Data for the study were taken from the first 5 years (1976- 1980) of the National Youth Survey, a projected, longitudinal, household-based, general population, multistage cluster sample survey of Americans who were 11- to 17-years-old in 1976. A total of 1,725 youths were interviewed, an estimated 73 percent of the original sample. The dependent variable in the study was the felony assault scale described in Elliott et al. (1989), consisting of nonprofitable violent offenses: aggravated assault, gang fighting, and sexual assault. Separate analyses were performed for frequency of offending in the general adolescent population and, consistent with the criminal career paradigm, separately for prevalence of offending and for frequency of offending in the criminally active population. The results are consistent with previous tests of the integrated theory. The integrated theory combines elements of strain, social control, and social learning theories. According to the integrated theory, social and demographic background variables influence perceived and objective opportunities to attain academic, occupational, and other goals, and may also be associated with differences in socialization. Consequently, some individuals enter adolescence with weak bonds to social institutions (external controls) and weak personal beliefs or commitment to conventional, conforming, or law abiding behavior (internal controls). The findings of this study show the utility of the integrated theory for the explanation of violent illegal behavior. 8 tables and 30 references