Drinking, smoking, and educational achievement: cross-lagged associations from adolescence to adulthood

Abstract

Background: Adolescent substance use is associated with lower educational achievement but the directionality of the association remains uncertain. We analyzed data on drinking, smoking and educational achievement to study the associations between substance use and education from early adolescence to young adulthood.

Methods: Longitudinal data from four time points (ages 12, 14, 17, and 19-27 years) from a population-based cohort study of Finnish twin individuals were used to estimate bivariate cross-lagged path models for substance use and educational achievement, adjusting for sex, parental covariates, and adolescent externalizing behavior. A total of 4761 individuals (49.4% females) were included in the analyses. Educational achievement was assessed with teacher-reported grade point average at ages 12 and 14, and with self-reported student status and completed education at age 17 and in young adulthood. From self-reported questionnaire items, frequency of any drinking, frequency of drinking to intoxication, any smoking and daily smoking were analyzed.

Results: Alcohol use and smoking behaviors at ages 12 and 14 predicted lower educational achievement at later time points even after previous achievement and confounding factors were taken into account. Lower school achievement in adolescence predicted a higher likelihood of engaging in smoking behaviors but did not predict later alcohol use. Higher educational attainment at age 17 predicted more frequent drinking in young adulthood.

Conclusions: Adolescent drinking behaviors are associated with lower future educational achievement independently of prior achievement, whereas smoking both predicts and is predicted by lower achievement. Early substance use indexes elevated risk for poor educational outcomes.

Keywords: Adolescent; Alcohol use; Cross-lagged model; Educational achievement; Longitudinal; Smoking.

Figure 1

Full cross-lagged path model for educational achievement and substance use. d, residual variance.

Figure 2

Cross-lagged path model for educational achievement and frequency of drinking. Statistically significant standardized path coefficients are shown. The covariates sex, parental education, parental drinking and parental smoking were included in the model but are omitted from the figure. d, residual variance; * p < 0.05; ** p < 0.01; *** p < 0.001.

Figure 3

Cross-lagged path model for educational achievement and frequency of intoxicating. Statistically significant standardized path coefficients are shown. The covariates sex, parental education, parental drinking and parental smoking were included in the model but are omitted from the figure. d, residual variance; * p < 0.05; ** p < 0.01; *** p < 0.001.

Figure 4

Cross-lagged path model for educational achievement and any current smoking. Statistically significant standardized path coefficients are shown. The covariates sex, parental education, parental drinking and parental smoking were included in the model but are omitted from the figure. d, residual variance; * p < 0.05; ** p < 0.01; *** p < 0.001.

Figure 5

Cross-lagged path model for educational achievement and daily smoking. Statistically significant standardized path coefficients are shown. The covariates sex, parental education, parental drinking and parental smoking were included in the model but are omitted from the figure. d, residual variance; * p < 0.05; ** p < 0.01; *** p < 0.001.