Abstract
Smoking prevalence has decreased significantly among US adolescents and young adults in the past 20 years. It is possible that adolescent and young adult smokers were moving from core to peripheral positions in social networks and thus less influential as suggested in previous research on adult smokers. We construct five sample datasets to test these hypotheses but none of them receives much support. When the proportion of smokers is relatively higher in two sample datasets, smokers tended to be at more marginal network positions than nonsmokers, both smokers and nonsmoker could exert peer influence, and the magnitude of peer influence from smokers was even greater than that from nonsmokers. When smoking was less frequent in the other three sample datasets, smokers and nonsmokers were at random network positions and no peer influence on smoking behavior was detected. Therefore, core/periphery network positions are still the key linking smoking prevalence and peer influence among US adolescents and young adults but operating through a different mechanism from their adult counterparts. When scientists design and conduct prevention programs against adolescent and young adult smoking behavior, core/periphery network positions, smoking prevalence, and peer influence should all be taken into consideration.