Interference in Sentence Recall

Abstract

The 144 subjects learned either one list of sentences or three lists (of varied interlist topical similarity) by the method of free recall. After 24 hours a measure of semantic recall was taken. Significant retroactive (p < .01) and proactive (p < .025) inhibition were observed as a function of the number of lists. Significant retroactive (p < .05), but not proactive, inhibition occurred as a function of interlist topical similarity. This experiment examines recall of the content of sentences learned by the method of free recall. Variables of interest were (a) number of lists and (b) interlist similarity of topics with which the sentences dealt. The existing data for sentences are equivocal with regard to both of these variables as they affect recall. Hall (1955) failed to find retroactive inhibition in a task requiring recall of missing words from previously presented sentences, while Mehler and Miller (1964) found a negligible amount of retroactive inhibition for the semantic content of sentences learned by the method of free recall. Slamecka (1960, 1961) demonstrated both proactive and retroactive inhibition in the recall of single 20-to-30word sentences learned by serial anticipation. Slamecka's relearning measure, however, does not permit examination of the recall of meaning without verbatim recall of the sentence. In the experiment below, topical similarity was manipulated by using lists of sentences that dealt with either biology or Australian history. Number of lists was varied by having groups learn either one or three lists. Recall of semantic content was measured after 24 hours. An interference theory would predict that subjects learning a single list would exhibit greater recall than those learning multiple lists and that subjects recalling a list similar to their other lists would show poorer performance than subjects recalling a dissimilar list.