The moderator explains the Federal parole guidelines chart, which assigns points to persons being considered for parole, based upon factors found to be statistically related to successful parole outcomes. Those receiving high scores are more likely to be paroled. When questioned about the reliability of this chart, panelist Peter Hoffman, research director of the U.S. Parole Commission, notes that those receiving high scores have an average 6-percent recidivism rate, and those having low scores have an average 50-percent recidivism rate. When questioned about the moral and legal issues of using these parole guidelines, Hoffman argues that since the commission is charged with making decisions about the release of serious offenders, it is better to make those decisions scientifically than subjectively. Further, he indicates that the reliability of the chart makes it less objectionable than if it was a poor predictor of recidivism. Panelist Peter Greenwood of the Rand Corporation, who was involved in the development of the California parole guidelines, indicates that his scale is similar to the one used by the Federal Parole Commission. He advises that criminal behavior patterns differ from State to State, however, and that scales used by the various States should reflect this. Panelist John Monahan of the University of Virginia Law School advises that predictive scales based upon statistical analysis are more reliable in predicting criminality than are clinical judgments. All three panelists believe that predictive scales can and should be used as sentencing guidelines, but Monahan cautions that sentencing paramaters should be determined by the severity of the crime. The sentence given within those parameters (from the minimum to the maximum sentence for a given offense) would be determined by the guidelines.