The observable effects of curricular tracking were investigated for a cohort of students in six Vermont high schools during the years 1981-89. The outcome variable was persistence in school as predicted by grades in academic subjects from grade 4 through 9, or time of withdrawal, and the level of courses in which students were enrolled. Three groups were identified: (1) general track; (2) vocational track; and (3) college-bound track. Data were collected from student records for 421 students. Sixty percent of the general track students dropped out; an average of 3 to 15% of the vocational and college-bound groups dropped out. Many of these students could have been identified as at-risk; by knowing how students performed in mathematics and English from grades 4 through 9, the model was able to predict the students that would drop out, and it was right 88% of the time. Results strongly suggest that curricular tracking was related to dropping out of school. Twelve bar graphs and three tables contain study data. A 51-item list of references is included. (SLD)