We used one-way ANOVA, t-test, Pearson product-moment correlation and multiple regression statistics to exam the associations between staff's smoking status, smoking-related attitudes and knowledge, and their opinions of smoke-free schools.