Objective:
Numerous studies examining the question of race and delinquency have found ethnically unique processes accounting for adolescent problem behaviors. Other researchers have found evidence that while levels of family and school bonds by ethnic groups may differ, the same process variables underlie adolescent delinquency regardless of race. The purpose of this investigation is to examine whether, consistent with Hirschi's original formulation of social control theory, there is an independent effect of family and school bonds in early adolescent Caucasians and Hispanics predicting delinquent conduct. Second, this investigation examined whether the relationship among predictor variables differed by ethnic group.

Method:
The sample consisted of 1,021 6th and 7th grade Caucasian (73%) and Hispanic (27%) early adolescents (mean age = 12.7) from three southwestern United States middle schools. It was equally divided by grade and gender. Information provided by the Census Bureau indicated that families were representative of a wide range of socio-economic classes. A passive consent procedure was used in this study; parents had the option of not permitting their child to take part in the study. Self report data were collected in class rooms during a three-month period, from March to May 1992. Students were surveyed on two consecutive class days under the supervision of their teacher and at least one research assistant.

Delinquency was assessed using the Achenbach's Youth Self Report (YSR; 1). Adolescents rated statements about their own behavior (alpha=.85); parental monitoring and disciplinary practices were measured via eight items based on Patterson and Dishion's work (2; alpha =.79). Seven items assessed the quality of the affective bond in the parent-child relationship (attachment; alpha=.75). Academic performance was assessed by having each subject report an overall grade point average. Students rated a single item for a measure of after-school commitments, absenteeism, educational aspirations, and peer substance use. Finally, school substance use was assessed by two items measuring the degree to which adolescents were willing to engage in substance use behaviors at school (r = .58).

Results:
First, oneway ANOVA's indicated that Hispanics reported more delinquent conduct than their Caucasian age mates (mean scores = 7.49 and 6.54, respectively; p <.003). The two ethnic groups also differed on three predictor variables: Caucasian youth reported significantly higher educational achievement (mean grades = 3.2 versus 2.8), educational goals, and after-school commitments (mean number of commitments = 1.4 and 1.0; p <.001). Second, set hierarchical regressions found that both variable sets entered (family and school) were highly significant (p <.001) for adolescents in both ethnic groups. While the family bond accounted for 23% (Caucasians) and 14% (Hispanics) of the total variance explained in delinquent conduct, the set of school bond variables accounted for an additional 14% and 26%, respectively. Both sets combined accounted for 37% of the total variance in Caucasian early adolescent delinquency and 40% in Hispanics. Finally, LISREL analyses examined the underlying similarity of predictors. Initialcomparisons of predictor correlations with delinquency by ethnic groups confirmed a general pattern of similarity (average r = .7; p <.01). The comparison of the total 9X9 correlation matrices employing LISREL by ethnicity revealed an excellent fit and a nonsignificant chi square (GFI = .99 for Caucasians and .97 for Hispanics; X2(45) = 56.32, p <.12).

Conclusion:
Findings suggested that both family and school bonds have independent effects in predicting delinquent conduct in this ethnically diverse early adolescent sample. Together these factors consistently account for about 40% of the variance explained in delinquency.

Further, although mean level differences existed in delinquency and its predictors by ethnic group, consistent with recent findings comparing members of other ethnic groups (3), the variables predicting delinquency for Caucasians and Hispanics were similar. Specifically, parental involvement most consistently predicted delinquent conduct from the family domain, and peer substance use was the best predictor from the school variables, regardless of ethnic group membership. Finally, both the independent effects of family and school bonds as well as the similarity of underlying predictors for both ethnic groups were consistent with social control theory (4, 5). This study provides further empirical support for and a validation of social control theory across different age and ethnic groups.

A. T. Vazsonyi, D. J. Flannery: "An Examination of Early Adolescent Delinquency" (running head: "Early Adolescent Delinquency"). Paper presented at the First International Congress of Adolescentology, Assisi, Italy, October 22-24, 1993.

Key Words: Early Adolescence, Delinquency, Ethnicity, Social Control.

A. T. Vazsonyi M.S., D. J. Flannery Ph.D.: The University of Arizona, Family Studies Division Family and Consumer Resources Bldg. #210 Tucson, Arizona, 85721 U.S.A.