A study investigated the role that writing plays in eleventh grade students' understanding of literary texts. Classroom observations, collected student writing, interviews with the teacher, and interviews with six case study students--two from each of three classes--provided a portrait of teaching methods, writing tasks, and student responses to each. Students read four short stories with no teacher-sponsored discussion. No writing was done after the first story. For each of the remaining three, students completed one of three in-class writing tasks: (1) restricted writing, consisting of unelaborated descriptions, interpretations, and generalizations; (2) personal writing, in which students were to explain and elaborate upon their individual responses; and (3) formal writing, in which students were to interpret the story in an extended fashion, drawing inferences from the text alone. Case study students completed the same assignments, and their verbal composing protocols were recorded. Still without class discussion, posttests on each of the stories were administered three days after students had read them, and a final posttest was given at the end of the unit. The individual and final posttest results indicated that the positive effects of extended writing versus restricted writing increased over time. At all three levels (generalization, description, and interpretation), when students wrote extensively about a story in either a personal or a formal mode, they scored better on the posttests than when they wrote in a restricted mode. (HTH)