Aims: The first longitudinal investigation of the extent to which same-age and older schoolmates' smoking and non-smoking are associated with adolescents' smoking transitions during three grade intervals.
Design: Same-age and older schoolmates' smoking and non-smoking were assessed when adolescents were at grades 5 (age 10), 7 (age 12) and 9 (age 14). Adolescents' smoking transitions were assessed at three grade intervals: 5th-7th (age 10-12), 7th-9th (age 12-14) and 9th-12th (age 14-17).
Setting: Forty Washington State school districts.
Participants and measurements: Smoking questionnaire data were gathered on a cohort of adolescents (n = 4354 for same-age schoolmate analysis; n = 1833 for older schoolmate analysis) that was 49% female and 91% Caucasian.
Findings: No significant evidence that same-age schoolmates' smoking or non-smoking was associated with any of the adolescent smoking transitions at any of the three grade intervals. In contrast, the probability that each older schoolmate's smoking was associated with the adolescent making the transition to trying smoking was 1% (95% CI: 0.4%, 1.5%) and with the transition from trying to monthly smoking was also 1% (95% CI: 0.2%, 2.0%) during the 7th-9th grade (age 12-14) interval. Moreover, each older schoolmate's non-smoking was associated with a 1.001-1.006 (all P < 0.05) relative risk of an adolescent not trying smoking or escalating from trying to monthly smoking at several grade intervals.
Conclusions: Interventions should perhaps focus on the influence of both smoking and non-smoking older schoolmates during late childhood and early adolescence.