Table 2.
Summary of Regression Models Predicting Past-Year, Weekly, or Daily Cannabis Use in Adolescence from Child Executive Function, Parental Cannabis Use, and Child-, Family-, and Parent-Level Covariates
| Model 1: Adjusting for basic covariates |
Model 2: Adjusting for parent substance use |
Model 3: Adjusting for adolescents’ alcohol use |
||||
|---|---|---|---|---|---|---|
| Past-year1 cannabis use outcome: | ||||||
| b (SE) | 95% CI | b (SE) | 95% CI | b (SE) | 95% CI | |
|
| ||||||
| Executive function score2 | 0.45 (0.22)* | 0.02, 0.87 | 0.38 (0.22) | −0.05, 0.80 | 0.18 (0.16) | −0.13, 0.50 |
| Male | −0.57 (0.38) | −1.32, 0.18 | −0.70 (0.37) | −1.43, 0.02 | −0.83 (0.31)** | −1.45, −0.22 |
| Cognitive ability | −0.01 (0.09) | −0.18, 0.17 | −0.02 (0.08) | −0.19, 0.14 | −0.04 (0.08) | −0.18, 0.11 |
| Parent education | 0.05 (0.19) | −0.33, 0.43 | 0.18 (0.17) | −0.15, 0.52 | 0.15 (0.16) | −0.16, 0.46 |
| Parent cannabis use | - | - | 0.53 (0.25)* | 0.03, 1.03 | 0.41 (0.21)* | 0.01, 0.82 |
| Parent alcohol use | - | - | 0.69 (0.26)** | 0.19, 1.19 | 0.37 (0.22) | −0.06, 0.80 |
| Adolescents’ alcohol use | - | - | - | - | 2.35 (0.43)*** | 1.50, 3.20 |
|
| ||||||
| Weekly cannabis use outcome: | ||||||
| b (SE) | 95% CI | b (SE) | 95% CI | b (SE) | 95% CI | |
|
| ||||||
| Executive function score2 | 0.69 (0.23)** | 0.24, 1.15 | 0.64 (0.24)** | 0.17, 1.12 | 0.51 (0.20)* | 0.12, 0.89 |
| Male | 0.01 (0.43) | −0.84, 0.86 | −0.12 (0.43) | −0.97, 0.73 | −0.35 (0.40) | −1.15, 0.44 |
| Cognitive ability | −0.01 (0.11) | −0.21, 0.20 | −0.03 (0.11) | −0.24, 0.19 | −0.04 (0.10) | −0.24, 0.16 |
| Parent education | −0.33 (0.24) | −0.80, 0.14 | −0.20 (0.23) | −0.66, 0.26 | −0.24 (0.23) | −0.70, 0.21 |
| Parent cannabis use | - | - | 0.66 (0.30)* | 0.07, 1.25 | 0.52 (0.27) | −0.01, 1.05 |
| Parent alcohol use | - | - | 0.66 (0.29)* | 0.09, 1.24 | 0.35 (0.29) | −0.21, 0.91 |
| Adolescents’ alcohol use | - | - | - | - | 2.03 (0.46)*** | 1.12, 2.93 |
|
| ||||||
| Daily cannabis use outcome: | ||||||
| b (SE) | 95% CI | b (SE) | 95% CI | b (SE) | 95% CI | |
|
| ||||||
| Executive function score2 | 0.76 (0.28)** | 0.20, 1.31 | 0.65 (0.25)* | 0.15, 1.15 | 0.59 (0.25)* | 0.11, 1.08 |
| Male | −0.16 (0.54) | −1.22, 0.90 | −0.30 (0.54) | −1.36, 0.75 | −0.59 (0.54) | −1.64, 0.46 |
| Cognitive ability | −0.20 (0.13) | −0.47, 0.06 | −0.22 (0.13) | −0.47, 0.03 | −0.22 (0.13) | −0.47, 0.03 |
| Parent education | −0.39 (0.30) | −0.99, 0.21 | −0.12 (0.28) | −0.66, 0.43 | −0.14 (0.28) | −0.69, 0.42 |
| Parent cannabis use | - | - | 0.81 (0.39)* | 0.05, 1.56 | 0.75 (0.39) | −0.01, 1.51 |
| Parent alcohol use | - | - | 0.81 (0.35)* | 0.13, 1.50 | 0.57 (0.35) | −0.12, 1.26 |
| Adolescents’ alcohol use | - | - | - | - | 1.60 (0.49)** | 0.64, 2.56 |
Notes. b (SE) = unstandardized beta (standard error); betas are in the logit scale and represent the increase in the log odds of cannabis use associated with a one unit increase on the predictor. CI = confidence interval. In two-level models, the dichotomous cannabis use outcome was regressed on individually grand-mean centered age at the within-person level. In Models 1–3, within-person effects of age were all b (SE) = .32 (.08) for past-year cannabis use, and ranged from b (SE) = .57 (.12) to .60 (.13) for weekly cannabis use, and from b (SE) = .72 (.14) to .75 (.16) for daily cannabis use, all p < .001. Results above denote the regressions of the random threshold of cannabis use on predictors and covariates at the between-person level. The threshold is interpretable as the expected value of the latent response variable at which an adolescent transitions to cannabis use at age 17.62 years (the mean age of the sample across the waves in adolescence). Models adjusted for mean age of observation at the between-person level (betas not shown).
Any use in the past year measured repeatedly across adolescence.
Higher scores indicate poorer executive function.
p < .05.
p < .01.
p < .001.