Smoking onset in adolescence: a person-centered analysis with time-varying predictors.

Predictions from smoking-specific versus contextual models of smoking onset were tested with data from a 4-wave sample with 1,364 adolescents. Predictor variables were derived from stress-coping theory, social influence theory, and problem-behavior theory. In addition to groups of abstainers and experimenters, cluster analysis of smoking data indicated 3 groups who showed onset either in 7th grade (early onset), 9th grade (intermediate onset), or 10th grade (late onset). Almost all study variables discriminated the smoking groups from the abstainers. The onset groups were discriminated by Group X Time interactions showing differential changes in predictors (increases in risk factors and declines in protective factors), which occurred just prior to onset. The results generally support a contextual model of the onset process.

Role of parent support and peer support in adolescent substance use: a test of mediated effects.

This research tested comparative effects of parent and peer support on adolescent substance use (tobacco, alcohol, and marijuana) with data from 2 assessments of a multiethnic sample of 1,826 adolescents, mean age 12.3 years. Multiple regression analyses indicated that parental support was inversely related to substance use and that peer support was positively related to substance use, as a suppression effect. Structural modeling analyses indicated that effects of support were mediated through pathways involving good self-control, poor self-control, and risk-taking tendency; parent and peer support had different patterns of relations to these mediators. The mediators had pathways to substance use through positive and negative recent events and through peer affiliations. Effects for gender and ethnicity were also noted. Mechanisms of operation for parent and peer support are discussed.