Current smoking among young adolescents: assessing school based contextual norms.Tob Control. 2004 Sep; 13(3):301-7.TC
To extend research on the relation of school based contextual norms to current smoking among adolescents by using three analytic techniques to test for contextual effects. It was hypothesised that significant contextual effects would be found in all three models, but that the strength of these effects would vary by the statistical rigor of the model.
Three separate analytic approaches were conducted on baseline self report student survey data from a larger study to test the relation between school level perceived peer tobacco use and individual current smoking status.
A representative sample of 5399 sixth through eighth grade students in 14 midwestern middle schools completed the survey. All enrolled sixth through eighth grade students were eligible to participate in the survey. The student participation rate was 91.4% for the entire sample, and did not differ significantly between the schools (range 82-100%).
MAIN OUTCOME MEASURE
Thirty day cigarette smoking prevalence.
A level 2 only model based on aggregated individual responses indicated that students in schools with higher average reported peer tobacco use were more likely to be current smokers than students in schools with lower average peer tobacco use. Using a level 1 only model based on individual responses indicated that the effect of school level perceived peer tobacco use on current smoking was significant when individual perceived peer tobacco use was excluded from the model but was non-significant when individual perceived peer tobacco use was added to the model. A multilevel model also indicated that the effect of school level perceived peer tobacco use on current smoking was not significant when individual perceived peer tobacco use was added to the model.
The analytic approach used to examine contextual effects using individuals' reports of peer tobacco use norms that were aggregated to obtain a context measure of the school norms may produce statistical artefacts that distort the association of the school context in general, and peer tobacco use norms in particular, with increased risk for current smoking beyond the risk associated with individual factors.