Background Research suggests greater densities of tobacco retail outlets in residential neighbourhoods are associated with higher rates of smoking among adolescents. Policies to limit the impacts of tobacco outlets on health requires better understanding of how they affect smoking behaviour. This analysis assesses the relationship between outlet density and Scottish adolescents’ awareness of tobacco products, beliefs about cigarette prices, attitudes towards smoking and access to cigarettes from retailers.

Methods We analysed data from 20,466 13 and 15 year old respondents in the 2010 Scottish School Adolescent Lifestyle and Substance Use Survey (SALSUS), a nationally representative cross-sectional survey. Outlet density was described by 10,161 addresses on the Scottish Tobacco Retailers Register, 2012. The postcodes of the children’s home addresses were linked to a spatially-weighted Kernal Density Estimation (KDE) measure of the density of outlets located within 400 m of their residence, aggregated to quintiles. Children’s knowledge of cigarettes was represented by the number of brands they could name, their beliefs regarding costs by estimated price of a packet of twenty, smoking opinions by counts of responses to 13 statements regarding smoking and ease of access to cigarettes by whether smokers had recently purchased cigarettes from shops themselves or through adult proxies. Pathway models described the links between outlet density and these outcomes, plus age, sex, ethnic group, family economic status, family structure, parental and friend smoking, neighbourhood deprivation and urban-rural status. Models were stratified by smoking status (never smoker/ever smoker/current smoker).

Results The pathway models indicated significant (p < 0.05), but moderate, direct effects of outlet density on knowledge of cigarette brands: with each increase in outlet density quintile the number of brands named increased by 0.04 and 0.08 among never and current smokers respectively. Outlet density did not have significant direct effects on any other outcomes, but in some models did have significant effects on smoking behaviour of children’s parents and friends, which in turn were related to some study outcomes.

Discussion Scottish adolescents living in areas with high outlet densities had greater knowledge of cigarette brands, but did not believe cigarettes to be cheaper, have more positive opinions about smoking, or report that cigarettes were easier to purchase. Outlet densities may influence adolescents indirectly though effects on their parents and friends smoking behaviour. This study is limited by its cross-sectional design; longitudinal data could assess whether outlet density is related to changes in smoking status over time.