Control theory implies that social services will decrease levels of deviant behavior by strengthening the bond between the adolescent and society. Labeling theory implies the reverse, that the process of formal adjudication through the juvenile court will first stabilize and then increase levels of deviant behavior. Diversion programs were originally developed as an application of labeling theory, with the objectives of minimizing involvement with the juvenile court, referring adolescents to less stigmatizing social services, and ultimately reducing levels of deviant behavior. An additional issue has been the effect of gender on service delivery to adolescents in the juvenile justice system. This paper examines these four issues using panel data and multiple regression of follow-up on baseline variables.
Classical control theory hypothesizes that two key elements of the social bond, attachment and involvement, are related to lower levels of deviant behavior. Despite some empirical support for these relationships (e.g., Gray-Ray & Ray, 1990; Hirschi, 1969); Thornberry (1987) and Thornberry, Lizotte, Krohn, Farnworth, and Jang (1991) have criticized classical control theory for failing to allow for bidirectional associations among the elements of the bond or between these elements and deviant behavior. Interactional theory, developed to address these issues, specifies bidirectional associations among the bonds and between the bonds and deviant behavior, and describes delinquency as the development of a criminal career over time (Thornberry, 1987). Thus, interactional theory implies the use of longitudinal data and nonrecursive models to examine control theory.
Using longitudinal data to estimate both lagged and instantaneous nonrecursive models, Liska and Reed (1985) found that parental attachment negatively affected delinquency, delinquency negatively affected school attachment, and that school attachment positively affected parental attachment in a circular model that held for white, upper-class, and lower-class respondents. Conversely, using longitudinal data to control for prior delinquency, Agnew (1991) found that parental attachment had no direct instantaneous effect on delinquency and had only a trivial indirect instantaneous effect on delinquency via school commitment. Elliott, Huizinga, and Menard (1989) found that family involvement was negatively related to marijuana use but unrelated to delinquency or polydrug use. Thornberry et al. (1991) found that for younger adolescents, parental attachment and delinquency were reciprocally related, but that for older adolescents the association was unidirectional with delinquency predicting parental attachment. Using panel data, Krohn, Thornberry, Collins-Hall, and Lizotte (in press) found that drug use predicts school dropout (weakening a bond to society), but school dropout has no effect on subsequent drug use. Further, delinquency was found to be neither an antecedent nor a consequence of school dropout.
Longitudinal studies have reported complex results for the association among control theory variables and deviant behavior. However, reciprocal relationships among the elements of the social bond and deviant behavior over time have been found. Thus, in falling to take into account interactional theory and nonrecursive modeling, cross-sectional studies cannot assign the correct causal direction to an observed empirical association. However, very few studies have examined bidirectional associations among control theory variables and deviant behavior using a longitudinal data set and a nonrecursive model (Thornberry et al., 1991). The present study used a longitudinal data set to examine hypothesized bidirectional associations among elements of the social bond and drug use by adolescents.
According to labeling theorists, the self is viewed as a social process subject to the reactions of others (Mead, 1934; Cooley, 1962; Becker, 1963). …