THE TABLE IS BASED ON 55 STUDIES INVESTIGATING PREDICTORS OF RECIDIVISM IN 138 SAMPLES OF PAROLEES, AND ON 16 STUDIES OF PREDICTORS IN 39 SAMPLES OF PROBATIONERS. A NUMBER OF STUDIES REPORTED SUFFICIENT INFORMATION FOR TWO VARIABLES--TYPE OF INSTANT OFFENSE AND AGE AT FIRST ARREST--TO PERMIT DETAILED ANALYSES OF THESE VARIABLES. ALTHOUGH TYPE OF INSTANT OFFENSE IS A STABLE PREDICTOR OF RECIDIVISM, AUTO THEFT IS THE ONLY SPECIFIC OFFENSE CATEGORY CONSISTENTLY AND SIGNIFICANTLY ASSOCIATED WITH RECIDIVISM. THE RELATIVE EFFECTIVENESS OF OTHER SPECIFIC OFFENSE CATEGORIES AS PREDICTORS OF RECIDIVISM VARIES FROM JURISDICTION TO JURISDICTION OR FROM TIME PERIOD TO TIME PERIOD. FIRST ARREST BEFORE AGE 18 IS CONSISTENTLY RELATED TO RECIDIVISM, AND FIRST ARREST AFTER AGE 21 IS CONSISTENTLY RELATED TO NONRECIDIVISM. THE FREQUENCY TABLE INDICATES THAT THE FOLLOWING ITEMS ARE THE MOST STABLE PREDICTORS OF RECIDIVISM AND SHOULD BE INCLUDED IN PRESENTENCE REPORTS AND PROBATION AND PAROLE FILES: AN OFFENSE OF AUTO THEFT, PRESENCE OF PRIOR CONVICTIONS, STABILITY OF EMPLOYMENT, AGE AT FIRST ARREST, LIVING ARANGEMENTS, CURRENT INCOME, HISTORY OF OPIATE USE, AND HISTORY OF ALCOHOL ABUSE. CERTAIN OTHER VARIABLES, SUCH AS PRIOR PROBATION ORDERS AND TYPE OF JOB, APPEAR TO REQUIRE MORE INVESTIGATION, WHILE OTHERS (SUSPENDED SENTENCE, PROBATION, LENGTH OF PRISON TERM) DIFFERENTIATE RECIDIVISTS FROM NONRECIDIVISTS BUT MAY NOT BE USEFUL FOR TREATMENT-OFFENDER MATCHING. A LIST OF REFERENCES IS INCLUDED. (LKM)