After statistical corrections for selection into employment, job quality was found to reduce the likelihood of economic and non-economic criminal behavior among a sample of released high-risk offenders. None of the most salient alternative explanations--sample selection, human capital accumulation, personal expectations, external labor market effects or prior criminality--appeared to diminish the job quality effect. Assuming that assigning offenders to high-quality jobs would reduce their likelihood of recidivism, policymakers must determine if they can justify allocating the best jobs (or the training required to access them) to the least deserving members of a large and needy underclass population. Even if cost-benefit analysis revealed a net economic gain to such a program, it would be difficult to justify on equity grounds without extending such opportunities to other groups. Perhaps work programs for ex-offenders would be most effective if they were embedded in a comprehensive, and correspondingly costly, national employment, and training strategy. The paper suggests that entry of ex-offenders into high-quality jobs may increase social controls, decrease the motivation to commit crime, and thereby alter the relative attractiveness of legal and illegal activities. Notes, tables, figure, references, appendix