According to self-efficacy and relapse theories. abstinence self-efficacy (ASE) ratings should be context-specific: they should vary across situations. This variability may be important, as it could signal high-risk for relapse situations. In this study, confirmatory factor analysis was used to identify situational variability in a novel ASE assessment (Relapse Situation Efficacy Questionnaire. or RSEQ). Results supported a hierarchical structure, where both context-specific and unidimensional measures of ASE exist within the assessment. Context-specific factors included Negative Affect, Positive Affect, Restrictive Situations (to smoking). Idle Time, Social-Food Situations, Low Arousal, and Craving. Multiple context-specific factors and the aggregate factor predicted cessation outcome, even after controlling for concurrent smoking rate. However, the context rated with the least confidence proved to be the best outcome predictor, suggesting the existence of "Achilles' heel" situations. These data indicate the internal and predictive validity of the RSEQ.