The use of cumulative cloze procedure to investigate contextual build-up in deaf versus hearing readers

The purpose of this study was to compare and to describe deaf versus hearing readers' sensitivity to contextual build-up by examining each group's successive approximations of deleted noun meanings as constructed in cumulative cloze tasks.

The methodology of this study, a 2 X 5 X 2 factorial design, focused on successive predictions of deleted noun meanings in five cumulative cloze tasks completed by five deaf and five hearing readers at fourth-, sixth-, eighth-, tenth-, and twelfth-grade levels. Both for deaf and hearing readers, sentences within the five individual cumulative cloze passages where the greatest gain in comprehension occurred (MG sentences) comprised an alternative cloze test for five other readers at all five grade levels.

The results indicate that both deaf readers and hearing readers predict meaning more accurately given passage-level versus sentence-level contextual constraints. On this particular cumulative cloze task, deaf readers at grades eight, ten, and twelve performed similarly to hearing readers at grades eight, ten and twelve respectively. However, deaf readers in grades four and six performed significantly lower than their hearing counterparts. Hearing readers at all five grade levels and deaf readers at the eighth-, tenth-, and twelfth-grade levels increased in cumulative cloze accuracy as the number of exposure point in passage-level text increased. By the sixth, final exposure points in passage-level text, no significant differences were found in deaf versus hearing readers' predictions of target nouns at grade levels eight, ten, and twelve.

Deaf readers tended to abandon correct choices after predicting target nouns more often than did hearing readers. Additionally, deaf readers returned to abandoned correct choices less often than did hearing readers. Both semantic and grammatical acceptability of cumulative cloze responses improved for deaf as well as for hearing readers as the amount of available contextual information increased in passage-level text. MG sentences, proved to be more semantically and grammatically acceptable under passage-level versus sentence-level contextual conditions.

The results of this study indicate that given passage-level contextual constraints, deaf as well as hearing readers construct meaning similarly.