Does the association between different dimension of social capital and adolescent smoking vary by socioeconomic status? a pooled cross-national analysis.

Objectives: To analyze how dimensions of social capital at the individual level are associated with adolescent smoking and whether associations differ by socioeconomic status.
Methods: Data were from the 'Health Behaviour in School-aged Children' study 2005/2006 including 6511 15-year-old adolescents from Flemish Belgium, Canada, Romania and England. Socioeconomic status was measured using the Family Affluence Scale (FAS). Social capital was indicated by friend-related social capital, participation in school and voluntary organizations, trust and reciprocity in family, neighborhood and school. We conducted pooled logistic regression models with interaction terms and tested for cross-national differences.
Results: Almost all dimensions of social capital were associated with a lower likelihood of smoking, except for friend-related social capital and school participation. The association of family-related social capital with smoking was significantly stronger for low FAS adolescents, whereas the association of vertical trust and reciprocity in school with smoking was significantly stronger for high FAS adolescents.
Conclusions: Social capital may act both as a protective and a risk factor for adolescent smoking. Achieving higher levels of family-related social capital might reduce socioeconomic inequalities in adolescent smoking.