This article empirically examines the effect of social background, education, and school organization on adolescent tobacco use in the Netherlands. We test theories of norm enforcing and horizon expanding social networks and distinction by examining the relationship between daily smoking behavior and school organization. Using the 2007 Dutch National School Survey on Substance Use, we find that individual student characteristics are more important predictors than school characteristics. Importantly, social background effects are clearly mediated by school tracks, suggesting that tracking helps us to explain social gradients in substance use. However, school context plays almost no role in adolescent smoking behavior. One exception concerns students in the general track, for whom we find that smoking is further reduced when they are placed in the same school organization as students of the vocational track. This is in line with the theory that tracking differences are magnified in a context where interaction between students from different tracks is promoted. (Contains 2 tables, 1 figure, and 2 notes.)