Examining cohort effects in developmental trajectories of substance use

Abstract

The current study demonstrates the application of an analytic approach for incorporating multiple time trends in order to examine the impact of cohort effects on individual trajectories of eight drugs of abuse. Parallel analysis of two independent, longitudinal studies of high-risk youth that span ages 10 to 40 across 23 birth cohorts between 1968 and 1991 was conducted. The two studies include the Michigan Longitudinal Study (current analytic sample of n = 579 over 12 cohorts between 1980 and 1991 and across ages 10–27) and the Adolescent/Adult and Family Development Project (current analytic sample of n = 849 over 11 cohorts between 1968 and 1978 and across ages 10–40). A series of nonlinear, multi-level growth models controlled simultaneously for cohort and age trends in substance use trajectories. Evidence was found for both age and cohort effects across most outcomes as well as several significant age-by-cohort interactions. Findings suggest cohort trends in developmental trajectories of substance use are sample and drug-specific in the adolescent and early to mid-adult years. Thus, studies that do not control for both trends may confound cohort and developmental trends in substance use. For this reason, demonstration of one analytic approach that can be used to examine both time trends simultaneously is informative for future multi-cohort longitudinal studies where change over time is of interest.