A TEST OF LIFE‐COURSE EXPLANATIONS FOR STABILITY AND CHANGE IN ANTISOCIAL BEHAVIOR FROM ADOLESCENCE TO YOUNG ADULTHOOD*

Using longitudinal data from a sample of 236 young adults and their romantic partners, we tested a life‐course model that integrates social control and peer influence arguments with the idea of assortative mating. For both males and females, adolescent delinquency and affiliation with deviant peers predicted having an antisocial romantic partner as a young adult. Involvement with an antisocial romantic partner, in turn, had both a direct effect on crime as well as indirect influence through adult peer affiliations. For females, quality of the romantic relationship also predicted crime. The analyses revealed several moderating influences in addition to these mediating effects. For females, a conventional romantic partner, strong job attachment, and conventional adult friends all served to moderate the chances that a woman with a delinquent history would graduate to adult crime. In contrast, only conventional adult friends served this function for males.