The data are taken from Project Metropolitan, a longitudinal study of a Stockholm birth cohort comprising 15,117 cases (males and females) studied during a 30-year period. Of the females 791 cases or 11 percent acquired records of delinquency, drug use, or crime up to age 26. The delinquents were divided according to assumed seriousness and compared to the nondelinquents on two dimensions of the social bond called Attachment to school and Committment to education. The results show, first, that delinquency varies with the strength of the social bond and, second, that the ability to predict future delinquency with knowledge of these aspects of the bond is limited, as the variance proportions accounted for are low. However, when comparing the explained variances to some well-known studies, we find that the differences are not so extensive. 5 tables, 21 references. See NCJ 124605-124612 for individual papers. (Author abstract)