Task interspersal and performance of matching tasks by preschoolers with autism

The current study examined the effects of task interspersal on the performance of matching-to-sample tasks by three children with autism. A pre-baseline assessed each child's mastery level of a large body of matching stimuli. These matching tasks included matching identical and non-identical animals, numbers, letters, and shapes. Through this assessment mastered and non-mastered matching-to-sample stimuli were determined empirically. Following a baseline condition that presented only non-mastered stimuli in succession, treatment was introduced in a multiple-baseline design across children. During the treatment condition, trials with mastered stimuli were interspersed with trials with non-mastered stimuli. For all three children, the percentage of correct matching responses to the non-mastered stimuli increased systematically with the introduction of the interspersal procedure. Following treatment, a third condition was conducted that reduced the total number of reinforcers available per session to baseline levels. The data demonstrated that all three participants maintained treatment levels of correct responding during this third condition. Thus the increased reinforcement density during treatment was not needed for maintenance of correct responding. The discussion addresses additional control procedures that would be needed to evaluate the role of reinforcement density during treatment.