Identification and Prediction of Drinking Trajectories in Early and Mid-Adolescence

Abstract

The aim of this study was to identify subgroups of early and mid-adolescents with different drinking trajectories. In addition, we examined whether gender, parental, and peer factors predicted adolescents' membership of these drinking trajectories. We used longitudinal data of 428 families (fathers, mothers, mid-adolescents, and their younger siblings). Latent Class Growth Analyses were performed to identify drinking trajectories. Four drinking trajectories emerged for early adolescents: abstainers, light drinkers, increasers, and heavy drinkers. For mid-adolescents, we identified a fifth group (stable drinkers) in addition to the four trajectories identified for early adolescents. Our results showed that being a boy, having a best friend or father who drinks heavily, and having parents who are permissive toward adolescents' alcohol creates increased risk for both siblings to attend the more heavy drinking trajectories.

Rutger Engels was supported by a fellowship from the Netherlands Organization for Scientific Research (NWO) during the preparation of this article.

Notes

¹The similarities in alcohol consumption of friends and adolescents could be due to either the influence of peers or to the selection of drinking friends (e.g., Bot, Engels, Knibbe, & Meeus, 2005).

²A correlation table including all model variables is available from the first author.

Note: BIC = Bayesian Information Criterion; AIC = Akaike's Information Criterion; H = entropy measure; LMR-LRT = Lo-Mendell-Rubin Likelihood-Ratio-Test.

Note: The drinkers in the left column are the reference classes. BF = Best Friend; F = Father; M = Mother; ∗p < .05. ∗∗p < .01. ∗∗∗p < .001.

³In preliminary analyses, we also controlled for parents' and adolescents' education. Because the education of all the family members was not significantly associated with any of the classes for both adolescents, we omitted these variables from further analyses.

⁴We also conducted multinomial logistic regressions with alcohol-specific rules based on the reports of the fathers and those of the mothers. The results based on parental reports were similar to those with the perspectives of the adolescents.

Note: The drinkers in the left column are the reference classes. ∗p < .05. ∗∗p < .01. ∗∗∗p < .001.

⁵In our study the older siblings drank significantly more often with their fathers than the younger siblings, t _T1(420) = 4.28, p < .001; t _T2(416) = 3.11, p < .01; t _T3(420) = 2.51, p < .05.

⁶In our study, both siblings drank significantly more often with their fathers than with their mothers, t _T1oldest(420) = 3.27, p < .01; t _T1youngest(419) = 2.10, p < .05; t _T2oldest(420) = 3.33, p < .01; t _T2youngest(422) = 2.85, p < .01; t _T3oldest(424) = 3.00, p < .05; t _T3youngest(416) = 3.11, p < .05.