Consistent with previous studies, the findings of the current study show that the various factors measured were linked to adolescent smoking when examined separately; however, in an ecological setting, factors related to the individual and leisure activities emerged as the strongest predictors. These factors consisted of attitudes toward smoking, perceived behavioral control, and evenings spent with friends. School and family factors were mediated by the individual and leisure factors. When examined alone, however, they explained a substantial amount of the variance and a considerable amount of school differences in adolescent smoking. Although factors related to leisure and the individual were more strongly related to adolescent smoking than factors related to school and family, the latter factors might be important background factors that reinforce the effect of factors related to the individual. This study is based on data from the Control of Adolescent Smoking study (Wold et al., 2004), which is in turn linked to "Health Behavior in School-Aged Children," a cross-national study conducted by the World Health Organization. The current study sample (1,670) was drawn from a population of 15-year-old students in Norway. The Students' Survey addressed their smoking behavior, their perceptions of teachers' smoking behavior, academic achievement, attitudes toward school, best friend's smoking, evenings with friends, and engagement in sports, as well as parents' smoking behavior and other family factors. Teachers reported on their involvement in smoking, and attitude toward school policy against teachers and students smoking on school premises. 2 tables and 41 references