THE POLICE JUVENILE DIVERSION PROGRAM OF RICHMOND, CALIFORNIA, IS USED TO DEMONSTRATE EVALUATION METHODOLOGIES. THE PROGRAM INCLUDES THE FOLLOWING COMPONENTS: POLICE CRISIS INTERVENTION AND MEDIATION; POLICE COORDINATED AND INITIATED REFERRALS TO COMMUNITY SERVICE AGENCIES; PAID TUTORING EXPERIENCE FOR BOTH OFFENDERS AND NONOFFENDERS; POLICE DRUG EDUCATION AND COUNSELING; SPECIALIZED COUNSELING BY SECONDARY SCHOOL COUNSELORS, POSITIVE POLICE INVOLVEMENT IN THE SCHOOLS, AND THE DISSEMINATION AND PROMOTION OF PROGRAM INFORMATION AND ACTIVITIES. FOR THE PURPOSES OF EVALUATION THE PROGRAM IS DIVIDED INTO HIGH LEVEL RELATIONSHIPS, WHICH HAVE TO DO WITH THE OVERALL OBJECTIVE OF THE PROGRAM; INTERMEDIATE LEVEL RELATIONSHIPS, WHICH HAVE TO DO WITH INDIVIDUAL COMPONENTS OF THE PROGRAM; AND LOW LEVEL RELATIONSHIPS, WHICH AFFECT THE EFFECTIVE FUNCTIONING OF THE VARIOUS COMPONENTS. EACH OF THESE LEVELS OF THE PROJECT ARE EVALUATED, USING THE PROCEDURE DEEMED APPROPRIATE FOR THE ACTIVITY CONSIDERED. EXPERIMENTAL, QUASI-EXPERIMENTAL, PREEXPERIMENTAL, AND EX POST FACTO EVALUATION DESIGNS ARE EMPLOYED. THE SIGNIFICANT RESULTS OF THE NONEXPERIMENTAL DESIGNS SUGGESTED BUT COULD NOT CONCLUSIVELY VALIDATE THE PROGRAM'S ROLE IN REDUCING RECIDIVISM. IT IS NOTED THAT ALTHOUGH THE EXPERIMENTAL PROCEDURES WERE SUCCESSFULLY INITIATED IN THE COMMUNITY SETTING, THE ATTAINMENT OF A SUFFICIENT NUMBER OF SUBJECTS FELL SHORT OF THE SIGNIFICANCE MARK. DATA TABLES ARE INCLUDED. (RCB)