This prospective, longitudinal study showed that an intense educational program significantly improved academic performance in underserved students who were selected solely for their desire to participate in the program and their parents' willingness to support them. Mean performance in six representative components of the Woodcock-Johnson Tests of Achievement-Revised improved at a rate that substantially exceeded one year for each year in the program. During three years of attendance, performance levels increased from below to more than two years above grade level. (Contains 5 tables.)