In separate mixed effect logistic regression models, we tested whether age at first cannabis use, DSM5 cannabis use disorder (CUD) diagnosis,34 and maximum duration of daily cannabis was associated with cannabis-related experiences, controlling for sex, age, genotyping array, lifetime use of any other illicit substances (including hallucinogens, stimulants, sedatives, opioids), and 10 genetic ancestry principal components as fixed effects, and for family ID as a random intercept. We also tested whether these risk factors were associated with reporting at least 1 of the 5 cannabis-related experiences (hereafter referred to as “any” cannabis-related experience). Next, we tested whether the schizophrenia PRS was associated with cannabis-related experiences; we controlled for the same covariates as above, and CUD diagnosis (a binary variable) as another fixed effect. Statistical analyses were conducted in R.35
