ATTEMPTS AT APPROXIMATING AN EVALUATIVE EXPERIMENTAL DESIGN THAT USES OUTCOME MEASURES APPLIED TO RANDOMLY SELECTED EXPERIMENTAL AND CONTROL GROUPS USUALLY RESULT IN COMPROMISING BOTH THE STATISTICAL INTEGRITY OF THE EXPERIMENT AND THE INVOLVED AGENCY'S STANDARD OPERATING PROCEDURES. ADDITIONALLY, SUCH AN APPROACH IS EXPENSIVE, IN ITS USE OF BOTH TIME AND MONEY. IF PROGRAM EVALUATION IS TO BECOME THE RULE INSTEAD OF THE EXCEPTION, A MECHANISM MUST BE DEVELOPED TO PROVIDE STANDARDIZED PROGRAM EVALUATIONS WITHOUT THE HIGH COST OF DISTINCT DATA-GATHERING PROJECTS. A COMPUTERIZED OFFENDER-BASED TRACKING SYSTEM (OBTS) PROVIDES SUCH A MECHANISM. THE ARIZONA OBTS RECORDS A VARIETY OF INFORMATION ON THE CHARACTERISTICS OF ITS INMATES FOR THE PURPOSES OF IDENTIFICATION AND INSTITUTIONAL CLASSIFICATION. THE DATA BASE INCLUDES ALL OF THE ELEMENTS REQUIRED TO EVALUATE PROGRAM EFFECTIVENESS: (1) POST PROGRAM OUTCOME DATA, (2) PROGRAM PARTICIPATION VARIABLES, AND (3) OFFENDER CHARACTERISTICS VARIABLES. THE OUTCOME MEASURES FOR WHICH DATA ARE PROVIDED ARE COMPATIBLE WITH THE NATIONAL ADVISORY COMMISSION ON CRIMINAL JUSTICE STANDARDS AND GOALS 'CORRECTIONS' STANDARD 15.5, WHICH SPECIFIES RECIDIVISM (PRECISELY DEFINED) AS THE PRIMARY EVALUATIVE CRITERION. THE ARIZONA DEPARTMENT OF CORRECTIONS IS DESIGNING A SERIES OF COMPUTER PROGRAMS WHICH WILL PRODUCE STANDARDIZED EVALUATIONS OF A WIDE VARIETY OF ONGOING OPERATIONAL PROGRAMS. THE PROPOSED PROGRAM WILL CREATE AN EVALUATION TOOL WHICH ANALYZES THE INTERRELATIONSHIPS BETWEEN THE CHARACTERISTICS OF THE OFFENDERS AND TREATMENT PROGRAM THEY RECEIVE WHILE UNDER THE JURISDICTION OF THE STATE DEPARTMENT OF CORRECTIONS. THE SOFTWARE DEVELOPED UNDER THIS PROJECT IS DESIGNED TO BE EASILY TRANSFERRABLE FOR USE ON ANY DATA BASE WHICH INCLUDES DESCRIPTIVE DATA ON SUBJECTS AS OF THE TIME THEY ENTER A PROGRAM, DATA ON PROGRAM PARTICIPATION AND DATA ON POSTPROGRAM PERFORMANCE. THE MODEL IS DESCRIBED IN DETAIL. (RCB)