One phase of an investigation of the high attrition rate of junior colleges, this study sought to determine whether it is possible to predict, before classes begin, if an entering student will stay through or withdraw during the first semester. Data obtained in Phase One of the study of 653 entering students were augmented with high school test data and parental occupational information. Two data decks--one for persisters and one for withdrawers--were used in the discriminant analysis. The variable, Social Studies Florida Twelfth Grade score, emerged as the most significant in discriminating between persisters and withdrawers. Student status, i.e., full-time or part-time students, was generally second in importance, but was significant in only one of the five analyses, with the full-time student always more likely to persist than his part-time peer. Third in significance was generally race, with blacks less likely to persist than whites or "others." It was found that with the Florida Twelfth Grade excluded, high school rank entered significantly, the student having the better high school standing more likely to persist than one with a lower standing. When all academic data were excluded from the analysis, none of the other variables were significant in discriminating between persisters and dropouts. It was found that the data available would lead to predictions of about 65% accuracy. (DB)