That work is extended by identifying a broader range of individual-level mechanisms of social control, specifying group and organizational processes for controlling delinquency, conceptualizing classical criminological theories as special cases of a general interactionist framework, and testing the interactionist model empirically. The analysis used here focuses on the first three waves of data for 918 male respondents, between the ages of 11 and 17, who were interviewed in their homes in 1977, 1978, and 1979. A structural equation model of differential social control and delinquency consisted of six blocks of variables: background, prior delinquency, commitments to convention roles and parental objections to delinquency, role-taking, and outcome. The findings indicated that delinquency results largely from the variables that measure role-taking. Commitments to conventional roles, structural locations, and residential characteristics affected delinquency indirectly through role-taking. Delinquency also stemmed in part from association with delinquent peers. Results partially supported the labeling hypothesis of secondary deviance. While strong ties to conventional institutions exerted significant total effects on delinquency, these effects were mediated by role-taking. 1 table, 2 figures, 16 notes, 93 references, and 2 appendixes