This report documents the long term cost benefits to society of the Self Controlled Interactive Learning Systems (SCILS), a program based on cybernetics and designed to teach early reading skills to children ages 3 to 6. SCILS required children to spend not more than 20 minutes daily using a "talking typewriter," a "talking page," and a "voice mirror" to practice letter recognition, typing letters from dictations, and reading words aloud. Teachers also worked with children on individual experience stories. The major assumptions of the cost benefit analysis were that documented high reading achievement in fifth grade for subjects who completed at least 30 hours of SCILS would continue and that none of them would be functionally illiterate or drop out before completing high school. The first chapter of the report extensively reviews the benefits to the United States and other countries of investing in preschool, primary, secondary, and postsecondary education, both for the advantaged and for those disadvantaged by gender or race. The next chapter describes the cost benefit analysis used in the study. Benefits to individuals were calculated in terms of individual differential lifetime earnings, and social benefits were calculated in terms of increases in the gross national product. Costs included those incurred by providing more high school education and by offering the SCILS system. Findings indicate an eight-to-one return on investment in SCILS for the individual and a two-to-one return for society. Other findings include: (1) regardless of initial IQ, children who had 10 or more hours of SCILS achieved at or above grade level in reading skills and children who had 30 or more hours of SCILS achieved at or above grade in reading comprehension and arithmetic; (2) For children in SCILS, IQ did not predict achievement; (3) achievement was sustained beyond grade three; and (4) there is a positive correlation of time in SCILS with achievement that increases as children go through school. The last chapter presents a longitudinal data analysis of the reading achievement of children participating in SCILS. (CB)