This study involved 137 participants who were assessed on 12 relevant predictor variables and then randomly assigned to social support or self-control treatment. Outcome across treatments was predicted by cotinine levels, treatment history, previous abstinence, happiness, self-efficacy, and perceived social support for quitting. Social support treatment was more effective than self-control treatment for participants with high baseline self-control orientation scores and participants with high self-efficacy scores. All other hypothesized Subject x Treatment interaction effects were nonsignificant. The study provided an example of a coherent approach to matching research and demonstrated the difficulty involved in providing treatments that are different enough from each other to benefit different smokers. Matching research has important theoretical value but may have limited potential for improving smoking treatment effectiveness.