“The association between adolescent lifetime cannabis use and educational attainment was no longer significant after adjusting for potential confounders including alcohol use.”