Using the Louisville Behavior Checklist and a test of human figure drawings, the researchers assessed whether statistically significant differences existed between the groups and whether reliable identification of sexual abuse victims could be made using scores from these two instruments. Discriminant function analysis showed significant differences between the groups on variables including internalized behavior problems, externalized behavior problems, and emotional indicators. However, use of these scores to classify individual children as sexually abused or nonabused produced substantial misclassifications, particularly between the sexually abused and clinic-referred groups. While, as a group, the sexually abused children scored higher on the Louisville subscales and drew more emotional indicators in their human figure drawings, the variability of individual scores within each group allowed some sexually abused subjects to score lower than children in either the clinic or community groups. 3 tables and 27 references