Donald J. Leu

Differences between Oral and Written Discourse and the Acquisition of Reading Proficiency

Examines discourse conflicts between the structure of written text and the oral expectations of young readers. Twenty-eight second grade students read and retold two stories: one with oral and one with written discourse structures. ANOVA results suggest that written discourse stories were more difficult to comprehend. In addition, an interaction appeared between familiarity with written discourse structures and comprehension of the two story versions. As familiarity with written discourse structures increased, the interference effect on reading comprehension decreased. Oral reading error results paralleled these findings, suggesting that expectations based on a knowledge of oral discourse structures may, in some cases, actually interfere with the reading comprehension of young readers.

THE USE OF CONTEXTUAL AND GRAPHIC INFORMATION IN WORD RECOGNITION BY SECOND-, FOURTH-, AND SIXTH-GRADE READERS

This study evaluated the relative importance of graphic and contextual information in word recognition as well as the relative importance of several specific sources of contextual information (lexical, semantic, syntactic, discourse) among students at three different grade levels (second, fourth, and sixth). In addition, word recognition behavior was compared at both target word locations where various aspects of context were disrupted and at non-target word locations where naturally occurring oral reading errors appeared. Results indicated that all students used greater graphic than contextual information at target word locations. The use of graphic information during word recognition tended to increase with development. The use of contextual information tended to decrease with development. These results are interpreted as being generally consistent with recent work by Stanovich (1980) and others who emphasize the importance of automatic context-free word recognition ability. Context use at non-target word locations occurred largely at function word locations that were rich in syntactic context. These results suggest that word recognition processing varies not only with the proficiency level of a reader, as Stanovichs model suggests, but also with the amount of contextual richness that surrounds any particular word.