Discrete trial teaching is a systematic form of instruction found to be effective for children diagnosed with autism. Three areas of discrete trial teaching warranting more research are the effectiveness and efficiency of various prompting procedures, the effectiveness of implementing teaching in a group instructional format, and the ability of children with autism to observationally learn from their peers. This study evaluated the effectiveness of a no-no prompting procedure implemented in a group instructional format to teach five children, four of whom were diagnosed with an autism spectrum disorder, to correctly label facial expressions. Additionally, this study evaluated whether participants observationally learned to expressively label facial expressions not directly taught to them but taught to their peers. Using a multiple baseline design, results of this study indicated that all participants learned to expressively label facial expressions taught to them directly with a no-no prompting procedure. The participants also learned through observation to expressively label facial expressions taught only to their peers using the same procedure. (Contains 5 tables and 5 figures.)