to establishing empirical frequency cutoffs for the evaluation of oppositional defiant disorder (ODD). (2) To examine the effects of changes in the definition of ODD between DSM-III-R and DSM-IV. METHOD: The Great Smoky Mountains Study is a general population study of 9-, 11-, and 13-year-olds. Subjects and their parents were interviewed with the Child and Adolescent Psychiatric Assessment at baseline and again 1 year later. RESULTS: Ninetieth percentile frequency cutoffs for ODD symptoms are given. Although rates of ODD were little different between DSM-III-R and DSM-IV, fewer than half of those who met criteria by one or the other definition met criteria according to both. DSM-IV defined a more disturbed group of children than did DSM-III-R. Requiring only two or three ODD symptoms plus impairment identified children with substantial evidence of disturbance who did not otherwise meet criteria for any diagnosis. CONCLUSIONS: The DSM-IV criteria represent an improvement over DSM-III-R. However, a reduction in the number of ODD symptoms required for diagnosis is indicated. Symptom frequency criteria for ODD symptoms are suggested for clinical use.