ALTHOUGH ONLY A FEW STUDIES WERE AVAILABLE PERTAINING TO REVOCATION AND RECIDIVISM IN PROBATION, AND THEREFORE THE SAMPLE REVIEWED WAS SMALL, A WIDE DISPARITY IN THE DEFINITION OF REVOCATION AND RECIDIVISM EXISTED SUCH THAT REVOCATION/RECIDIVISM RATES HAVE LITTLE COMPARATIVE VALUE. IN ADDITION, CRITERIA OF PROBATION EFFECTIVENESS ARE NOT WELL DEFINED. A LONGITUDINAL SYSTEMS DESIGN IN THE RESEARCH, WHICH IS BELIEVED REQUIRED FOR ASSESSING PROBATION EFFECTIVENESS, IS ALSO LACKING. THERE IS CONFUSION IN THE STUDIES OVER THE DIFFERENCE AND RELATIVE IMPORTANCE BETWEEN PROBATION 'OUTCOME' AND PROBATION 'IMPACT'. 'OUTCOME' IS CONSIDERED THE SUCCESS-FAILURE RATE OF THE PROBATIONERS UPON RELEASE FROM PRISON, WHILE 'IMPACT' INVOLVES THEIR POSTPROBATION SUCCESS-FAILURE RATE. THE LINK BETWEEN OUTCOME AND IMPACT IS CONSIDERED CRUCIAL IN MEASURING THE EFFECTIVENESS OF PROBATION, BUT THIS IS OFTEN IGNORED IN STUDIES. A RESEARCH STRATEGY IS RECOMMENDED AS THE PRINCIPAL NEED IN THE SUBJECT AREA, BASED UPON COMPARABLE RECIDIVISM STATISTICS DEVELOPED SO THAT ALL PROBATION AGENCIES CAN HAVE A COMMON MEASURE FOR EVALUATING THEIR SERVICES. IT IS FURTHER CONCLUDED THAT A MEASUREMENT OF PROBATION 'SUCCESS' OR 'FAILURE' IS NEEDED THAT GOES BEYOND THE VIOLATION OF A CONDITION OF PROBATION. IN ORDER TO OBTAIN THIS OBJECTIVE, THE ADOPTION OF LONGITUDINAL EVALUATIVE STATISTICS IS RECOMMENDED TO REPLACE THE TRADITIONAL 'HEAD COUNT' STATISTICS. (RCB)