ANALYSIS OF LONGITUDINAL DATA WITHIN ACCELERATED LONGITUDINAL DESIGNS

Using latent growth curve methodology and combining information from 4 different overlapping age cohorts, each measured at 1-year intervals over a 3-year period, a single developmental trajectory for adolescent alcohol use was formed spanning 12-17 years of age (N = 461). At the same time, data collected from the same study sample on a single cohor t (N = 81) over 6 annual assessments (ages 12-17) was examined to dete rmine whether the accelerated design adequately approximated the devel opmental trend within a true longitudinal design observed over 6 years of data collection. Comparison of separate models and a combined mult isample model revealed no significant differences between the 2 design s in their representation of initial status and growth, as well as sim ilar effects for the covariates of gender, deviance, and parent alcoho l use on initial status and growth.