AbstractsPsychology

A review of observational coding systems indicated the need for an instrument which is easy to learn and implement in applied settings, and which can be applied to diverse populations of disturbed children. A two phase study was conducted to evaluate the reliability and validity of observational ratings derived from the Direct Observation Form (DOF) of the Child Behavior Checklist. The DOF contains 96 behavior problem items (drawn from the Child Behavior Checklist) and a measure of on-task behavior. Observational periods are 10 minutes in length and samples of behavior are taken on six occasions. Observers write a narrative description of behavior noting the occurrence and frequency of any behavior problems. On-task behavior is rated every minute for five seconds. After each observational period the 96 checklist items are rated on a 0-1-2-3 scale. Scores are derived by summing the ratings of on-task and problem behavior across the six sessions. In Phase one of the study, the behaviors of 25 referred boys (aged 6-11) were rated by two observers in classroom settings. Checklist data from parents and teachers were also collected. Observer agreement on ratings of on-task and problem behavior was very good. Concurrent validity and the discriminant validity of individual behavioral items among observers, parents, and teachers was adequate considering the differences in setting and amounts of child contact. The temporal stability of ratings was poor, reflecting the high variability of behavior among referred children. In Phase two, 15 matched pairs of teacher-referred and nonreferred boys (aged 6-11) were observed in regular classroom settings. Checklist data were also collected from teachers. Significant differences were found between the groups for problem and on-task behavior; which suggests predictive validity. Relations among teacher and observer ratings were indicative of concurrent validity. Based on a crude classification strategy, screening efficiency was good for problem behavior (misclassification rate of 16.5 percent) and poor for on-task behavior (misclassification rate of 43 percent). Overall it was concluded that the DOF represents a reliable and valid indicator of children's problem behavior. Ways of improving screening efficiency are discussed. In addition, it was suggested that the development of syndromes and construction of norms would increase the usefulness of observational ratings for clinicians and researchers.