We estimate the effect of education and student status on the propensity to smoke. Our estimation strategy accounts for the endogeneity of education by "differencing out" the impact of unobserved characteristics correlated with educational attainment. This is accomplished by exploiting education differences between similarly selected groups that are one year apart in their life cycle. The results indicate that an additional year of education does not have a causal effect on smoking. Unobserved factors correlated with educational attainment entirely explain their cross-sectional relationship. We do find, however, that being a student reduces the likelihood of smoking. This may be a peer effect, which prior research shows has a significant impact on smoking decisions.