Studies of the returns to education have generally investigated the relationship between individuals' investments in formal education and on-the-job training and their labor-market productivity. It is well recognized, however, that other factors besides formal education and training contribute to a person's effective stock of human capital (and hence to productivity); these factors include early childhood environment, parents' behavior (see Dugan 1969), and associations with other individuals. There is general agreement, for example, that a child's development is affected by the ability and performance of peers in school. At a later stage, a significant part of an individual's college and graduate education appears to result from association with fellow students, the more able students contributing to the education of all. After formal schooling is completed, close associates are likely to continue to affect an individual's further educational development and to influence the rate of depreciation of the individual's stock of knowledge. One of the more persuasive explanations of the observed strong positive relationship between formal education and labor-market productivity is that, in addition to providing specific skills, formal education improves the individual's ability to acquire and assimilate information, to perceive and understand changing conditions, and to respond effectively.' From