抄録
Main idea comprehension is an important aspect of successful reading (Grabe & Stoller, 2011). Readers understand the main idea of a paragraph based on the subordinate details (e.g., Ushiro, Nakagawa, Kai, Watanabe, & Shimizu, 2008) and a superordinate main idea summarizing the whole text (e.g., Murray & McGlone, 1997). Although readers are not likely to rely on either of these types of comprehension alone, few previous studies have examined them simultaneously. Therefore, the present study investigates how EFL readers' representations reflected these types of main idea comprehension, comparing complete texts. Seventy Japanese EFL university students read six texts, each including a text-level main idea, paragraph-level main ideas, and details. They then completed an immediate recall task with either a detail (micro cue), a text-level main idea (macro cue), or no cue (control). Two weeks later, 40 of them recalled the texts on the basis of the same cues. The micro and macro cues improved (a) total recall rate in the delayed task and (b) recall rate of the paragraph-level main ideas regardless of recall time. The macro cues were more effective than the micro cues in these cases. The results demonstrate that both subordinate and superordinate comprehension, especially the latter, contributed to representations in the delayed task and representations of paragraph-level main ideas.
