TO EVALUATE SEVERAL HYPOTHESES DERIVED FROM CRIMINALOGICAL THEORY FOR THEIR USEFULNESS AS DETERMINATES OF THE EFFECTIVENESS OF PAROLE, TWO CONTRASTING GROUPS OF PRISONERS WERE EXAMINED. THE FIRST GROUP CONSISTED OF RECIDIVISTS RETURNED TO PRISON AFTER THE COMMISSION OF ADDITIONAL CRIMES DURING THEIR PAROLE. THE SECOND GROUP HAD BEEN PAROLED AND HAD NOT EXPERIENCED ANY SUBSEQUENT LEGAL PROBLEMS. THE 96 MEN IN THE FORMER GROUP AND THE 56 MEN IN THE LATTER GROUP ALL COMPLETED QUESTIONNAIRES AND WERE INTERVIEWED. THERE WERE NO STATISTICALLY SIGNIFICANT DIFFERENCES FOUND BETWEEN THE MEMBERS OF THE TWO GROUPS ON VARIABLES OTHER THAN THOSE USED IN THE PREDICTION ANALYSIS. DATA WERE COLLECTED FROM THE 2 GROUPS ON 83 VARIABLES FOR THE PREDICTION RESEARCH, AND 15 VARIABLES WERE STRONGLY RELATED TO THE PAROLE OUTCOME. THE LENGTH OF CRIMINAL RECORD, THE IDENTIFICATION WITH CRIMINAL PERSONS, AND THE ORIENTATION TOWARDS CRIMINAL MEANS OF GOAL ATTAINMENT WERE CORRELATED MOST CLOSELY WITH SUCCESS OF PAROLE AS MEASURED BY THE AVOIDANCE OF RECIDIVISM. IT IS BELIEVED THAT THESE MEASURES CAN CONTRIBUTE SIGNIFICANTLY TO PAROLE PREDICTION EFFORTS. THE INTERCORRELATION BETWEEN ORIENTATION TO CRIMINAL MEANS AND IDENTIFICATION WITH CRIMINALS WAS LOW ENOUGH TO SUGGEST THAT THESE MEASURES ARE TO SOME DEGREE INDEPENDENT OF EACH OTHER. OTHER MEASURES SUCH AS THE DEGREE OF 'PRISONIZATION' AND ASPIRATION LEVELS ALSO MAY INCREASE THE ACCURACY OF PREDICTIONS OF PAROLE OUTCOME. THE LIMITED EFFECTIVENESS OF PREVIOUS ATTEMPTS AT OUTCOME PREDICTION IS ATTRIBUTED TO A LACK OF ADHERENCE TO THE BASIC PRINCIPLES OF SCIENTIFIC RESEARCH, AND THE CORRECT USE OF SUCH PRINCIPLES IN THE DEVELOPMENT OF RELEVANT VARIABLES SHOULD OVERCOME THOSE LIMITATIONS. (TWK)