During the second semester of the 1989-90 school year, all of Virginia's 65,000+ sixth-grade students were the first to take literacy tests in mathematics, reading and writing as part of a new Virginia Assessment Program mandated by the legislature. Passing scores on all three of these literacy tests is now mandatory for admission to ninth grade. The writing portion of the assessment requires that students construct a writing sample in response to a writing prompt. For the three years prior to 1989-90, school systems within the state could participate voluntarily in a fourth-grade baseline test to determine student potential for failing the sixth-grade assessment. Students whose papers fall into the bottom quartile of all papers scored each year are considered at-risk for failing the Literacy Passport Test at the sixth-grade level. This study examines the writing behaviors and the characteristics of the papers written by four fifth-grade students identified by the Virginia Department of Education as at-risk for failing the Literacy Passport Test. The author chose to function both as researcher and as participant/observer in the study, functioning in both of these roles for a twenty-one week period during the fall and early winter of 1989-90. Data was collected during a three hour per day, three day a week time period. Collection sources included field notes, interviews with students and teachers, and student papers, including the fourth-grade baseline assessment, papers written during the twenty-one weeks of data collection and a simulated Literacy Passport Test writing sample. Findings include a description of each student's approach to writing and an analysis, both analytical and domain-based, of the writing of selected papers of each of the four students. Implications for teaching, as well as suggestions for further research, are included in this document. / Ed. D.