Discrete-trials teaching (DTT) is a valuable tool within applied behaviour analysis that is used to teach skills to children with autism. We evaluated the effectiveness of a revised DTT self-instructional manual for training individuals to teach 3 tasks to children with autism. Using a modified multiple-baseline-across-participants design, 3 female university students with no previous experience in DTT: (a) attempted to teach 3 tasks to a confederate role-playing a child with autism (baseline), (b) studied the revised self-instructional manual (training), and (c) attempted to teach the 3 tasks to a confederate again (post-training). Participants required an average of 6 hours and 41 minutes to master the manual, and mean accuracy in conducting DTT increased from 45% in baseline to 82% post-training. Two of the 3 participants achieved the mastery criterion (80%) on all 3 teaching tasks on their post-training assessments; the other participant achieved the mastery criterion on all 3 teaching tasks after a session of feedback and demonstration. Participants then attempted to teach the 3 tasks to a child with autism (generalization), and achieved a mean DTT accuracy of 83%.