L'hypothèse de la capacité limitée des processus de traitement de l'information. L'apprentissage distribué et le rappel libre

Abstract

Summary The hypothesis of limitations on information processing capacity permits one to predict : a) that the recall will be a constant when the total time accorded to learning is itself a constant ; b) that this relationship will be produced within certain limits of fractionation of the total time as a resuit of the number of presentations. The purpose of this study is to verify if the recall tallies with these predictions when learning is distri-buted, that is, when the total constant time is mode up in variable proportions of presentation times and temporal intervals. In the first experiment, the time of presentation represents 25, 50, 75 and 100 % (relations TE/TT) of the total time of the learning process of a series of 8 consonnant trigrams. This variable does not constitute a source of significant variation, in conformance with what the hypothesis of limited capacity leads one to expect. On the other hand, when the distributed learning process is realized in four presentations instead of one, a significant improvement is noted. A second experiment demonstrates that this improvement is not more significant under a massed learning situation. The discussion leads to the conclusion that within a constant total time, learning distributed over one presentation is more difficult than learning distributed over four presentations : in this case, the hypothesis of limited capacity predicts that the recall constants will not have the same importance in the two situations. The results obtained from the first experiment and notably, the absence of interaction between the number of presentations and the TE/TT relation, are in accordance with this prediction.