Early-life characteristics and educational disparities in smoking

•Educational gradients in smoking start in adolescence, before schooling is completed.•Childhood characteristics explain only a modest share of smoking initiation gradients.•Adolescent characteristics account for a large share of gradients in initiation.•Disparities in quitting exist for men and are not explained by life-course factors.

Educational inequalities in adult health outcomes are well-established, but it remains unclear when and how these disparities emerge across the life course. We use the 1970 British Cohort Study (BCS) to examine the links between early-life characteristics (cognitive, social and emotional, familial, peer, socioeconomic, school- and health-related) and emergent disparities in smoking. We examine whether characteristics from childhood explain differences in smoking initiation at ages 16 and 26 by the education respondents eventually obtain. Using characteristics from ages 0, 5, 10, 16, and 26, our results reveal that cognitive, socioeconomic, and health-related factors together explain little of the education-smoking link. However, characteristics measured during adolescence such as school attachment and having friends who smoke account for about half (49%) of the education-smoking association at age 16 and most of the association (67%) at age 26. Models of smoking initiation between ages 16 and 26 reveal a similar pattern of results. In contrast, educational disparities in quitting at age 26 exist only for men and are not explained by observed life-course characteristics, including those from adolescence. Further analyses using an extended vector of life course covariates from age 10 uphold our findings.