PERHAPS the most popular issue in the field of criminal justice today is selective incapacitation—the policy of singling out for detention those criminals who are most likely to repeat their crimes. Given the limited capacity of prisons and jails, it is impossible to incarcerate all those who deserve it. Thus, it is argued, particular attention must be paid to detaining those who pose the greatest risk of repeated, serious offenses. The question, of course, is how to identify these offenders, and one answer is through statistical “prediction.” Sophisticated batteries of predictive variables have been developed to try to determine whether a given defendant is likely to offend again. Advocates urge the use of forms of prediction tests both before adjudication (to detain the most likely offenders) and after (to incarcerate them).