This study is the fourth in a series of reports from the Reading-to-Write Project, a collaborative study of students' cognitive processes at one critical point of entry into academic performance. The study of task representation reported here explores how students and teachers perceive the same writing assignment by comparing the reported representations of a reading-to-write task from freshman writers with the essay evaluations from trained judges. Subjects, 72 students enrolled in four sections of a freshman composition course, received a follow-up assignment to revise their first draft of a time management essay. The experimental group received instruction in "organizing plans" and an introduction to writing issues in task representation, and completed a Self-Analysis Checklist (S-AC). The control group received the lecture (minus the revision assignment) and an introduction to the self-analysis procedure. Trained judges then rated the essays produced by both groups. Comparison of the judges' and the students' representation of the task indicated that judges and students disagreed far more than they agreed on features in the final product, but that these rival perceptions could be tempered by prompting and instructing students to "interpret with a purpose" when they revised a first draft. Results also indicated that the bases for the different perceptions and expectations are the more interesting and practical findings from this study. (Five figures and three tables of data are included; the Reading-to-Write study reference list, the S-AC, and an appendix of data are attached.) (RS)