Based on the correlation between offender data and self-reported crime for 2,200 inmates in California, Michigan, and Texas, Greenwood identified 7 factors that significantly impact the frequency with which an offender commits crime. The factors relate to prior juvenile and adult arrests, employment patterns, and drug abuse. Greenwood theorizes that an instrument based on these factors could identify offenders likely to commit a high rate of crimes in the future so as to ensure their imprisonment (selective incapacitation) To test the effectiveness of the selective incapacitation theory, Greenwood cites California data to show that more certain and lengthy incarceration has reduced crime over time in that State. Greenwood also suggests using preventive intervention for high-risk groups through such programs as Head Start, parent training, and more effective schools. Panelists' comments on Greenwood's presentation address his tendency to focus on the past as an automatic predictor of the future rather than on the current circumstances of the offender, the appropriateness of basing sentencing on factors other than seriousness of the current offense, and the need for experimental testing of incapacitation theory.