In his report, Greenwood claims to have developed a classification scheme that would enable criminal justice practitioners to determine which offenders should receive long, 'incapacitating' prison sentences and which can be sentenced to alternative programs or safely released to the community. If implemented in its purest form, he says, selective incapacitation could result both in a net reduction of crime in the community and in the number of offenders who would need to be incarcerated. Since the release of the Rand report, most criminologists agree that Greenwood's findings are incomplete, methodologically flawed, and do not justify his policy proposal. Some have also raised moral and legal objections to it. In this article we outline the history, criticism and impact of selective incapacitation. We find that there are no clear and forceful answers to the dilemmas posed by the Greenwood saga. However, the pressure on the research community to come up with quick and easy answers to complex social problems, we suggest, is less than subtle, particularly when money, reputation and the ability to do research hang in the balance. (Author abstract)