Model Similarity and Observer Learning

Abstract

During Phase I 5 groups of 9 Ss each were instructed to say words at a signal to obtain a baseline of animal responses. 4 of the groups then said words intermittently with a taped model during Phase II. Two groups were exposed to a “normal” model and two groups were exposed to a “retardate” model. Half of the Ss under each taped condition received vicarious reinforcement, i.e., hearing “good” from an E when the model gave a critical response. Half under each tape condition received no vicarious reinforcement. The fifth group, a control group, heard no tapes and received no reinforcement. Model-observer similarity emerged as a significant factor. Reinforcement facilitated learning under both model conditions. The control group showed no increment in critical responses across trials.