AS ORIGINALLY CONCEIVED, THE RESEARCH DESIGN CALLED FOR A STUDY SAMPLE OF 200 CASES IN EACH STATE, DIVIDED EQUALLY INTO 100 EXPERIMENTALS AND 100 CONTROLS BY A PROCESS OF RANDOMIZATION. HOWEVER, SEVERAL PROBLEMS OF ESTABLISHING THE MUTUAL AGREEMENT PROGRAMS, SELECTION OF CLIENTS, AND CLIENT ATTRITION LIMITED THE NUMBER AND NATURE OF SUBJECTS. IN SPITE OF THESE LIMITATIONS, THE RESEARCH WAS CARRIED OUT USING THE AVAILABLE DATA. BASED ON THESE RESULTS, THE AUTHOR HAD EARLIER CONCLUDED THAT THE CONTRACTUAL MODEL COULD BE USED AS A VEHICLE FOR ABBREVIATING THE PERIOD OF INCARCERATION, SINCE A SIGNIFICANTLY GREATER NUMBER OF EXPERIMENTALS HAD BEEN RELEASED WHEN COMPARED TO CONTROLS. HOWEVER, A REANALYSIS OF THE DATA ON CLIENT BACKGROUNDS SHOWED THAT THE EXPERIMENTAL GROUP ACTUALLY SERVED MORE TIME THAN CONTROL SUBJECTS, LEADING THE AUTHOR TO CONCLUDE THAT THE EARLIER RELEASE PHENOMENON WAS PROBABLY A SELECTION EFFECT AND NOT THE RESULT OF MAP. THE AUTHOR DISCUSSES THE IMPLICATIONS OF THESE FINDINGS TO THE FUTURE OF MAP, AND URGES CONTINUED DEVELOPMENT OF MAP.