Table 3.
Logistic Multilevel Models Predicting Log Odds of Ever Using Stimulants to Get High
| Parameter | b | SE | p | OR | 95% CI OR | |
|---|---|---|---|---|---|---|
| Intercept | − 4.14 | 0.24 | < .001 | 0.02 | (0.01, | 0.03) |
| Age | 0.20 | 0.06 | < .001 | 1.22 | (1.09, | 1.36) |
| Peer Modeling | 0.17 | 0.02 | < .001 | 1.19 | (1.14, | 1.23) |
| Parental Monitoring | − 0.04 | 0.02 | .020 | 0.96 | (0.92, | 0.99) |
| School Performance | − 0.08 | 0.07 | .250 | 0.92 | (0.80, | 1.06) |
| Region | ||||||
| Northern Plains | 0.16 | 0.25 | .526 | 1.17 | (0.70, | 1.95) |
| Northwest | 0.58 | 0.59 | .333 | 1.79 | (0.53, | 6.10) |
| Southeast + Texas | − 0.78 | 0.80 | .342 | 0.46 | (0.09, | 2.42) |
| Upper Great Lakes | 1.14 | 0.27 | < .001 | 3.13 | (1.80, | 5.42) |
| Southwest | ||||||
| Female gender | 0.08 | 0.19 | .682 | 1.08 | (0.75, | 1.56) |
| Male gender | ||||||
| Prescribed by doctor | 2.17 | 0.21 | < .001 | 8.79 | (5.86, | 13.18) |
| Not prescribed by doctor | ||||||
Note. OR = odds ratio.CI = confidence interval. ORs are exponentiated parameter estimates (b’s). Continuous predictors were centered at the grand mean. Models estimate the log odds conditional on the random effects. Bolded typeface indicates fixed effects that were significant at the p < .05 level.