TO DETERMINE THE VALIDITY OF FACTORS THE PAROLE BOARD USES IN PREDICTING PAROLEE SUCCESS DURING A 2-YEAR PERIOD FOLLOWING RELEASE, RESEARCHERS CLASSIFIED THE SAMPLE OF PAROLEES ACCORDING TO 21 PAROLE SELECTION ITEMS, INCLUDING THE FREQUENCY AND SERIOUSNESS OF THEIR PREVIOUS CRIMINAL RECORDS, TOTAL TIME SPENT UNDER LEGAL SUPERVISION FROM BIRTH UP TO ADMISSION TO THE WISCONSIN STATE PRISON, TYPE OF CRIME, AGE AT FIRST ARREST, TOTAL TIME SPENT UNDER LEGAL SUPERVISION SINCE BIRTH, NUMBER OF CITATIONS RECEIVED FOR MISCONDUCT WHILE AT THE PRISON, AGE AT LEAVING SCHOOL, AGE UPON ADMISSION TO THE PRISON AND RELEASE FROM THE PRISON, MARITAL STATUS, ETC. THE MEN WERE FURTHER CLASSIFIED ACCORDING TO WHETHER OR NOT THEY HAD BEEN ARRESTED AND CONVICTED FOR AN OFFENSE COMMITTED AFTER PAROLE RELEASE. CHI-SQUARE ANALYSIS WAS USED TO ASCERTAIN ASSOCIATIONS BETWEEN THESE SELECTION ITEMS AND RECIDIVISM. RESULTS INDICATE THAT MOST OF THE FACTUAL INFORMATION CONSIDERED BY THE PAROLE BOARD IS IRRELEVANT IN DETERMINING THE RISK IN PAROLING INMATES EXCEPT FOR ONE ITEM--THE TOTAL TIME THAT A PERSON HAS BEEN UNDER LEGAL SUPERVISION FROM BIRTH UP TO THE TIME OF ADMISSION TO THE WISCONSIN STATE PRISON. IT IS RECOMMENDED THAT THE PAROLE AUTHORITY BE PROVIDED A RESEARCH AND STATISTICS UNIT TO COORDINATE AND ANALYZE INFORMATION RELATING TO POTENTIAL PAROLEES SO THAT BETTER MEASURES OF PAROLE RISK CAN BE ASSESSED. ACTIVITIES FOR THIS UNIT ARE SUGGESTED. TABULAR DATA AND REFERENCES ARE PROVIDED. (DAG)