A normative sample of 1,114 children was contrasted with a sample of 620 sexually abused children and 577 psychiatric outpatients on the Child Sexual Behavior Inventory (CSBI), a 38-item behavior checklist assessing sexual behavior in children 2 to 12 years old. The CSBI total score and each individual item differed significantly among the three groups after controlling for age, sex, maternal education, and family income. Sexually abused children exhibited a greater frequency of sexual behaviors than either the normative or psychiatric outpatient samples. Test-retest reliability and interitem correlation were satisfactory. Sexual behavior problems were related to other generic behavior problems. This contributed to the reduced discrimination between psychiatric outpatients and sexually abused children when compared to the normative/sexually abused discrimination. The study concludes that the 38-item version of the CSBI, although it should never be used in isolation, is a reliable and valid indicator of sexual behavior that can be used to gather specific information during sexual abuse evaluations to make decisions about diagnosis and treatment on an individual basis. Tables, appendix, note, references