Background: In order to design effective tobacco control policy in low income countries, it is essential to understand smoking prevalence and predictors. In Ethiopia, most of what is known on the prevalence of smoking comes from studies in larger towns. Little is known about predictors of smoking in any Ethiopian setting.

Objectives: The analyses reported were designed to determine smoking prevalence and social factors associated with ever smoking in Butajira town.

Methods: Cross-sectional study nested within a large questionnaire-based survey undertaken in Butajira, southern Ethiopia, between February and April 2003.

Results: Prevalence data were available on 1895 individuals aged 15 years and over. 15.4% of men and 0.2% of women had ever smoked, and 11.8% and 0.2% respectively, were current smokers. Using logistic regression, male gender (p