NCJ Number
66907
Date Published
1978
Length
101 pages
Annotation
RESULTS ARE REPORTED FROM A TEXAS FEASIBILITY STUDY AND PRODUCT DEVELOPMENT PROGRAM TO IMPROVE THE ACCURACY OF PAROLE DECISIONS AND TO ORGANIZE THE AVAILABLE INFORMATION TO EXPEDITE THE DECISION PROCESS.
Abstract
THE INFORMATION DEVELOPED CONCERNS EXPECTED FUTURE BEHAVIORS OF PAROLEES. THE PRODUCT DEVELOPED PROVIDES DECISIONMAKERS WITH A SERIES OF ESTIMATIONS OF RELATIVE RISKS OF FAILURE OF A PAROLE ELIGIBLE INMATE; THAT IS, IF PAROLED, THE PROBABILITY OF A GIVEN PERSON BEHAVING IN MALADAPTIVE WAYS HISTORICALLY RELATED TO A RETURN TO CRIME AND PRISON IS INDICATED. THE INDEPENDENTLY ESTIMATED RISKS COVER THE POTENTIAL BEHAVIORS OF ASSAULTIVE ACTS, POORLY CONTROLLED AGGRESSIVE TENDENCIES, AND VARIOUS MALADJUSTMENTS TO THE RESTRICTIONS OF THE PAROLE SYSTEM, SOCIETY, AND WORK. THE STATISTICAL PREDICTIVE PROCCEDURES USED WERE STANDARD (MULTIPLE DISCRIMINANT FUNCTIONS ANALYSIS AND MULTIPLE REGRESSION ANALYSIS). THE VARIABLE-TO-SUBJECT RATIO WAS KEPT HIGH BY LIMITING THE NUMBER OF VARIABLES PERMITTED TO ENTER AN EQUATION, THUS REDUCING THE OVER-PREDICTION PROBLEM. LARGE OFFENDER SAMPLES, COLLECTED FROM ACROSS ALL PRISON UNITS OF THE TEXAS DEPARTMENT OF CORRECTIONS, WERE USED FOR EACH BEHAVIOR PREDICTED. THE COMPARATIVE 'OUTSIDE' SAMPLES USED WERE ALSO LARGE AND CONSISTED OF WORKERS ACROSS MANY BUSINESS AND INDUSTRIES OF THE TYPE AND AT THE JOB LEVEL PAROLEES WOULD BE EXPECTED TO SEEK AND OBTAIN. A BATTERY OF 11 PREDICTOR EQUATIONS WAS DEVELOPED. ALL WERE CROSS-VALIDATED WITH OTHER COMPARABLE SAMPLES AT STATISTICALLY SIGNIFICANT LEVELS AND WERE SUBJECTED TO RIGOROUS FIELD TESTS COMPARABLE TO CONDITIONS THAT MAY BE EXPECTED TO OCCUR. CURRENTLY, THE PREDICTORS ARE RECOMMENDED FOR USE ONLY WITH DIFFICULT CASES; I.E., WHEN THE EXISTING SYSTEM DOES NOT PROVIDE A CLEAR-CUT DECISION. FOLLOWING CAREFULLY MONITORED USE AND IMPROVEMENT, HOWEVER, THE PRODUCT CAN BE EXPECTED TO BECOME A PART OF THE OVERALL DECISIONMAKING PROCESS. TRIAL USE AND REVISIONS OVER A 2-YEAR PERIOD ARE RECOMMENDED. INSTRUMENTS AND DETAILED PROCEDURES USED IN THE PROJECT ARE PROVIDED IN THE APPENDIXES. (RCB)