Research on the high-probability (high-p) request sequence shows that compliance with low-probability (low-p) requests generally increases when preceded by a series of high-p requests. Few studies have conducted formal preference assessments to identify the consequences used for compliance, which may partly explain treatment failures, and still fewer have examined the impact of programmed reinforcement for compliance to high-p requests. The present study first investigated the effects of high-p request sequences, with and without programmed reinforcement, on compliance to low-p requests using a reversal design with three children with autism. Preferred stimuli were identified via formal reinforcer preference assessments, and compliance, latency to compliance, and task completion time were measured. Results demonstrated high-p request sequences were most effective in increasing compliance and reducing compliance latency and task completion time when implemented with programmed reinforcement. Generalization probes conducted with a second trainer indicated that compliance occurred for all but one of the participants' low-p requests. The further effects of inter-instruction intervals (10 s and 5 s) were examined using a combined alternating treatments and reversal design with one participant. Results demonstrated high-p request sequences were most effective in increasing compliance when implemented with 5 s inter-instruction intervals and with programmed reinforcement. (Contains 3 figures and 1 table.)