The current two-experiment study examined the effects of a regressive prompt-delay procedure on sight-word reading of four elementary school students. In contrast to traditional progressive prompt-delay procedures in which the latency of prompts is increased, the regressive prompt-delay latency is decreased over time. Data indicate that participants learned sight words quickly and maintained responding at high levels when instruction was withdrawn across both regressive and progressive prompt-delay conditions. Results are discussed in context of empirically supported prompting strategies and possible motivating operations that might be introduced when the learner is instructed to respond more quickly than the experimenter in regressive prompt delay (e.g., a game-like activity that potentially makes the activity educational, engaging, and enjoyable for students). [ABSTRACT FROM AUTHOR]