Short communicationAre area-level effects just a proxy for school-level effects? Socioeconomic differences in alcohol consumption patterns among Swedish adolescents
Introduction
The socioeconomic pattern of alcohol consumption is inconsistent compared to many other health behaviours (Harper and Lynch, 2007), particularly in the adolescent population (Hanson and Chen, 2007). Although certain studies have revealed a protective effect of growing up in families with low socioeconomic status, other investigations have highlighted this type of upbringing as a risk factor (Lemstra et al., 2008). In addition to these individual-level associations, researchers have investigated whether socioeconomic disadvantage at the contextual level can play a role in alcohol consumption. Two recent studies examined adolescents living in Oslo, Norway, and demonstrated that the highest levels of alcohol consumption and alcohol intoxication were reported by residents of more affluent city districts while the opposite pattern was observed for alcohol problems (Pedersen and Bakken, 2016, Pedersen et al., 2015). A third study revealed significant variation in adolescent alcohol use across neighbourhoods in the American city of Chicago, Illinois. However, except for immigrant concentration, which was negatively associated with the measured outcome, no indicator of neighbourhood socioeconomic disadvantage had any significant effect (Fagan et al., 2015). Despite the numerous differences between these studies with respect to measurements and design, these investigations shared one feature: neither study addressed schools, which provide another important context in adolescents’ lives. Other research has indeed identified significant school-to-school differences in alcohol consumption that have largely been driven by the lower rates of alcohol consumption in socioeconomically disadvantaged schools (Olsson and Fritzell, 2015). Given the burgeoning number of studies that have concluded that area-level effects in adolescent populations generally are simply a proxy for school-level characteristics (Bernelius and Kauppinen, 2012, Brännström, 2008, Rendón, 2014, Sykes and Musterd, 2010), the current investigation set out to simultaneously assess variations in adolescent alcohol use and binge drinking at the city district and school levels. Moreover, the aim was to analyse whether any such variations could be ascribed to the socioeconomic characteristics of the examined city districts, schools, and students.
Section snippets
Methods
The present study included data at the student level, the school level, and the administrative city district level of schools. The student-level data were derived from the Stockholm School Survey 2014, which was conducted among 9th and 11th graders attending schools located in the municipality of Stockholm, Sweden (n = 15,169). While all public schools in Stockholm municipality were urged to participate, private schools participated voluntarily. The response rate was 76% (n = 11,507). The present
Results
Table 2a demonstrates the results for alcohol use. The intra-class correlations (ICC) reflect the variation in alcohol use at the contextual levels. According to Model 1, the ICC for the city district level is 0.03 and significant, indicating that there is significant variation in alcohol use among city districts. Model 2 additionally accounts for the school-level ICC; although significant variation among city districts remains (ICC = 0.02), more variation is observed among schools (ICC = 0.04).
Discussion
Our findings suggest that it is more important to consider the school than the city district when assessing contextual variations in adolescent alcohol use in general and binge drinking in particular. Thus, failure to account for the school context may have caused past studies to overestimate city district-level differences in alcohol consumption among adolescents (Pedersen and Bakken, 2016, Pedersen et al., 2015). Furthermore, in an attempt to explain contextual variations in alcohol use and
Role of funding source
Nothing declared.
Contributors
Per Carlson stands behind the basic research idea, the statistical analyses and writing the methods section. Ylva B Almquist has gone through previous research and has written large parts of the introduction and the discussion. Both authors have approved the final article.
Conflict of interest
No conflict declared.
Acknowledgements
This study was funded by The Alcohol Research Council of ‘Systembolaget’ (SRA). Thank you to the City of Stockholm and The Social Development Unit for providing the data.
References (13)
- et al.
A multi-level analysis of the impact of neighborhood structural and social factors on adolescent substance use
Drug Alcohol Depend.
(2015) - et al.
Drinking pattern among adolescents with immigrant and Norwegian backgrounds: a two-way influence?
Addiction
(2005) - et al.
School outcomes and neighbourhood effects: a new approach using data from Finland
Making their mark: the effects of neighbourhood and upper secondary school on educational achievement
Eur. Sociol. Rev.
(2008)- et al.
Socioeconomic status and health behaviors in adolescence: a review of the literature
J. Behav. Med.
(2007) - et al.
Trends in socioeconomic inequalities in adult health behaviors among U.S. States, 1990S. States, 1990–2004
Public Health Rep.
(2007)