The subjects were 776 males and females in the 9th through 12th grades in 8 schools reflecting a wide range of locations, familial backgrounds, and socioeconomic levels. Students completed questionnaires that gathered information on demographic characteristics; involvement with marijuana, narcotic pills, alcohol, and cigarettes; parental control; involvement with youth activities; delinquency; attitudes toward drugs; and psychological problems. Subjects involved with marijuana were more likely than other subjects to have marijuana-using friends, to be male and members of youth movements, and to have less parental supervision. Data from a subsample consisting of the 65 youths most involved with marijuana and 35 other randomly selected youths from the original sample received further analysis. Two clusters of youths with high involvement with marijuana were identified: youths from low-income problem families with many children and poor parental control, and youths of high socioeconomic status who were seeking exposure to new thrills. The close connection between marijuana, tobacco, and alcohol involvement indicated the existence of an underlying factor: readiness to experiment with mood-changing substances. Those tending to drink more alcohol were youths with delinquency problems, involvement in youth activities, psychosomatic problems, and low parental control. All these are characteristics of the low socioeconomic segment of the population, in which parental problems do not leave time for parents to exercise proper education or control over their children. Tables and 10 references.