Individuals with autism often have deficient intraverbal repertoires. Previous research has found success in using a blocked trials procedure to facilitate discrimination training. A previous study (unpublished) from our laboratory extended this procedure to intraverbal training. The current study continued this line of research by exploring the outcomes of probing the criterion performance more frequently. Three children with autism, ages 7-13, participated. Eight question pairs were taught. One question was presented repeatedly until a specified number of consecutive correct responses occurred, then the other question was presented. Contingent on specific mastery criteria, the trial blocks were faded into smaller blocks until the questions were presented in quasi-random order. Between each step, a criterion probe was conducted to determine if further steps were necessary. The procedure has been successful for two of the three participants. Criterion probe performance showed that not all teaching steps were needed every time. The procedure may have facilitated acquisition over time, because the number of trials to mastery generally decreased over successive targets.