Objectives: To examine household structure when studying determinants of youth smoking, as the configuration of a family is an important factor in the etiology of adolescent problem behaviors.
Design And Sample: The study sample (n = 13,001) included respondents aged 12-19 years who were either living in two-parent households, single-parent households, or no-parent households, and with valid response to the smoking status questions from the Canadian Community Health Survey.
Measures: Multivariate logistic regression was used to test the presence and strength of the association between household structure and the likelihood of smoking while controlling for age, sex, household education, and exposure to secondhand smoking.