Abstract
The present study examined effects of using talking dice to train an 11-year-old boy with pervasive developmental disorders to engage in conversation. The boy was first trained in listening behavior, and then in speech behavior. In the training sessions on speech behavior, the length of time that he talked and the length of each of his sentences were measured under 5 conditions; (a) presenting a model of ending a conversation, (b) presenting a topic prior to starting a conversation, (c) providing enough time for him to think, (d) providing him with a choice of topics, and (e) providing him with an opportunity to talk about any topic he chose. The results suggested that presenting a model of ending a conversation, providing enough time for him to think, and providing him with a choice of topics were effective for increasing the length of his sentences, whereas presenting a topic prior to starting a conversation and providing him with an opportunity to talk about any topic he chose were effective in increasing the length of time that he talked.
