Building schemata for expository text through collaboration and an integration of reading and writing

This study was designed to test whether an intervention based on collaboration and the integration of reading and writing would build schemata for processing expository text, and if so, would the effects of the intervention differ systematically as a function of students' initial status, reading levels, prior knowledge and initial awareness of text structure. The intervention, which involved instruction in two rhetorical structures, compare/contrast and cause/effect, was a modification of a procedure designed for a previous study by McGee and Richgels. The collaborative component incorporated scaffolded instruction following the master/apprenticeship model, procedural facilitation using graphic organizers, and novice/peer group work. The reading/writing component involved writing passages from graphic organizers and reading passages to create graphic organizers.;Subjects for the study were 76 sixth graders from two urban middle schools. All the students were in heterogeneously grouped reading classes which met every other day for one trimester. Students were tested on five occasions for each structure on level of structure awareness, percentage of idea units recalled and comprehension. Analysis involved using a growth modeling procedure, hierarchical linear modeling, to determine individual rates of growth as well as the effects of reading level, prior knowledge and initial structure awareness on growth trajectories.;For both rhetorical structures, students in the Treatment Group outperformed the Control Group in growth of structure awareness and number of idea units remembered. Measures of comprehension were unreliable because of an early ceiling effect. However, on all of these measures after the pretest, the Treatment Group outperformed the Control Group. The findings of the study provide evidence that the combination of direct instruction, socially-mediated activities and integrated reading and writing activities contribute to the building of schemata for expository text structure. Furthermore, reading level, prior knowledge and level of initial awareness of text structure were found to influence initial status but not rate of learning.