Hypotheses from General Strain theory are addressed using data from a random sample of adults in Raleigh, NC. Analyses examine three issues: (1) whether strain predicts self-projected criminal behavior; controlling for past self-reported crime; (2) whether negative emotions mediate the relationship between strain and projected crime; and (3) whether social support and criminal peers serve as contingencies or mediators for strain in predicting criminality. Results are generally consistent with previous studies focusing on youth. Three of four measures of strain are found to predict the crime measures. However, that relationship is not mediated by negative emotion and the measures of social support and criminal peers do not act as contingencies or mediators. The results suggest that strain may not operate through negative emotions and that theoretical refinement is needed to identify which potential contingencies are likely to be operating under various circumstances.