Aims: To examine the effect of family and friends' smoking on uptake of regular smoking among young people from mid-adolescence to early adulthood; whether such effects are time limited, vary by social class and gender, and the extent to which uptake precedes or follows friends' smoking.
Setting: The West of Scotland.
Design and participants: A longitudinal survey of 1009 15-year-olds interviewed at baseline, followed-up at 16, 18, 21 and 23 years of age, using interviews and postal questionnaires.
Measurements: Self-reported measures of smoking were used, allowing analysis of the uptake of regular smoking (1 + cigarette per day) between 15 and 23, and in the periods 15-16, 16-18 and 18-21 years. Measures of parental and sibling smoking and social class were obtained at baseline and respondents' reports of friends' smoking at 15 and 18 years.
Findings: Regular smoking more than doubled between 15 (14%) and 23 (36%). Adjusted for other factors, no independent effect of parental smoking on uptake in any period was observed, an effect of sibling smoking being confined to uptake between 15 and 16. Friends' smoking at 15 increased the likelihood of uptake up to 10 times over the next year, but did not extend to later years; that at 18 increased it up to three times between 18 and 21. These effects did not vary by gender or social class. Further analysis revealed a strong effect of friends' smoking at age 18 on earlier as well as later uptake.
Conclusions: Contrasting with prevalent assumptions, the period from mid-adolescence to early adulthood is important for uptake of regular smoking, and in particular reveals friends' smoking to be of continuing significance, especially around school-leaving when friendship networks often change markedly.