THE USE OF PREDICTION BY JUNIOR HIGH REMEDIAL READERS IN INDIVIDUALIZED AND SMALL GROUP SETTINGS.

Abstract

The purpose of the study was to examine the use of a prediction technique during the reading of short stories with surprise endings. Investigated were the effects of instructional setting and content familiarity upon interest, overall reading comprehension, literal comprehension, and inferential comprehension. Verbal predictions and supportive evidence generated at the midpoint and prior the story climax was also examined. The subjects, 54 Chapter I remedial readers in a metropolitan school district in Tucson, Arizona, were randomly assigned to experimental or control groups. The study spanned three days. During this time, the students in the individualized treatment read three stories--one of familiar content, one of neutral content, and one of unfamiliar content. Each subject in this treatment generated hypotheses and support for predictions at both the story midpoint and climax. After each story, individual students completed an interest questionnaire and a comprehension assessment. Those in the group treatment followed the same procedure, with predictions and supportive evidence shared in a small group setting of three. Those in the control group read without predicting. The data analysis yielded these findings: (1) Overall comprehension and literal comprehension were not affected by the prediction treatment or story familiarity. (2) The control group surpassed the interest group on the number of inferential questions answered correctly; both the control group and the group prediction treatment subjects performed better than the individualized prediction treatment subjects on the inferential comprehension items. (3) Both the familiar and the neutral selection were more interesting to the students than the unfamiliar selection. (4) Interest scores for the control, familiar group were substantially higher than those for the group familiar treatment, the group, unfamiliar treatment, and the control, unfamiliar treatment. (5) A relationship did not exist between the interest scores and the total comprehension scores of the three stories. (6) Most predictions at the midpoint and prior to the story climax were inaccurate. (7) Predictions, though diverse, could be categorized into approximately 14 groups at the midpoint and 14 groups at the climax. (8) Most predictions were supported either with textual information or scriptal evidence; seldom were script and text ideas combined.