Early cannabis initiation and educational attainment: is the association causal? Data from the French TEMPO study

Int J Epidemiol. 2017 Oct 1;46(5):1641-1650. doi: 10.1093/ije/dyx065.


Background: Adolescent cannabis use has been reported to predict later educational attainment; however, results of past studies may be confounded by inappropriate control for factors that make some youths more likely to use cannabis precociously than others. We aimed to test the possibility of a causal relationship between early cannabis initiation and later academic achievement.

Methods: Analyses are based on data collected among TEMPO cohort study participants (France, 2009, n = 1103, 22-35 years). Participants were previously assessed in childhood (1991) and adolescence (1999); additionally, their parents had taken part in a longitudinal epidemiological cohort study (GAZEL). Early cannabis initiation was defined as use at age 16 or earlier. Educational attainment was defined as the completion of a high-school degree ('Baccalauréat'). Early (up to and including age 16 years) and late (after age 16 years) cannabis-use initiators were compared with non-users using logistic regression models controlled for inverse probability weights (IPWs) of exposure calculated based on participants' socio-demographic, juvenile and parental characteristics.

Results: In age- and sex-adjusted analyses, early cannabis initiators were more likely than non-users to have low educational attainment [odds ratio (OR): 1.77, 95% confidence interval (CI) 1.22-2.55]. In IPWs-controlled analyses, this association somewhat decreased (OR: 1.64, 95% CI 1.13-2.40). Late cannabis initiators did not have lower educational attainment than non-users. Early cannabis use and educational attainment appeared more strongly associated in young women than in young men.

Conclusions: Early cannabis can cause low educational attainment. Youths who initiate cannabis use early require attention from addiction and education specialists to reduce their odds of poor long-term outcomes.

Keywords: adolescent; cannabis; education; longitudinal cohort.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Academic Success*
  • Adolescent
  • Adult
  • Educational Status*
  • Female
  • France / epidemiology
  • Humans
  • Logistic Models
  • Longitudinal Studies
  • Male
  • Marijuana Smoking / epidemiology*
  • Marijuana Smoking / psychology*
  • Risk Factors
  • Sex Factors
  • Socioeconomic Factors
  • Young Adult