文章の記憶に及ぼす文脈情報の効果

Effect of Contextual Information on Memory for Narrative

岡 直樹

森 敏昭

桐木 建始

広島大学教育学部紀要. 第一部

Bulletin of the Faculty of Education, Hiroshima University. Part 1

The reduction of information presented in the narrative plays an important role in the process of comprehending and remembering a narrative. The contextual information or the title given to a narrative is considered one of the important factors to affect this process. Researches on the use of titles showed that knowledge of contextual information prior to the presentation of materials distorted many of events in the materials compared to no-title condition (e.g. Kozminsky, 1976; Owens, Bower, & Black, 1979). The present experiment was, therefore, designed to demonstrate how contextual information about narrative influences the way of readers' comprehension and memory.

A mixed design was used with retention interval (immediate and 24-hour) as a between-subject variable and contextual information (Before, After and Control condition) as a between-subject variable.

The subjects were 60 students at Hiroshima University.

The materials was a sequence of five event sequences about a character, called Yukiko, making a cup of coffee, visiting a doctor, attending a lecture, going grocerly shopping, and attending a party. These stimulus materials were the same as those of Owens, et al. (1979).

Subjects in the Control condition read the five events with the name of the character. Subjects in the Before and After conditions read the contextual information just before or after they read the five events. The subjects were given 2.5 min. to read the narrative,and were asked to recall the narrative immediately or after 24-hour interval. Immediately following the recall test, the subjects were given the recognition memory test. They were instructed to rate each of 50 statements on a 7-point scale (1: sure old; 7:sure new). Fifty statements consisted of 10 old statements, 10 new statements, 5 plausible, 5 implausible statements for the contextual information and 20 neutral statements.

Recall data indicate that the subjects in the After condition recalled more sentences from unrelated events than from related events. These results were interpreted as an evidence that the contextual information provided at output phase influenced the reconstruction processes of recall.

The main findings of recognition test are that the subjects in Control and After conditions were more accurate in recognizing plausible statements than the subjects in Before condition. Implausible statements, on the other hand, tended to be rated as old statements. These results were interpreted as showing that contextual information provided at input phase affected comprehending and storage pahses.

教育 [ 370 ]

日本語

広島大学教育学部

[ISSN] 0440-8705

[NCID] AN00213359