Table 4.
Interaction effect estimates from the alternative multinomial logistic regression * with main effects (age, gender) and three-way interaction (perceived academic achievement × gender × FAS).
| Gender and Family Affluence (FAS) Level in the Interaction | Cluster 2 Mainly Tobacco and Alcohol | Cluster 3 High Alcohol and Cannabis | Cluster 4 Poly-Users | ||||||
|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | OR | 95% CI | ||||
| Boys × Low FAS | 2.06 | 0.82 | 5.21 | 1.47 | 0.61 | 3.56 | 0.59 | 0.13 | 2.73 |
| Boys × Average or high FAS | 1.61 | 0.81 | 3.20 | 1.06 | 0.56 | 2.01 | 1.72 | 0.83 | 3.53 |
| Girls × Low FAS | 2.41 | 0.98 | 5.95 | 2.19 | 0.66 | 7.35 | 5.55 | 1.97 | 15.66 |
| Girls × Average or high FAS | 5.22 | 2.99 | 9.13 | 2.17 | 0.96 | 4.91 | 3.60 | 1.55 | 8.33 |
* FAS—family affluence; reference category Cluster 1—non-users; results significant at p < 0.05 are bolded; OR—adjusted odds ratio; CI—confidence interval; OR associated with poor perception of school achievements.