Generalization of Training Speaking Skills: The Role of Overt Activity, Feedback, and Child's Initial Level of Competence
72 children (mean age = 6–5) were assigned to 1 of 6 training conditions for speaking skills in a 2 (with or without feedback) × 3 (degree of overt activity) crossed design. Children in the active condition participated directly in three training exercises; each child in the passive condition observed the performance of the child in the active condition to whom he was matched; children in the active-passive condition received a combination of the two previous treatments. Children were given a pretest and immediate and delayed posttests, each of which was comprised of five tasks, four of which aimed at measuring possible transfer to other referential communication behaviors. Posttest evaluations showed that improvements in speaking behavior were related to the feedback condition, but not to the child's initial level of competence as assessed by global performance on the five pretest tasks. The reverse was true for the four transfer tasks. The amount of overt activity did not exert a significant effect on children's performance of any of the five tasks. Possible causes of the difficulty of obtaining transfer in training experiments are discussed.