The juveniles selected for this study had displayed some form of problem behavior or delinquent activity. The 229 juveniles participated in individual, face-to-face interviews, using a structured questionnaire. Measures of social control included moral belief, endorsement of Confucian ethics, endorsement of Daoist ethics, attachment to social workers, attachment to teachers, attachment to work, and parental control. Social-learning measures included perceived reward from offenses, learning from delinquent peers, friends' approval of offenses, and friends' disapproval of volunteer services to the community. Cognitive-development measures included problem-solving confidence, low irrational beliefs, theorizing about social problems, and self-control. Strain measures included perceived social inequality, low achievement, alienation from work or school, faith in capitalism, conflict with police, and frustration. A three-step regression analysis first identified significant background factors, followed by a further examination of the contribution of recent delinquent experiences. By controlling for these experiences, the third step estimated the unique effects of psychosocial processes by referring to the four theories under consideration for the integrated theory. These effects essentially revealed the change, increase or decrease, linked to each psychosocial factor that promoted the likelihood of delinquency. The findings show that theories of cognitive development, social control, and social learning are significant in predicting the likelihood of delinquency among marginal youth. They highlight the important influences of the juveniles' personal theorizing about social problems, attachment to work, moral belief, and friends' approval of delinquency. Their frustration and perceived social inequality were predictive of a lower likelihood of delinquency. The implications of these findings for delinquency prevention are discussed. 3 tables, 85 references, and appended description of items of measurement