PROMPTED BY THE CONVICTION THAT POLICE DATA ON DELINQUENCY COULD PROVIDE A BETTER FOUNDATION THAN COURT DATA FOR AN INDEX OF DELINQUENCY, RESEARCHERS SOUGHT TO REFORMULATE THE PRINCIPLES USED BY POLICE AGENCIES TO COMPILE DELINQUENCY STATISTICS. THESE PRINCIPLES, AS REFLECTED IN THE STANDARD CLASSIFICATION OF OFFENSES (SCO), WERE FOUND TO HAVE CERTAIN SHORTCOMINGS. IN CRIMINAL EVENTS INVOLVING MORE THAN ONE SEPARATELY DEFINED CRIME, ONLY THE CRIME RANKING HIGHEST IN THE SCO HIERARCHY OF SEVERITY IS COUNTED. IN ADDITION, THE SYSTEM IGNORES THE VARIETIES OF CRIME THAT ARE CONTAINED WITHIN EACH CLASS. USING RECORDS FROM THE PHILADELPHIA POLICE DEPARTMENT'S JUVENILE AID DIVISION, RESEARCHERS CONSTRUCTED A REVISED CLASSIFICATION BASED ON OFFENSIVE JUVENILE EVENTS THAT CAUSED ACTUAL PHYSICAL HARM TO A VICTIM AND/ OR PROPERTY LOSS OR DAMAGE. WHEN THE REVISED CLASSIFICATION WAS COMPARED WITH THE SCO, IT WAS FOUND THAT ONLY 166 OF THE 504 EVENTS CATEGORIZED AS CLASS I (INVOLVING BODILY INJURY THEFT OF PROPERTY, OR DAMAGE TO PROPERTY) IN THE REVISED SYSTEM WOULD HAVE BEEN LISTED AS INDEX CRIMES BY THE SCO. THE PROCESS BY WHICH THE REVISED CLASSIFICATION WAS CONVERTED INTO INDEXES IS DESCRIBED BRIEFLY, AND DIFFERENCES IN THE INDEXES PRODUCED BY THE REVISED SYSTEM AND THE SCO ARE POINTED OUT.