N his well-known review of the human capital approach to the study of income distribution Jacob Mincer makes the point that . . a better understanding of the relation between and earnings requires an understanding of the factors determining investment (1970, p. 18). The importance of this point has been lost somewhat in the past few years with the plethora of empirical examinations of the relationship between years of schooling (as the measure of human capital investment) and earnings. Nevertheless, it is clear that if schooling is itself the result of optimizing rather than random behavior, the typical earnings regression will overstate the contribution of schooling to earnings. A simple model that explains interpersonal differences in investments in formal schooling has been developed by Becker (1967, 1975) with more formal extensions of this model having been presented by Ben-Porath (1967) and Wallace and Ihnen (1975). Basically, Becker's model views the individual as maximizing the present value of his net earnings over the life cycle by investing in formal schooling up to the point at which the marginal rate of return from the equals the marginal financing costs. This may, in turn, be expressed in a conventional demand-supply framework. The demand for schooling is the product of two factors: the expectation of returns from a particular. level of schooling achievement and the probability that the particular individual will in fact succeed in attaining this level.' The first factor is largely determined by the exogenous forces of the labor market where individual differences arise because of imperfect knowledge. The second is largely a function of individual'capacities (ability), the schooling environment, and the extent to which the individual believes the schooling environment and curriculum will actually lead to an increase in his stock of human capital. The supply side of the model reflects the opportunity for which is determined, in large part, by the availability of financing funds. For youths enrolled in high school the major economic decision they face is whether to continue with the formal educational process. The primary alternatives to schooling are full-time participation in the labor force, service in the Armed Forces, marriage and work within the household. The data indicate that these alternatives are chosen, in varying degrees to the extent that only 75% of a schooling cohort that entered the fifth grade in 1964 graduated from high school in 1972 (U.S. Department of HEW, 1974, p. 14). A sizable proportion of young adults, therefore, terminate their educational before the completion of high school. This decision to drop out of high school has seemed to arouse considerable public and private concern. While there is now some evidence and concern regarding an overinvestment in college training (Freeman, 1975), the fact that young people drop out of high school usually raises the spectre of increased crime, drug usage, unemployment and a general alienation of youth from the adult community. The public result of this. concern has been a variety of dropout prevention programs which have been supported under Title VIII of the Elementary and Secondary Education Act together with a provision' of the Vocational Education Act. which directs Federal monies to areas of high need, including local areas with a high concentration of school dropouts. A variety of instructional methods have been supported in the dropReceived for publication May 9, 1977. Revision accepted for publication February 15, 1978. * University of South Carolina. The research reported herewas supported by the Office of the Assistant Secretary for Planning and Evaluation, Department of Health, Education and Welfare. That support, together with the comments of Caroline Clotfelter, Susan Cochrane, Elchanan Cohn, Linda Edwards, Robert Hauser and two anonymous referees, is gratefully acknowledged. The computational assistance of Jane Lee was also of great help. The conclusions expressed here do not necessarily reflect the position of the sponsoring agency. I This useful distinction is taken from Griliches (1973).