THE INDEX WAS DEVELOPED TO PREDICT THE RESULTS OF REFORMS IN THE ENGLISH PAROLE SYSTEM. IT WAS DEVISED USING MORE THAN 1,100 SUBJECTS AND WAS VALIDATED ON A SECOND, EQUALLY LARGE GROUP SELECTED RANDOMLY. USING A SIMPLE REGRESSION EQUATION, THE SCORES PRODUCED BY THE INDEX ARE CONVERTED TO PERCENTAGES OF LIKELIHOOD OF RECONVICTION WITHIN 2 YEARS; THE PREDICTIONS ARE FOUND TO COMPARE WELL WITH ACTUAL FIGURES. IT APPEARS THAT THE INDEX GIVES A GOOD PREDICTION OF THE LIKELIHOOD OF GENERAL RECONVICTION FOR NEARLY ALL TYPES OF PRISONERS, ALTHOUGH ITS BEST RESULTS ARE FOR MEN SENTENCED TO 4 YEARS OR LESS AND FOR PROPERTY OFFENDERS. IT COULD BE USEFUL BUT LESS ACCURATE FOR OTHER SUBJECTS. HOWEVER, EVEN WHEN USED IN THE EXACTING TASKS OF PICKING OUT PRECISELY WHICH OF THE MEN IN A GROUP WILL REOFFEND, THE INDEX PERFORMS WELL ENOUGH TO BE USED FOR RESEARCH AND ADMINISTRATIVE PURPOSES. IT IS INACCURATE IN THE PREDICTION OF THE RECIDIVISM OF SEX OFFENDERS, ALTHOUGH THIS MAY BE ATTRIBUTABLE TO SOME FACET OF THOSE OFFENDERS' BEHAVIOR OR TO THE SMALL SIZE OF THE SAMPLE; THE MOST PROBABLE CAUSE IS THAT THE BASE RATE OF RECONVICTION DEPARTS MARKEDLY FROM 50 PERCENT. DATA RELATING TO THE COMPUTATION OF THE PAROLE PREDICTION INDEX ARE APPENDED. NOTES AND REFERENCES ARE CITED. (KBL)