Source: (2003) Social Problems. 50(4): 592–617.Using self-projected measures of criminal behavior obtained from a telephone survey, we investigate the predictive ability of four variables central to individual-level application of Braithwaite’s shaming theory: participation in gossip, having been reintegratively shamed, having been the object of gossip, and having been stigmatically shamed. The results are partially supportive of the theory. Contrary to expectations, the theoretical inhibitory shaming variables (participation in gossip and being reintegratively shamed) do not reduce projected misbehavior. Consistent with predictions, however, theorized crime-generative shaming variables (having been the object of gossip and having been stigmatically shamed) do appear to increase the chances of several forms of misbehavior. In addition, individuals’ interdependency does not enhance the effects of the inhibitory shaming variables and has only limited and contradictory effects in enhancing the effects of crime-generative shaming variables. Results suggest that the theory needs clarification and that it may need further refinement to specify more carefully the conditions under which shaming processes inhibit or enhance criminal probabilities. Author's abstract.