Abstract
This is a statistical study of the association of parole outcomes with a set of possible predictor variables that is accomplished by multidimen sional contingency table analysis. It is keyed to prior studies conducted by the Parole Decision Making Project of the National Council on Crime and Delinquency. These studies, based on data of approximately 2500 prisoners in the federal corrections system, suggest nine variables presently employed by the U. S. Parole Board in parole decisions. The present analysis, employing a different multivariate data technique, indicates only four of the nine variables are required. It also demonstrates that representing the four variables by a sum of scores can hide discrimination power available when each prisoner profile is viewed as a four-dimensional array of data. A Technical Appendix appears at the end of the paper and describes the methodology of con tingency table analysis within the context of prisoner profile data and parole outcome.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have