Two experiments were conducted on four units of a residential facility for the multiplyhandicapped retarded in an attempt to improve daily care and training services. Experiment I compared the effects of two procedures in maintaining the work performance of attendants, using an A-B design on two units. One procedure consisted of implementing specific staff-resident assignments, the other consisted of allowing attendants who had met performance criteria to be eligible for a weekly lottery in which they could win the opportunity to rearrange their days off for the following week. Results showed that the lottery was a more effective procedure as measured by the per cent of time attendants engaged in predefined target behaviors, and by their frequency of task completion in several areas of resident care. Experiment II replicated and extended these results to the area of work quality on two additional units, using a multiple-baseline design. The performance lottery was found to be an effective economical procedure that could be implemented by supervisory staff on a large scale.DESCRIPTORS: work performance, institution, staff management, work behavior, reinforcement, lottery, time off, attendantsThe care, training, and habilitation of the institutionalized retarded has become a major social issue in recent years, as evidenced by reports and essays describing the substandard conditions under which many of these persons live (Blatt, 1970;Rivera, 1972); by the development of national accreditation standards for the operation of residential facilities (Accreditation Council for Facilities for the Mentally Retarded 1971); and by the growth of legal action concerning the rights of the retarded (President's Committee on Mental Retardation, 1975).One area of residential treatment in which substantial progress has been made is that of program development (Gardner, 1971; Watson, 1972a). However, as Kazdin (1973) (Harmatz, 1973) or in social play (Daily, Allen, Chinsky, and Veit, 1974). These desirable interactions account for less than 1% of staff time in some instances. It has also been reported that as much as 20% to 40% of their time may be spent in "leisure-time" activities such as reading newspapers or watching television (Bensberg and Barnett, 1966), and that typical attendant-resident interactions may actu-