Schedule:
Wednesday, May 28, 2014
Columbia A/B (Hyatt Regency Washington)
* noted as presenting author
The aim of this study was to estimate the prevalence of early drinking among Norwegian adolescents and to test models for predicting early drinking onset. The study included 1550 eight graders from 41 high schools in Oslo and neighbouring municipalities with a mean age of 13.5 (SD = 0.7) years and equal gender distribution. Data were collected using online questionnaires. Self-reports showed that 24% had drunk at least one glass of alcohol. Significantly more boys (29%) than girls (19%) reported drinking. Predictors of early drinking onset were identified by generalized linear mixed models with two multivariable models applied. Social and environmental variables were entered into the multivariable model at first level (first model). The amount of correctly classified subjects totaled to 82.2% whereas the first model correctly classified only 29.1% of those having been drinking alcohol at least once. Variables predicting drinking significantly comprised gender (F = 9.3, p < .01), religion (F = 8.7, p < .001) smoking (F = 97.8, p < .001), and having bullied others (F = 4.9, p < .01). By adding at the second level to the predictor set psychosocial variables such as intentions to drink, expectancies, attitudes and norms towards drinking to the multivariable model resulted into a significant improvement in model fit and increased the percentage of correctly classified subjects to 86.3%, in general. The second model correctly classified 56.2% of those having been drinking alcohol at least once. Except for bullying, which was non-significant in the second model, gender, religion and smoking were still found to be significantly associated with drinking, as well social positive expectancies towards drinking (F = 8.0, p < .01), global positive expectancies (F = 4.8, p < .05), positive attitudes towards drinking (F = 20.7, p < .001), intentions to drink (F = 6.3, p < .05) and subjective norms about drinking (F = 63.7, p < .001). Findings of the study are of relevance both for identifying adolescents at risk for early drinking and for preventive efforts. In particular, not only socio-economic and behavioral factors were found significantly associated with early drinking but also psychosocial variables. The results expand previous research about a consistent relationship between early drinking and role of adolescents’ expectancies, attitudes, norms and intentions and encourage the use of programs addressing different psychosocial variables in order to prevent early drinking.