The usefulness of providing specific information and a progress/feedback mechanism to control utility usage in community-based, halfway houses for dependent-neglected and for delinquent adolescents was explored. The investigation was carried out in a random sample of 12 houses of an Arizona facility, divided into equivalent groups of three houses. The observation lasted 11 months, with baseline periods varying from 6-9 months. Utility unit changes were measured every three days. Each resident in the test groups (schedules varied) was given a manual describing conservation measures, and passed a quiz on this material. In the feedback period each house was notified of increase/decrease changes in utility usage levels. Conditions and time periods varied according to group assignments. Patterns of usage and usage means for each group were analyzed. Percentage-change analyses were employed to detect any differences. Both manual and feedback conditions were effective in reducing utility usage from baseline levels. Change was particularly evident in groups with initially high usage patterns. (Author)