Multilevel analyses on the relations of school district''s characteristics to alcohol drinking in a 3-year national survey of adolescents

• Main Authors: I-Chin Lin, 林宜瑾
• Other Authors: Wei J. Chen
• Format: Others
• Language: en_US
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/09891620165291528069

碩士 === 國立臺灣大學 === 流行病學研究所 === 96 === Objective: To examine the effects of school-districts'' characteristics on adolescent alcohol use and whether the characteristics of school-districts can explain the school-district differences on adolescent alcohol use under the multilevel framework. Methods: In 2004-2006, a nationally representative sample of 53,597 school-attending adolescents from 148 school-districts was selected using multistage, random, cluster sampling. Information on adolescents'' alcohol use experiences, sociodemographic characteristics, and delinquent behaviors was collected through the self-administered questionnaire. Relations of three school-district characteristics (prevalence of perceived classmate'' alcohol use, density of convenience stores, and prevalence of drinking and driving) to adolescent alcohol use were examined by means of multilevel logistic regression analysis. Results: Prevalence of perceived classmates'' alcohol use, density of convenience stores, and prevalence of drinking and driving were significantly associated with current alcohol use among adolescents. The odds on current alcohol use would increase 1.45 times on average for adolescents in school-districts of higher risks comparing to the ones in school-districts with lower risks. While nearly 70% of residual heterogeneity among school-districts could be explained by the individual-level variables, only 10.20% to 12.24% of school-district variances were further explained by the three area-level variables. Conclusions: Adolescents whose schools located in townships with higher prevalence of perceived classmates'' alcohol use, higher density of convenience stores, or higher prevalence of drinking and driving had higher odds of using alcohol. Yet, prevalence of perceived classmates'' alcohol use, density of convenience stores, and prevalence of drinking and driving explained relatively limited school-district differences on adolescent alcohol use.