Scheier L M, Newcomb M D
Department of Psychology, University of California, Los Angeles 90024.
J Subst Abuse. 1991;3(3):277-99. doi: 10.1016/s0899-3289(10)80012-4.
Many psychosocial factors are associated with adolescent drug use, though most have not been tested as true predictors of drug use in prospective studies. Studies to date have also not differentiated predictors of drug use from abuse and have not addressed differential effects for specific substances. To address these concerns, we expanded the multiple risk-factor approach using 2-year longitudinal data from a sample of seventh graders. Frequencies of use for alcohol, cigarettes, marijuana, cocaine, and hard drugs were assessed at Time 1 and Time 2 and used to reflect latent constructs of polydrug use. From a set of 29 risk factors, unique predictors of any substance were separated conceptually according to whether they most related to initiation/experimental or problem/heavy drug use and were then summed into two-unit weighted indexes at each time. Distribution-free structural equation models were used to accommodate the nonnormal distributions of the illicit drug use measures. The problem risk index was strongly correlated with polydrug use at Time 1 and increased polydrug use at Time 2. Several specific relationships between risk and drug use across time also were noted.
