The influence of errors and decision-making on learning by performance and learning by observation

Abstract

Learning is commonly studied in situations which ideally are limited to transactions between an individual and some aspect of the environment which does not include another individual, for example, classical and operant conditioning experiments. Typically, interactions between the Subject (S) and the Experimenter (E), and S and S are eliminated, or at least minimized. The purpose of this thesis, however, is the study of learning in situations which involve not only transactions between S and his environment, but also between one S and another, that is, observational learning. In order to define observational learning, it is necessary to distinguish it from the related concepts of stimulus enhancement, social facilitation, allelomimetic behaviour, and imitation. Stimulus enhancement refers to a situation where one subject, acting as a model, draws an observer's attention to a particular cue. Spence (1937), in a review of experiments on learning by observation or imitation, pointed out that stimulus enhancement could explain a number of the results which had been obtained in problem-solving experiments. He suggested that the response of a model to a particular part of the apparatus has the effect of directing the attention and responses of the observer to this part of the apparatus, and thus the probability of the observer arriving at the correct solution is greatly increased. Thus, when stimulus enhancement alone is operating in a situation, what the observer learns from the model is the direction in which to concentrate his efforts.