This proposal is for training and research regarding the smoking behavior of a nationally representative sample of adolescents ages 12-18, surveyed as part of the National Longitudinal Survey on Adolescent Health. The analysis will examine differences among smoking youth by age, race/ethnicity, and gender. Further, this research will identify predictors of transitions to subsequent stages of smoking that could be used to target programs to adolescents most likely to benefit from them. Ancillary variables such as alcohol use will be studied in the context of mapping the natural history. The longitudinal nature of the study will subsequent examination of time-variant characteristics of the sample and their impact on predicting smoking behavior.