With regard to smoking, our findings of bidirectional associations with educational achievement replicate previous research (Pennanen et al., 2011; Tucker et al., 2008). Whereas earlier studies have not extended their prediction of educational attainment into young adulthood, we found daily smoking at age 17 predicted lower education in young adulthood, when tertiary education is undertaken. Adjusting for adolescent behavior problems, however, reduced this path to non-significance. Interestingly, our findings for any current smoking and daily smoking were partly different. This suggests that experimental or infrequent smoking is differently associated with educational achievement than is regular and frequent, such as daily, smoking.
