The study uses as a point of departure for defining deviance MacNaughton-Smith's first and second codes: the official behavioral norms fixed by law and the variable rules for attribution of deviance. While violations of the second code may entirely replace violations of the first code as the reason for labeling, various causes of labeling may be differentiated within each code, such as multiple violations of the first code. An algebraic formula is provided to define the relationship between code violations. A series of path models also describes the relationship of official and unofficial labeling factors still more exactly. The sample group to test the validity of the formulas consists of 104 juvenile offenders and a control group of 100 vocational school students. Questionnaires test code 1 violations (type, seriousness, and frequency of offenses) and code 2 violations (broken families, large families, parents' professions, living area, size of residence, scholastic failure, placement in a home, and vocational training or lack of it). The model proposed successfully associates characteristics typical for the socialization background of delinquent juveniles with indicators of delinquency. It also differentiates independently of deviant behavior between groups officially labeled and those not labeled. Using the model, the probability of certain types of offenders being labeled as deviant can be partially predicted. However, selection factors should not be overemphasized. The model must be improved by more precise definition of socialization criteria, consideration of longitudinal data on the development of deviance, and resolution of the oversimplification and harmonization tendencies of the approach. A bibliography of over 100 entries, notes, tables, and diagrams are supplied.