“…Because age (OR = 1.35, 95% CI: 1.02-1.78), peer smoking (OR = 1.85, 95% CI: 1.46-2.35), and gender (OR = 0.42, 95% CI: 0.22-0.81) were statistically significant and important a priori covariates, they were retained in all subsequent models run. However, because socioeconomic status was not significant (OR = 1.26, 95% CI: 0.93-1.70), and a priori research has shown equivocal results on socioeconomic status being a risk, protective, or unrelated factor for tobacco use in Chinese adolescents (e.g., Johnson et al, 2006a;Gong et al, 1995;Hu et al, 1990;Pan, 2004;Sun and Ling, 1997;Zhu, Liu, Shelton, Liu, and Giovino, 1996), it was removed from subsequent models. …”