The public pressure for a more aggressive response to crime in conjunction with the severe constraint of current limited prison capacity has yielded a search for means of increasing the crime-control effectiveness of the criminal justice system. Since the principal scarce resource is prison cells, this search can involve finding means for more efficient allocation of those cells. Pursuing that strategy inevitably introduces some conflict between the considerations of individualizing incarceration to maximize an incapacitation effect (reducing crimes against the public by constraining those most likely to commit such crimes) and those of ensuring equal, just, and uniform punishment. If an effective discrimination instrument were available, and if it were applied only to those convicted of offenses while the imposed punishment was no more severe than could reasonably be applied for the conviction offense, then most of the legal and philosophical objections to selective incapacitation can be accommodated. The crucial technical issue is the potential effectiveness of the decision instrument. The most immediate research task is to validate and test the generality of current findings on the characteristics of high-risk offenders, notably the relevant Rand research. Additionally, the research should be replicated, using samples from other States and focusing on offenders not in prison. As the results of the Rand work and related research identify improved selection criteria for candidates for incarceration, those criteria should be compared to current practice.