Evaluating the use of programmed reinforcement in a correction procedure with children diagnosed with autism

Abstract

Background: Procedures that reduce errors while learning a repertoire play an important role in Applied Behavior Analysis for people with autism due to the detrimental effects that excessive exposure to error may have on learning. Previous studies have investigated the effects of correction procedures that require active student response after a trial with error. Some intervention manuals recommend against reinforcing responses after correction to prevent the establishment of prompt dependence. This study directly investigated the effect of reinforcement after an active-response correction procedure during tact training in four children with autism. An echoic-to-tact training procedure was used to train tacts. A “no reinforcement after correction” (NRC) condition was compared to a “reinforcement after correction” (RC) condition, using an adapted alternated treatments design. Results: All participants needed less correction trials in RC than in NRC, and considering all 26 sessions in which both training procedures were implemented, participants’ performance was higher with RC than without in 17 sessions and was the same in 3 sessions. Conclusions: We discuss the effectiveness of reinforcing correct responding after an active-response correction procedure, the absence of prompt dependence, and the implications of better correction procedures for applied settings.