A LIST OF VARIABLES WITH SEEMINGLY GENERAL RELEVANCE TO SELF-REPORTED DELINQUENT BEHAVIOR WAS DEVELOPED ON THE BASIS OF PAST RESEARCH. THESE VARIABLES INCLUDE SOCIAL CLASS, PARENT-CHILD RELATIONSHIPS, SCHOOL EXPERIENCES, FUTURE-ORIENTED STRAIN OR FRUSTRATION, DELINQUENT ASSOCIATES, DELINQUENT VALUES, AND DETERRENCE THROUGH FEAR OF THREATENED PUNISHMENT. THESE FACTORS WERE REFINED AND WOVEN INTO AN EXPLICIT CAUSAL SCHEME. THE MODEL CAN BE CHARACTERIZED AS AN ATTACHMENT BRAND OF SOCIAL CONTROL THEORY, COMBINED WITH DIFFERENTIAL ASSOCIATION PROCESSES. DATA USED TO TEST THE MODEL WERE PRODUCED BY MEANS OF A SELF-ADMINISTERED ANONYMOUS QUESTIONNAIRE. THE INSTRUMENT WAS COMPLETED BY 734 SOPHOMORES IN 3 SEATTLE HIGH SCHOOLS IN APRIL 1975. MULTIPLE INDEXES WERE DEVELOPED FOR MANY OF THE CONCEPTS AND SPECIAL CARE WAS TAKEN TO AVOID CONCEPTUAL CONFUSION. IN THE FINAL EMPIRICAL MODEL EVOLVED, DELINQUENT ASSOCIATES APPEAR TO HAVE THE GREATEST INFLUENCE ON LAW VIOLATION. MOREOVER, THIS EFFECT IS ACCENTUATED WHEN INCREASED DESIRE FOR PEER APPROVAL PRODUCES A RISE IN SUSCEPTIBILITY TO PEER INFLUENCE. DELINQUENT VALUES ALSO PLAY A RELATIVELY POWERFUL ROLE. THEREFORE, IT IS ADVISABLE TO CONCEPTUALIZE A CONTINUUM OF VALUES IN RELATION TO THE ACCEPTABILITY OF CERTAIN ILLEGAL ACTS RATHER THAN TO EXPECT ADOLESCENTS TO ESPOUSE EITHER CONVENTIONAL OR DEVIANT VALUES. IN CONTRAST TO ORIGINAL EXPECTATIONS, PARENT ATTACHMENTS PLAY ONLY A MINOR ROLE IN THE OVERALL CAUSAL SCHEME, WHILE SCHOOL TIES ARE QUITE IMPORTANT. THIS EFFORT TO INCORPORATE UNRELATED AND COMPETING CONCEPTIONS AND PROPOSITIONS HAS LED TO A TENABLE GENERAL THEORY OF INVOLVEMENT IN DELINQUENT BEHAVIOR. IN ADDITION TO ITS BENEFITS OF INSIGHTS GAINED ABOUT THE RELATIVE IMPORTANCE AND INTERRELATIONSHIPS OF SELECTED VARIABLES IN THE CAUSATION OF DELINQUENT BEHAVIOR, THE MODEL SHOULD BE USED IN FUTURE RESEARCH. THE MEASURING DEVICES ARE APPENDED. NOTES, REFERENCES, TABULAR DATA, AND AN INDEX ARE INCLUDED. (MRK)