TO INVESTIGATE THE HYPOTHESIS THAT PSYCHOPATHIC OFFENDERS WOULD SHOW LESS SUPPRESSION OF BEHAVIOR AS A FUNCTION OF PUNISHMENT AT VARYING LEVELS OF PROBABILITY, RESEARCHERS DEVELOPED A GAME WITH 10 DIFFERENT LEVELS OF PUNISHMENT PROBABILITY, PUNISHMENT BEING BASED ON THE RESPONSE-COST TECHNIQUES OF REMOVING REINFORCERS (CHIPS REDEEMABLE FOR MONEY). INMATE SUBJECTS TAKING PART IN THE EXPERIMENT WERE SCREENED ACCORDING TO THEIR SCORES ON A CHECKLIST FOR PSYCHOPATHY (THIS PROCESS IDENTIFIED 50 OFFENDERS SORTED INTO PSYCHOPATHIC AND NONPSYCHOPATHIC GROUPS OF 25 EACH); 19 NONOFFENDERS WERE CHOSEN AS A THIRD GROUP. A MEASURE OF SUPPRESSION WAS OBTAINED FROM THE REDUCTION OF SUBJECT'S RESPONSE RATE DURING THE GAME. TWO ADDITIONAL MEASURES WERE THE NET WINNINGS OF EACH SUBJECT AT THE END OF THE GAME AND THE SUBJECT'S TOLERANCE FOR PUNISHMENT CALCULATED BY SUMMING UP THE LONGEST CONSECUTIVE SEQUENCE OF PUNISHMENT CARDS PLAYED IN EACH OF THE 10 DECKS USED. RESULTS SHOW THAT PSYCHOPATHIC OFFENDERS EVIDENCE A GREATLY DIMINISHED RESPONSIVENESS TO PUNISHMENT WHEN THE PROBABILITY OF PUNISHMENT IS HIGHLY UNCERTAIN. ADDITIONAL DATA ARE PRESENTED TO INDICATE THAT THE RESULTS ARE BEST EXPLAINED IN TERMS OF COGNITIVE FACTORS, WITH THE ELEMENT OF MAGICAL OR SUPERSTITIOUS LOGIC PROPOSED AS A MAJOR PATHOGNOMIC CHARACTERISTIC OF PSYCHOPATHY. STUDY DATA AND REFERENCES ARE PROVIDED. (DAG)