THE EVALUATION WAS CONDUCTED AFTER THE COMPLETION OF THE PROJECT AND DID NOT INVOLVE MONITORING ACTIVITIES. THE PHILADELPHIA POLICE DEPARTMENT'S JUVENILE AID DIVISION UNDERTOOK THE PROJECT UNDER THE ASSUMPTION THAT A RELATONSHIP EXISTS BETWEEN TRUANCY AND DAYLIGHT JUVENILE CRIME. THE PROJECT OBJECTIVE WAS TO PREVENT AND, THUS, REDUCE JUVENILE DAYLIGHT CRIME BY REMOVING TRUANTS FROM THE STREETS AND RETURNING THEM TO SCHOOL OR TO THEIR PARENTS. THE MAJOR THRUST OF THE PROJECT WAS APPLIED FROM DECEMBER 1976 THROUGH MAY 1977. THE 16 JUVENILE OFFICERS AND SUPERVISOR INVOLVED IN THE PROGRAM RECEIVED 112 HOURS OF INSERVICE TRAINING FROM POLICE COMMAND PERSONNEL AND REPRESENTATIVES OF THE DISTRICT ATTORNEY'S OFFICE AND COURT SYSTEM. THE FOLLOWING TOPICS WERE COVERED: CRIMINAL LAW, PROCEDURES AND POLICY, PSYCHOLOGY, DELINQUENCY, TRUANCY, INTEGRITY, COMMUNICATONS, GANG CONTROL, CRIME PATTERNS, RACE RELATIONS, REFERRAL AGENCIES, PREVENTION, COURT PROCEEDINGS, INTERVIEW AND INTERROGATION, SOCIOLOGY, INVESTIGATIVE TECHNIQUES, AND SELF-DEFENSE. VARIOUS GEOGRAPHIC LOCATIONS THROUGHOUT THE CITY, SELECTED FOR THEIR EXTREME CRIME CONCENTRATION, WERE THEN ASSIGNED TO EIGHT UNITS OF TWO OFFICERS EACH. TRUANTS FOUND ON THE STREETS WERE APPREHENDED AND RETURNED TO SCHOOL OR TO THE PARENTS. ON THE AVERAGE, 40 PERCENT OF AN OFFICER'S TIME WAS SPENT IN TRUANT INVESTIGATIONS; 39 PERCENT IN CRIME PATTERN INVESTIGATIONS; 19 PERCENT IN ADMINISTRATION; AND 2 PERCENT IN COURT. THE DATA COLLECTED REVEALED ITS EXACT TIME. TO CONDUCT THE EVALUATION, IT WAS NECESSARY TO ISOLATE JUVENILE CRIME BY TIME OF DAY. THE IDEAL DATA NEEDED TO PERFORM THE EVALUATION ARE REPORTED CRIMES ATTRIBUTED TO JUVENILE AND ADULTS INDIVIDUALLY. SINCE THIS DATA WAS UNAVAILABLE, THE SECOND BEST ALTERNATIVE IS TO USE ARREST DATA TO IDENTIFY JUVENILE AND ADULT OFFENSES AND REFER BACK TO THE CORRESPONDING REPORTED CRIME TO ESTABLISH THE TIME OF OCCURRENCE. ARREST DATA WERE COLLECTED AND, WITH THE AID OF A COMPUTER, CLASSIFIED IN TERMS OF THE FOLLOWING CHRACTERISTICS: POLICE DISTRUST, TIME OF DAY, TYPE OF CRIME, STATUS (JUVENILE OR ADULT), MONTH, AND YEAR. THE FLEXICAST COMPUTER PROGRAM WAS USED TO ISOLATE AND ELIMINATE FROM THE DATA PREDICTABLE CRIME COMPONENTS, SUCH AS SEASONAL VARIATION, TREND-CYCLE EFFECTS, AND AUTOREGRESSIVE EFFECTS. THE REMAINING DATA WERE ANALYZED. IT IS CONCLUDED THAT THE PROJECT HAD A POSITIVE EFFECT ON THE REDUCTION OF CRIME. OTHER ACTIVITIES IN THE COMMUNITY MAY WELL HAVE CONTRIBUTED TO THE SIGNIFICANT RESULTS. THE LIMITATIONS OF THE DATA USED FOR EVALUTION ARE DISCUSSED. RECOMMENDATIONS FOR FUTURE EVALUTATIONS ADVISE THAT THE FOLLOWING BE CONSIDERED: (1) DOCUMENTATION OF THE MAGNITUDE OF THE TRUANCY PROBLEM SHOULD BE SOUGHT; (2) A MEASURE OF THE AMOUNT OF CRIME ATTRIBUTABLE TO TRUANTS SHOULD BE OBTAINED; (3) REPORTED CRIME DATA SHOULD BE USED TO DETERMINE THE EXACT TIME AT WHICH CRIMINAL ACTIVITIES TAKE PLACE; AND (4) POLICE DISTRICTS SHOULD BE SELECTED AND MATCHED. DATA ARE TABULATED AND REPRESENTED GRAPHICALLY. (JCP)