Memory schemata for descriptive texts.

Abstract

A theory describing the use of schemata for representing prose In semantic memory was adapted and applied in assumptions regarding the recall of descriptive texts. Different texts that discussed the same general topic were defined as similar, and it was assumed that similar texts could share the same schema representation in memory while dissimilar texts could not. In each of five trials, university students were asked to read, and after two minutes, write from memory a presented text. Correct free recall of the words in each text was the dependent measure, and differential effects of prior trials were sought under three presentation conditions. Each condition was specified by the similarity of texts presented in successive trials. One condition presenting similar texts, Included a topic shift after the third trial. Comparisons were made between this condition and two control conditions: a condition with similar presentations without a topic shift; and a condition with dissimilar presentations. Assumptions regarding the mediation of transfer effects in the recall of texts sharing a schema specified recall probabilities across trials in each condition. None of the predicted trends in performance were substantiated as significant differential effects of prior trials were not obtained. Also, recall performances in the three conditions were indistinguishable.