mains: Fathers attributes, father-child relationship, adolescents personality, and environmental factors. The father attributes include his HIV status, illegal drug use, and methods of coping with HIV or the risk of having HIV. The measure of the fathers illegal drug use was derived from a combined score of the fathers report of his illegal drug use and the childs report of the fathers illegal drug use. It is found in previous studies that by combining the parent and childrens responses to measures provides a greater predictability than using one source alone. The father-child relationship domain include measures of warmth/affection, parenting variables (such as mothers parenting style and fathers parenting style such as rules and discipline), childs identification with both parents (admiration, emulation), father-child conflict, and the amount of time the father and child spent together. Except for the childs identification with the father scale, which is solely from the childs scores, all the father-child relationship scales were from a combined score of each fathers and his childs questionnaires. The adolescent personality domain includes intolerance of deviance, rebelliousness, delinquency, aggression, sexual activity, and other measures of problem behavior. These measures were all taken only from the childs self-reports. The last domain, environmental, included measures of school environment, victimization, and gang membership. These measures were also taken only from the reports of the child. Analyses: Pearson correlation coefficients were computed between the scales in the two domain chosen (i.e., father attributes and father-child relationships) and the adolescents past-year marijuana use. For the purpose of this paper, I have chosen only the variables from the father attributes and father-child relationship domains for the interaction-regression analyses. This will allow us to examine the effect of a variable from one domain in...
