The instructor produced two aided AAC models when a play scenario began to provide the child with immediate examples of how to produce multi-symbol messages. Then, aided AAC models were provided after the child took a communica- tive turn and/or completed an action during play; as with the baseline phase, the instructor waited a minimum of 1 s before providing a model, to ensure the child’s turn was complete. To provide an aided AAC model, the researcher (a) touched a combination of two symbols on the child’s AAC system, (b) labeled each of the two symbols while touching each symbol (for voice output systems, the speech synthesizer provided the label), and then (c) provided a spoken model akin to the spoken models in the baseline phase. For example, in the tea party scenario, if a child pretended that the dog spilled some tea, the instructor selected the symbols DOG and SPILL while saying dog and spill and then provided a spoken model (The dog spilled the tea!). A minimum of 30 aided AAC models were provided within each 15-min session. The intervention phase continued for each parti- cipant until criterion was reached (i.e., a minimum of 12 two-symbol messages during a 15-min play session over three consecutive sessions).