Response reliability and the study of adolescent substance use progression

Studies of adolescent substance use progression typically infer a sequence of initiation from self-reported ages at first use of alcohol, tobacco, mar ijuana, and hard drugs. This paper examines the reliability of this procedu re for a sample of 892 New Jersey youths interviewed on two occasions separ ated by three years. Individual responses on the second occasion differed s ubstantially from those provided on the first. However, the inferred sequen ces were consistent as long as 1) first use of alcohol and/or tobacco was c onsidered a single stage, and 2) cases in which individuals initiated the u se of two substances in the same year were considered as ambiguous regardin g order. The sequences reported were also consistent with the gateway theor y that suggests alcohol/tobacco precedes any possible use of marijuana and hard drugs.