Relationship of Reinforcement by Student Clinicians and Peers to Accuracy of Imitated Grammatical Constructions during Language Training

Percentages of correctly imitated grammatical constructions were examined within a framework of reinforcement theory. Four small treatment groups, each of 4 language-delayed children, were repeatedly observed interacting with different student-clinicians. The Therapy Reinforcement Schedule was used to obtain frequency counts of verbal and non-verbal reinforcements so that within each language group the students could be objectively divided into those who frequently and infrequently were reinforcing, and the children into those for whom peer-reinforcement was high or low. The resulting 4 groups were compared for proportions of accurately imitated constructions with a repeated-measures analysis of variance design (student-clinicians' reinforcement × peers' reinforcement × trials). Correctly scored imitations increased significantly over trials. In addition, children interacting with frequently reinforcing students received higher imitation scores than those with infrequently rewarding ones. Significant main effects of reinforcement by the peer group were not observed, although a reliable interaction of student-clinicians' reinforcement × peers' reinforcement was present. Regression analyses indicated the importance of certain types of reinforcements in predicting accuracy of imitation.