Substance use clusters in a college sample: a multitheoretical approach.

Substance use factors were examined and cross-validated in a cluster analytic approach in two independent samples of college students: N = 291 and N = 602. Reported frequency, intensity, and amount of substance use were examined for beer, hard liquor, marijuana, amphetamines, barbiturates, psychedelics, cocaine, and heroin. Variables were reduced using Principal Components Analysis (PCA) to form four substance use composites. Composite scores were entered into two different methods of cluster analysis that each identified four distinct clusters of substance use groups. External validity was obtained by showing that these four groups differed on a set of relevant variables. The four groups served as levels of the independent variable, substance use type, in four MANOVAs examining group differences on peer and family influence, psychosocial functioning, habit acquisition, and self-efficacy. Findings indicate that as frequency and intensity of substance use increased, individuals reported more problems in living, although a causal direction cannot be established. The findings are of potential value in early identification, prevention, and education regarding substance use among college populations.