Overall, results of multivariate and chi-square analyses revealed that gender and race were significant predictors of recruitment to the smoking study. Female and White students were more likely to participate in the smoking study than male and nonwhite students. None of the demographic, smoking experience, or psychosocial variables were significantly related to smoking study retention. The findings suggest that tobacco use researchers who wish to establish a well-balanced sample might need to over-select potential male and non-White participants. The most appropriate recruitment strategy seems to be the use of a screening measure with an implied consent protocol. Participants were 5,298 8th and 10th grade students in the Chicago metropolitan area who completed a brief screening survey during 2000. Based on the survey results, 1,457 students who ranged from susceptible nonsmokers to experimenters to regular smokers were invited to participate in a longitudinal study on adolescent smoking behavior; 48.9 percent agreed. Survey questionnaires focused on demographic factors, smoking measures, nicotine dependence, and psychosocial measures. Follow-up research might learn more about barriers to non-White participation through focus group methodologies. Tables, references