Certain behavioral and emotional factors at the individual level have been found to predict tobacco use longitudinally. In addition to the association between poor school performance and tobacco use, Wang and colleagues (1999) also found that non-smoking adolescents who exhibited signs of depression at ages 12 to 15 years had an increased risk of becoming a regular smoker by follow-up at ages 15 to 18 years. A second study, of more than 2,000 14- and 15-year-old students in Australia, also found that signs of depression and anxiety predict smoking initiation longitudinally, but only among those adolescents whose friends smoke (Patton, et al., 1998b). The results of this Australian study further suggest that the relationship between depression and tobacco use is most pronounced among adolescent girls. While the reasons for this relationship have not been established, it is possible that depressed adolescents may try to self-medicate with tobacco and/or they may have a heightened susceptibility to peer smoking influences (Patton, et al., 1998b). Other behavioral and emotional factors may also be important. Lynskey and Fergusson (1995), for example, followed more than 900 New Zealand children for seven years, and found that children with conduct problems at age 8 were nearly twice as likely to become daily smokers by age 15, when compared with their peers without conduct problems at baseline. In an additional study, 5th graders and 7th graders (15% Asian, 32% Hispanic, 38% white, 4% black, 11% other or unknown) who displayed disruptive behavior were more likely than were their peers to initiate a smoking habit by 8th or 9th grade (Cohen, Richardson, & LaBree, 1994).