Broadly defined, a "broken home" consists of a family in which at least one biological parent is missing. Although numerous studies from several disciplines have found at least some relationship between broken homes and delinquency, the mechanism that links broken homes to delinquency remains theoretically vague and empirically unknown. Nonetheless, three leading theoretical perspectives propose plausible mechanisms through which broken homes may affect delinquency at the social psychological level. These include the control perspective, which encompasses failures in the areas of social bonding, parental direct control, and juveniles' self-control; the learning perspective, which implicates the peer group and peer relations; and the strain perspective, which involves preventing individuals from achieving and maintaining positively valued goals or imposing negative stimuli. The empirical study reported in this paper involved data from the first (1976), second (1977), and third (1978) waves of the National Youth Survey, a panel study that uses a national probability sample of 1,725 adolescents between the ages of 11 and 17 at the time of their first interview. The dependent variables measured were delinquency rates and delinquency counts. The independent variables were broken homes, demographic control variables, parenting, family bonds, conventional beliefs, peer relations, and strain. The findings suggest that divorce/separation early in the life course of a child may be more strongly related to delinquency than prior research implies and that remarriage during adolescence may be strongly associated with status offending. Overall, results also suggest that association with deviant peers and attitudes favorable to delinquency account for the broken homes/delinquency relationship better than do a number of alternative explanations. 6 tables and 60 references