Do the rates at which adolescents use substances differ by the timing, duration,k curricula, or other components of the intervention programs in which they are enrolled; Are longitudinal changes in young in young adult binge drinking related to prior patterns of substance use during early adolescence? These questions can be answered using individual growth curve modeling, a powerful analytic tool for examining the relationship between individual change over time and potential predictors of that change. New, highly-effective statistical methods for integrating individual growth curve modeling into the framework of covariance structure analysis have been developed and demonstrated through application to problems in education, psychopathology, medicine, and economics. These methods remain relatively unknown and underutilized by prevention researchers, despite the great potential for examining the antecedent and concurrent predictors of the onset and change in substance use over time. The primary thrust of the proposed work is to introduce these methods to the substance use prevention community through empirical analyses of (a) existing substance use data on children and adolescents from the National Longitudinal Survey of Youth and (b) data on substance use in adolescence from the Adolescent Alcohol prevention Trial. In addition, the methodology will be expanded in several directions that have particular relevance for addressing substance use and prevention issues. These include (1) the explicit modeling of mediating effects on development, whereby a predictor may not act directly on change but indirectly through intermediaries, with implications for estimating the attenuating effects of protective factors on the progression of substance use for high-risk individuals; (2) the simultaneous modeling of individual change in multiple domains, with implications for estimating relationships among trajectories of polysubstance use and engendering investigations of profiles of change; (3) modeling discontinuous function s for individual change, with implications for modeling intervention processes; and (4) applying growth methods to accelerated longitudinal designs, where the developmental curve over a long period is recovered by using convergence techniques to link data from several overlapping age-staggered cohorts observed for shorter segments of time. This provides an unprecedented opportunity to study the age trajectory and substance use from childhood to young adulthood.
Showing the most recent 10 out of 443 publications
