WITH THE INCREASING AVAILABILITY OF COMPUTERS AND PROGRAM PACKAGES, THE DEVELOPMENT OF REGIONAL PREDICTION TABLES IS QUITE POSSIBLE AS WELL AS DESIRABLE. BY USING A LOCALLY DEVELOPED TABLE, OFFICIALS CAN TAKE INTO ACCOUNT REGIONAL FACTORS AFFECTING PAROLE OR PROBATION SUCCESS, CHANGES IN PENAL POLICY, AND CHANGES IN THE PRISON POPULATION. TO BE PRACTICAL, SUCH TABLES MUST HAVE THEIR VALIDITY CAREFULLY TESTED. TWO PERSONS SHOULD BE ABLE TO ARRIVE AT THE SAME SCORE WHEN THE TABLE IS USED CORRECTLY. THE TABLE SHOULD BE EASY TO USE AND MAKE MAXIMUM USE OF INFORMATION WITHOUT INTRODUCING COMPLEX CALCULATIONS. IT SHOULD NOT INCLUDE A PERSONAL JUDGMENT FACTOR. EXPERIENCE AND PERSONAL OPINION SHOULD BE USED IN CONJUNCTON WITH THE TABLE. THE COLLECTION STEP SHOULD INVOLVE GATHERING ALL INFORMATION ABOUT ALL AVAILABLE CASES IN THE REGION FOR A GIVEN PERIOD OF TIME. IT IS IMPORTANT THAT THE SAME TIME PERIOD BE OBSERVED FOR ALL CASES TO INSURE VALIDITY OF THE COMPUTATION. NEXT, THE PREDICTORS SHOULD BE DETERMINED. TO AVOID CHANCE RELATIONSHIPS, AT LEAST 30 CASES SHOULD BE AVAILABLE PER PREDICTOR VARIABLE. CHOOSING THE STATISTICAL METHOD TO BE USED IS DISCUSSED IN DETAIL. THE METHOD SELECTED WILL THEN DETERMINE THE CONSTRUCTION AND VALIDATION OF THE TABLES. THIS METHOD IS EXPLAINED STEP-BY-STEP USING A HYPOTHETICAL REGION. SUCH TABLES SHOULD BE EXAMINED EVERY 2 YEARS TO MAKE SURE THE ASSUMPTIONS ARE STILL VALID AND THE VARIABLES STILL GOOD PREDICTORS. PREDICTION TABLES ARE OFTEN USED TOGETHER WITH OTHER INFORMATION TO ESTIMATE THE PROBABILITY OF AN INDIVIDUAL'S SUCCESSFUL COMPLETION OF PROBATION. A MORE VALID USE OF SUCH TABLES WOULD BE THE ASSESSMENT OF VARIOUS PROGRAMS. THE TABLE COULD BE USED TO PREDICT SUCCESS RATES OF GROUPS, THEN SUCCESS RATES OF PERSONS WHO HAVE PARTICIPATED IN VARIOUS PROGRAMS COULD BE COMPARED WITH THEIR EXPECTED RECIDIVISM RATE. ESTIMATIONS OF PROBABLE NUMBER OF RECIDIVISTS CAN ALSO HELP CORRECTIONS OFFICIALS ALLOCATE MANPOWER AND OTHER RESOURCES. (GLR)