We conducted cross-lagged path analyses on four waves of data from age 12 to young adulthood in a population-based sample of Finnish twins. Our aim was to investigate the directionality and magnitude of associations between educational achievement and drinking and smoking behaviors across adolescence and young adulthood. In order to remove potential confounding by externalizing behaviors, analyses adjusting for behavior problems in adolescence were also conducted. Based on previous findings, we hypothesized that the associations of educational achievement with drinking and smoking behaviors would be bidirectional. Further, we hypothesized that stronger associations would be found for more severe (i.e., drinking to intoxication, daily smoking) than for less severe (i.e., any drinking, any smoking) substance use behaviors. Theory and previous findings did not permit clear hypotheses concerning the relative strength of the associations with drinking and smoking behaviors.
