Descriptive statistics were calculated using the Statistical Package of Social Sciences version 21.0 for Windows (SPSS Inc., Chicago, IL). All associations were examined using Mplus v6.0 (Muthén and Muthén, 1998–2012), which allows for clustering of data due to familial relatedness via estimation of robust standard errors. Separate multinomial logistic regression analyses, with and without covariates, were used to examine the relationship between cannabis involvement and SI, and suicide plan/attempt. To examine whether the strength of the association (i.e. ORs) across levels of cannabis involvement (e.g. cannabis use vs. 1–2 CUD symptoms vs. ≥3 symptoms) was significantly different (i.e. a dose response relationship), the magnitude of the association (i.e. OR) between STB and each level of cannabis involvement was equated to all other levels of cannabis involvement (e.g. ORcannabis use=OR1–2 CUD symptoms; Satorra, 2000).
