DescriptionThe current study drew on the Theory of Planned Behavior (TPB) to model self-change behavior in a sample of self-defined smokers over a 6 month follow-up period. The study tested the validity of the TPB model and the validity of the Expanded Model which proposed the addition of Impulsivity and Implicit Smoking Self-Identity as predictors of smoking outcome. The sample was recruited via an online advertisement and consisted of current smokers over18 years of age reporting no terminal illness. The baseline sample consisted of 81female and 49 male participants and the follow-up sample consisted of 28 female and 14 male participants. The study was administered online. Data consisted of demographic variables, dependence severity, and perceived stress. The TPB model variables were Smoking Attitude, as measured by Semantic Differential scales, Social Norm and Intent to Quit, measured using item ratings, and Self-Efficacy. Smoking outcome was measured as time to first quit attempt, number of quit attempts in the last 6 months, and longest period of abstinence in the last 6 months (dichotomized into high and low abstinence groups because of severe skew). A linear regression was run to test the first step of the TPB, predicting Intent to Quit from Personal Attitude, Social Norm, and Self-Efficacy. The current study found support for the first step of the TPB model- Personal Attitude and Social Norm predicted Intent to Quit. A series of regressions were performed to test the second step of the model, predicting Smoking Outcome from Intent to Quit, Self-Efficacy, Impulsivity, and Smoking Self-Concept. Partial support was found for the second step of the model, with Intent to Quit and Impulsivity predicting smoking outcome. Smoking Self-Concept was found to be predictive at a trend. Self-efficacy was not found to be predictive of either Intent to Quit or smoking outcome. An attrition analysis was performed to investigate predictors of study participation at follow-up. Support for the construct validity of the Smoking Self-Identity IAT was found; it was uncorrelated with Explicit Attitude and Social Desirability Bias and negatively correlated with smoking outcome. Limitations of the sample and implications for future research are discussed.