Although smoking relapse is the most frequently reported outcome of smoking cessation, with reported rates as high as 83% to 89% depending on the treatment format tried (Fiore et al., 2000), understanding the factors associated with relapse is not simple. The purpose of this study was to identify predictors of smoking relapse in a Canadian sample of former smokers. A secondary analysis was conducted of data collected in the Survey on Smoking in Canada, a national survey panel with four-cycle intervals in 1994-1995. In this study, all respondents who reported that they were former smokers at cycle one (N = 3,875) were included and divided into two groups. One group (N = 3,582) remained abstinent from smoking throughout the survey's four cycles, and a second group (N = 293) experienced a smoking relapse sometime between cycles one and four. Multiple logistic regression analysis of sociodemographic variables indicated that age, education, marital status, and employment status were associated with relapse. Differences were found between the two groups (relapsers versus non-relapsers) with respect to the presence of other smokers in the household. Differences were also noted between the groups when comparing subjects' exposure to smoke from cigarettes (p < .001); if subjects were bothered by cigarette smoke (p < .001); and if cigarette smoke caused any physical irritation (p < .001). Relapsers also differed from non-relapsers in that they reported smoking initiation at a much younger age. In all four cycles of follow-up, relapsers reported that "stress" was the primary reason for their smoking relapse.