Generalization and Maintenance of Language Responses

Publisher Website
Google Scholar
Abstract
Manipulating the reinforcement contingencies in training and generalization settings facilitated school-to-home generalization of language responses. Four developmentally disabled preschool children were trained in a one-to-one school setting to point to two sets of pictures in multiple baseline fashion. Initial generalization in the presence of a second trainer in school and the mother at home was documented in both no-reinforcement and intermittent reinforcement probe conditions (in which correct responses to nonprobe items were reinforced on a VR-3 schedule). High levels of correct responding with the second trainer at school were maintained in both the no-reinforcement and intermittent reinforcement conditions, regardless of the sequence of conditions. The reversal design showed that for three of the four children, intermittent reinforcement was necessary to maintain high levels of correct responding at home. Deterioration, increased variability, and, in some instances, extinction occurred when the no-reinforcement condition was in effect in the home setting.