This report describes the use and results of a “double conditioning” operant treatment paradigm designed to modify the behavior of chronic schizophrenics. Essentially, this paradigm involves the training of these patients through operant conditioning procedures to act in turn as behavioral therapists for fellow chronic schizophrenics.Twenty-seven patients participated in this study. These patients were divided into nine trios with two members of each trio assigned to serve as “guardian therapists” for their more regressed “charge” patient. A standardized hierarchy of response levels, ranging from simple eye contact to complex forms of social behavior, was constructed for all charge patients. Daily 45-minute behavior therapy (BT) sessions were held in which the guardians together with two staff systematically administered both social (e.g., praise) and primary reinforcements (e.g., ice cream, soda, cigarettes, chocolate, etc.) to the charge patient to induce him to progress to advanced levels in the response hierarchy. An arbitrary figure of 80 per cent correct responses constituted the criterion used to gauge when it was appropriate to thin out the differential reinforcement rate, as well as when to proceed to the next higher level in the hierarchy.Each of the BT sessions was divided into 3-minute time blocks with guardians alternating turns in conditioning their charge. As guardian patients worked, they were given considerable praise, encouragement and instruction from assigned staff. At the end of each time block, they would be given one coupon for effort and one for performance if they earned it. These coupons (secondary reinforcers) were negotiable for assorted “goodies” and “treats.”Over the course of time, we found that many of the guardian patients were able to master most of the intricacies and technicalities of the conditioning procedures with some developing a competence comparable to that of staff. Some guardians were entrusted with more and more responsibility and even took over several “exclusive” staff functions, such as keeping time, recording results on data sheets and selecting appropriate reinforcers before each session. Moreover, almost all charge patients showed variable gains as they moved up the levels in the hierarchy of responses. Despite the relative success of this program, a number of limitations to the operant conditioning paradigm were to emerge. These limitations are discussed fully in the report.