抄録
The objective of this study was to examine which factors affect the generalization of mand for instruction. A boy with autism (7 years old) participated in this study. First, the participant was taught a mand for instruction in a “labeling photo cards” task. After successful acquisition of the mand, a generalization test was conducted. In this test, the participant was presented with various tasks (e.g., “labeling illustration cards,” “classifying cards,” “labeling objects,” and “matching”). These tasks were divided into four types on the basis of the three-term contingency. Tasks of type A had similar discriminative stimuli and response to the trained task, tasks of type B had similar discriminative stimuli to the trained task but response was dissimilar, and tasks of type C had a similar response to the trained task but discriminative stimuli was dissimilar. Tasks of type D had neither similar discriminative stimulus nor response. The results showed that mand for instruction generalized to tasks of types A and C. Thus, the primary factor for the generalization of mand for instruction is the similarity of the required task response.
