This study investigated the feasibility of developing reliable, valid criteria for measuring and training the skills necessary to teach autistic children. The behaviors of 11 teachers and 12 autistic children were recorded in a series of different teaching situations. Teacher-training was initiated at different times for different teachers. The results showed: (1) it was possible to assess empirically whether a teacher was correctly using defined behavior-modification techniques; (2) generally, for any given session, systematic improvement in the child's behavior did not occur unless the teacher working in that session had been trained to use the techniques to a high criterion; (3) all 11 teachers were rapidly trained to use these techniques; and (4) the teachers learned generalized skills effective with a variety of children and target behaviors.

No abstract

The present study investigated the use of a compliance-training procedure and its effect on untreated deviant child behaviors. Three children, each generally noncompliant to adult requests and with several additional problems, such as crying, aggression, and self-injurious behavior, were trained in the compliance procedure under a multiple-baseline design across therapists. Compliance was defined as the correct response to prespecified requests. Other classes of deviant child behavior were measured continuously throughout the study but not directly reinforced. The results of the study showed that (a) increases in compliance to requests were directly related to the contingencies employed; (b) decreases in untreated deviant behaviors occurred when compliance increased, even though no direct contingencies had been placed on these behaviors; and (c) the relationship between untreated deviant behaviors and compliance appeared to be maintained by a different set of events in each of the three children. The results are discussed in terms of behavioral covariation and generalization.