five features of role-taking, then, are important individual-level mechanisms by which delinquent behavior is produced. While analytically distinct, they should overlap considerably in reality. This raises the important question of whether these features of role-taking affect delinquency at all, and if so, whether they affect delinquency uniquely or jointly. This individual-level process leading to delinquency might be termed "differential social control," since self control incorporates reference groups and is thus social control and can lead either to delinquency or conformity depending on the direction of control (Matsueda 1992).Figure 1 depicts a structural equation model of differential social control and delinquency. The model consists of six blocks of variables: (1) background variables measuring location in the social structure; (2) a measure of prior delinquency; (3) variables measuring commitments to conventional roles, parental disapproval of delinquency, and appraisals by parents (objective labels); (4) role-taking measured by delinquent peer associations; (5) other variables representing the role-taking process; and (6) an outcome variable of subsequent delinquent behavior (Heimer and Matsueda 1994). The model specifies a sequence of variables in which role-taking, the proximate determinant of delinquency, mediates the effects on delinquency of role commitments and social structural position. Thus, persons develop commitments to conventional roles in part based on their objective location in the class, residential, or ethnic structure. (See Figure 1.)Tittle's Control Balance Theory portrays the probability of deviance as a function of the ratio of control exercised to that experienced, and it stipulates the type of deviance likely to be committed as a joint function of the magnitude of one's control imbalance and the seriousness of potential deviant acts (Empey, Stafford, Hay 1999). The magnitude of a control imbal...
