The objective of this study was to develop a low-cost predictive and observational tool for the early detection of disturbed children and the prevention of childhood mental illness. A self-administered questionnaire with approximately 30 items was filled out by parents concerning their children before the child entered nursery school. Following entry into school and again at the end of the semester, the teacher completed a 1-paragraph description of each child according to the parameters outlined in the behavior inventory guideline. On the basis of parent questionnaires and teacher behavior inventories, independent raters experienced in early childhood development and the psychopathology of childhood made predictions about the child's probable emotional difficulties, based on a 4-point scale, 0 to 3. In a sample study, at the end of the first two weeks of school, 4 out of 14 were correct within one rank of prediction of emotional problems including separation problems. By the end of the first semester of school, thirteen out of fourteen were correct within one rank of prediction. It is suggested that this method may be an improvement over the transcribe-interview method used in previous prediction studies. (CS)