Family processes accounted for some of the family structure variation in these risk behaviors, as shown in Models 1b-3b. Adolescents outside of two biological married families remained more likely to smoke, although the magnitude of the effects declined modestly (Model 1b). The odds of smoking were about 1.3 times higher for adolescents in married stepfamilies or single-mother families and roughly double for teens in cohabiting stepfamilies relative to those in two biological married parent families. Yet, the odds for those who live in a cohabiting stepfamily no longer differed from the odds for those who reside in either a married stepfamily or single-mother family, suggesting this effect was driven by poorer socialization, socioeconomic disadvantage, and maternal smoking (all three sets of factors were operating as controlling for each individually did not reduce the positive effect of cohabiting stepfamilies [result not shown]). Maternal support and control were both negatively associated with smoking, whereas maternal smoking was positively related to smoking. Adolescents whose mothers had a college degree were less likely to smoke.
