infer causation from education to smoking and find strong evidence that education, whether it be measured in years of …, more particularly the probability of smoking regularly. Interestingly, however, while we find that more education … substantially increases the probability of never smoking, our other main finding is suggestive that increased education has a …