The purpose of this study was to test a hypothetical model of alcohol problems in German adolescents. Among 180 offspring, family history of alcoholism, parenting styles, behavioral and emotional problems, peer-group characteristics, feelings of self-esteem, behavioral problems and psychiatric comorbidity of the parents were examined. Data were generated from the Study of Health in Pomerania (SHIP), in which families were randomly selected if 12-18 year old biological offspring were members of the household; a smaller group of subjects was selected from local outpatient treatment centers. Members of 133 families, including 180 (50.6% male) offspring who were appropriate for the current analyses, received personal semistructured diagnostic interviews and several self-rating questionnaires. Analyses compared offspring with alcohol problems (AP; n = 40) and with no alcohol problems (NAP; n = 140), and used structural equation modeling to test a hypothetical model. The comparisons revealed that the AP group had significantly more behavioral problems (e.g., aggression/delinquency), more perceived parental rejection and less emotional warmth, a higher amount of alcohol consumption, were more likely to associate with substance-using peers and more often received a diagnosis of conduct disorder or antisocial personality disorder. Whereas the family history of alcoholism did not differ significantly between groups, parents of offspring with an alcohol use disorder had significantly more additional diagnoses on DSM-IV Axis I. The evaluation of the model supported the importance of aggression/delinquency and association with substance-using peers for alcohol problems in people. An additional diagnosis in the parents was directly and indirectly (through aggression/delinquency) related to alcohol problems of the adolescents. The data indicate that alcohol problems in the offspring are associated with several domains of influence in their environment. Prospective studies measuring both biological and environmental factors using sufficient sample sizes will be needed for optimal understanding of the development of alcohol problems in youth.