In 1979, humanities faculty at Reading Area Community College introduced the writing sample as the means of placing entering freshmen into appropriate writing classes. In response to expressed concern about the objectivity and validity of the writing sample approach, the humanities faculty engaged in a 6-year effort to adjust the writing sample model in an attempt to improve the validity and reliability of writing sample placement results. The first part of this project involved five phases: (1) compiling placement test evaluation data; (2) comparing internal grading criteria used by individual faculty evaluators; (3) establishing relationships between evaluator placement criteria with entrance requirements used for higher-level composition courses; (4) identifying and minimizing evaluator problems and sensitivities; and (5) soliciting suggestions for improving the writing sample procedures. During the second part of the project, a profile of students with borderline writing samples was developed in order to investigate the descriptive, academic characteristics shared by these students; to monitor their success in writing classes; to assess their performance in college programs; and to reveal factors affecting student retention among this group. Appendices include instructions for writing sample evaluators, checklists for placement criteria, and interview questions used to solicit evaluator's internal criteria. (EJV)