In Experiment I, three autistic children were taught the arbitrary matching-to-sample through a modified matching task, with five sets of visual stimuli (A, B. C. D, E), to examine whether stimulus equivalences were formed. Each set consisted of three stimuli. A11 of the subjects were first taught BA matching and were then able to perform AB matching in symmetry tests. After CA and DA were taught directly, CB/BC, and CD/DC performances emerged immediately for all of the subjects. Two of three subjects needed some additional training to establish DB equivalent relation. After AE was trained, EB/EC/ED were formed for two of three subjects without additional training. Each subject did not assign a common label to the stimuli in the same class, in equivalence tests and in naming test. In Experiment II, in which only two comparison stimli were randomly selected out of three stimulus elements in each set, the result also showed the equivalent relations between the stimulus sets. These results were discussed in terms of the conditions to establish the stimulus equiva1ence and the effect of response mediation.
View full abstract