The Effects of Peer Influence and Social Network on Adolescent Alcohol Use

Introduction: Adolescence is a critical period for understanding the nature, evolution, and etiology of alcohol use problems. A series of genetic studies indicated that the contribution of social environment has gradually declined in the progression from drinking initiation to alcohol use disorder, suggesting that social factors are particularly significant in shaping earlier stages of alcohol involvement. Research on peer influence in underage drinking has made prominent progress in the two decades, but is not yet fully comprehensive. On the basis of the longitudinal study in Taiwan, this study aims to investigate the extent to which peer characteristics and social network in early adolescence may affect subsequent alcohol drinking.

Methods: We used two waves of longitudinal data from the Alcohol-Related Experiences among Children (AREC II). The baseline sample comprised 1926 7^th graders (~13 years old) from 11 public middle schools in Taipei via multi-stage sampling (response rate=55%); subsequent follow up was conducted at 9^th grade (follow-up rate: 97%). Data concerning individual sociodemographic background, family structure and drinking behaviors, peer characteristics, and drinking experience were collected via web-based self-administered questionnaires. Building upon the maximally five friends nominated by young respondents, the social network structure and characteristics were obtained via the UCINET. Complex surveys analyses were used to evaluate relationship estimates.

Results: At 9^th grade, approximately two thirds of students have tried alcohol drinking at least once, with 8.54%_wt having ever drunk 4 or more times (moderate drinking) in the past year. Having observed peer drinking or having received alcohol offer from peers at 7^th grade was consistently associated with 2~3-fold of alcohol drinking behaviors at 9^th grade. After statistical adjustment for an array of confounding factors at 7^th grade, having peers with attitudes against drinking was linked with reduced alcohol drinking by 12% for occasional drinking (drinking for 1-3 times) and 26% for moderate drinking (p <0.001). In terms of social network attributes, the students with higher betweenness centrality were more likely to engage in drinking behaviors (e.g., drinking 4 or more times; aOR= 1.09, p<0.05); the higher degree centrality is linked with subsequent risk for moderate drinking (aOR= 1.06, p<0.05). 

Conclusions: Our results demonstrate that peer characteristics and social network attributes may affect one’s involvement of alcohol drinking through adolescence. To devise program targeting underage drinking and problems, preventive strategies may consider address peer influence and integrate social network via the lens of dynamic socioecological model.